Intelligent Machines 889 transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Leo Laporte [00:00:00]:
It's time for Intelligent Machines. Geoff and Paris are here. Our guest, Patrick Hillman, has a new kind of AI. It's not an LLM, it's based on physics. We'll talk about that. And of course, new models from Anthropic, OpenAI, the Chinese company Xiaomi. It's model extravaganza. And Paris says she's got a favorite.

Leo Laporte [00:00:23]:
Next on Intelligent Machines. Podcasts you love. From people you trust. This is TWiT. This is Intelligent Machines with Paris Martineau and Geoff Jarvis, episode 889, recorded Wednesday, September 23rd, 2026. We are the Neanderthals. It's time for Intelligent Machines, the show where we cover the latest in AI and robotics, and we have lots to talk about. Let me introduce our panel first.

Leo Laporte [00:00:57]:
Of course, the fabulous The wonderful Paris Martineau. We missed you, Paris, last week. So good to have you back. Feeling fit as a—

Paris Martineau [00:01:05]:
Glad to be back.

Jeff Jarvis [00:01:06]:
Fully oxygenated Paris Martineau.

Paris Martineau [00:01:08]:
It's true. I've got a lot of oxygen going to my brain and we don't know what that's gonna do to this show.

Leo Laporte [00:01:14]:
It could be anything. An investigative journalist at Consumer Reports where she covers poison.

Paris Martineau [00:01:22]:
True.

Leo Laporte [00:01:22]:
Foodborne illness. And toxins and all that stuff. It's good to see you. It's nice to have you back. We missed you. Jeff Jarvis is also here. I can't— never gives me a chance to miss him. You took a red-eye to get back here for the show, and I am very grateful to you.

Leo Laporte [00:01:39]:
Thank you, sir. Thank you.

Jeff Jarvis [00:01:41]:
If I fall asleep in the middle of the show, I hope you just ignore it. I did. I did.

Paris Martineau [00:01:46]:
Good. Yes.

Leo Laporte [00:01:46]:
Jeff is the, uh, former emeritus professor of journalistic innovation at the Craig Newmark Graduate School of Journalism at City University of New York. He's also the author of Hot Type.

Jeff Jarvis [00:02:01]:
So guess who I was on the plane with going out to San Francisco?

Benito Gonzalez [00:02:03]:
Who?

Jeff Jarvis [00:02:04]:
Craig Newark.

Leo Laporte [00:02:05]:
Don't say it.

Jeff Jarvis [00:02:05]:
And his wonderful wife.

Leo Laporte [00:02:06]:
Say it 3 times, he appears.

Paris Martineau [00:02:08]:
He'll appear, and I think that's beautiful.

Leo Laporte [00:02:14]:
Uh, Jeff, uh, wow, was that planned that you were on the same flight?

Patrick Hillmann [00:02:18]:
No, it's—

Jeff Jarvis [00:02:18]:
we happened to go out and we happened to be on the same flight, so it was great to see him.

Leo Laporte [00:02:21]:
That's kismet.

Jeff Jarvis [00:02:22]:
It is. Well, indeed.

Leo Laporte [00:02:23]:
What were you out here for?

Jeff Jarvis [00:02:25]:
I was moderating a panel for a new product at Handshake. A friend is there and asked me to just do it as a favor on education and AI, which was really interesting.

Leo Laporte [00:02:37]:
We have a very interesting guest. This is going to be part of, I think, a series of— we're going to be doing of AI that isn't LLMs, which is kind of an intriguing proposition. Our guest, Patrick Hillman, is with Logical Intelligence, spent 20 years as a a professor— professional explainer of things that were on fire. Crisis communications at Edelman, he was at GE, and then— and you may have heard his name at this time— he was in the hottest seat in crypto as Binance's first chief communications officer, later chief strategy officer, uh, during the FTX collapse. He left before the DOJ closed in, he's proud to say. Now COO and chief strategy officer of something that's not going to be nearly as crisis-laden, I hope, Logical Intelligence. But it is a little provocative. A San Francisco startup that says large language models are the wrong tool for anything where being mostly right means being wrong.

Leo Laporte [00:03:38]:
And they have something called an energy-based reasoning model, an EBM called Kona, which shipped earlier this year. Logical Intelligence was founded last year by a quantum physicist, Eve Bodnia, who's an expert on dark matter and particle physics. And Yann LeCun is the founding chair of its technical research board. Jeff's always bringing up Yann LeCun as the alternative to LLMs. Welcome. It's great to have you, Patrick.

Patrick Hillmann [00:04:06]:
Thanks for having me on. I'm excited.

Leo Laporte [00:04:08]:
So what's wrong with LLMs?

Patrick Hillmann [00:04:11]:
Nothing.

Leo Laporte [00:04:12]:
That's a relief because I'm spending a lot of money on them right now.

Jeff Jarvis [00:04:17]:
Yeah.

Patrick Hillmann [00:04:18]:
Look, I think there's a misconception that people believe that you're either going to have one or the other in AI. And we are religious as a company with the idea that the future of the AI ecosystem is just diversity of models. There's many different types of architectures. Most of us, when we think about AI, they automatically start thinking about large language models. Your audience, you guys know how they work. They create what we call chain of thought by guessing tokens, words in a phrase. They're really good at mimicking intelligence, but language is—

Leo Laporte [00:04:56]:
Mimicking is a key phrase, a key word in that sentence, actually. Mimicking, yeah.

Patrick Hillmann [00:05:02]:
Mimicking is a really important phrase there because when you ask an LM a question, it doesn't actually understand the question you're asking. It doesn't understand why you're asking the question. It can't go additional steps beyond just what the bare-faced question is and what it thinks should be the answer that it gives you.

Leo Laporte [00:05:21]:
Yeah, it doesn't have any understanding at all.

Patrick Hillmann [00:05:25]:
No, and that's a feature, not a bug, to be fair. They're not meant to understand. They're just not built that way.

Jeff Jarvis [00:05:32]:
We—

Patrick Hillmann [00:05:33]:
you train an LLM to be able to predict tokens more effectively by literally downloading everything you possibly can in text form, from the internet to now we're even scanning library books so that it can get better at it. But that's just not how we think about most tasks in our lives. When you get into your car and you drive and something— steps in front of your vehicle, you don't talk out loud. You are perceiving distance. You're perceiving temperature, possibly. You're thinking about grip on the road because it's raining. And you react immediately based on that. Language never comes into play.

Patrick Hillmann [00:06:07]:
And so because of the way that these LLMs are architected, they are really bad at what we call combinatorial tasks, when you have to think outside of just 1 plus 1 plus 1, and it's if this, then that, then this, then that, but not that, they fall apart.

Leo Laporte [00:06:25]:
Yeah, I fell apart. I couldn't quite follow that, but okay. They don't handle complex reasoning is what you're saying.

Patrick Hillmann [00:06:34]:
No, no, because they're not meant to. They're not meant to. And over the last couple years, the industry has tried to solve that problem because I think there was a lot of promise and a lot of excitement. Because when you first logged on to ChatGPT and you used it, it felt like a human was talking to you because that's what it's built to do. But the more you use it, the more the uncanny valley starts to slip in. The way I explain it to people is it's kind of like a really confident intern. I'm not knocking interns. I was one myself.

Patrick Hillmann [00:07:02]:
But it talks to you, and it sounds really smart, and it gives you an answer really quickly, and it's very confident in its answer. But if you know the topic, you understand that it's a little bit wrong. And because once— because they're operating in this linear fashion, it's thinking in a linear fashion, once it gets a little bit wrong, It very quickly gets very wrong. So the more complex the task, the more complex the decision-making tree it's creating, the more likely it is to hallucinate, and the more expensive it gets to try and solve those problems. So what do we do? We create agent layers, we pile agents on top of agents to try and take this specific language-based architecture and make it work against its nature, right? The way I try to explain to people is it's kind of like a Formula 1 car. Those things are amazing machines. They're the fastest on the planet. And if you have a flat road, those things can take turns at speeds that no other machine on the planet can.

Patrick Hillmann [00:08:03]:
But when you think about what an enterprise business environment is like, and this is where the rubber meets the road, for lack of a better term, they're not like an F1 track. It's more like an off-road track. And piling agents on top of these cars is basically like taking an F1 car and putting off-road tires on it, taking the spoiler off and putting some slightly better shocks, and expecting that it's going to be able to run a Baja race. It's not.

Jeff Jarvis [00:08:30]:
Yeah.

Patrick Hillmann [00:08:31]:
And so I think the market is now changing. So our belief is that you're going to have different types of architectures. I'll talk about one of ours in a little bit. But these architectures will use LLMs as an interface. You'll talk to it, you'll prompt it, and it'll give you feedback. But underneath the hood, there'll be different types of reasoning systems that don't utilize language at all. You're going to have world data companies like Jan's company, OMI Labs, which is just intelligence taking physical, real-world inputs and making it machine-ready so that an AI can actually utilize and think through it.

Paris Martineau [00:09:03]:
Yeah.

Patrick Hillmann [00:09:03]:
And then you'll have models like ours that have latent reasoning. Right now, we have a latent reasoning energy-based model that we use that'll basically provide constraints. on the AI system. And so constraint is really the name of the game. And that's the second point we're really religious about. AI has to be able to be constrained, not just from a safety standpoint, because it's obvious from a safety standpoint, but businesses can't have AI in their systems that isn't repeatable and predictable. That's why I think it was McKinsey just released a report maybe a month or two ago that said that around 80% of companies have some sort of form of deployed AI LLM-based unit within their company. But 90% of pilots that are actually tasked with critical infrastructure, critical networks, never got a pilot phase because you can't predict, you can't gamble on an outcome of a critical system.

Jeff Jarvis [00:10:01]:
And so you have to— The prediction machine is not predictable.

Patrick Hillmann [00:10:03]:
Exactly. The predictable machine is not predictable. Even though we know what the odds are of a dice roll coming up 6, It doesn't mean that if you roll that many times, 6 will always come up in that number. And that's the problem.

Jeff Jarvis [00:10:15]:
So I've read the papers about the energy-based model and—

Leo Laporte [00:10:18]:
Kona. This is the Kona model.

Jeff Jarvis [00:10:20]:
I would love for you to explain what the hell energy is.

Leo Laporte [00:10:23]:
Yeah, I would too.

Jeff Jarvis [00:10:24]:
And how it's different from the architecture that everybody's been paying attention to.

Patrick Hillmann [00:10:29]:
So energy-based models have actually existed for like 30 or 40 years. Jan McCune, again, in our company, on our board, has been working with energy-based models for a long time. What makes an energy-based model different— and I want to remind everyone here that I wear with a badge of honor that I am the dumbest person in my company. I'm one of the very few that don't have a PhD, a Fields Medalist, a Turing Award. So let's keep that in mind. But when you think about the difference between an LLM and an EBM, it really comes down to training and how it utilizes data to think. With an EBM, you take a very specific data set that you want to work within. And you take that dataset and you map it across essentially a representation of vector space, a physical landscape, as it were.

Patrick Hillmann [00:11:15]:
And then an engineer comes in and says, okay, I want these outcomes to happen. I don't want other outcomes to happen. There are things you want to do and things you don't want to do. The model then creates a physical landscape with outcomes we want being low energy points, and the outcomes we don't want being high energy points. And the model looks at the entire map of the landscape, and it scores every possible outcome. And it says, this has the lowest score because in physics and in mathematics, everything wants to lower its energy. And so that's how you have more deterministic AI output utilizing an EVM versus an LLM, which I discussed before, is a giant decision-making tree.

Jeff Jarvis [00:12:01]:
Can you give an example of how that gets applied?

Patrick Hillmann [00:12:04]:
Sure. So, for instance, right now, our company is working with a major materials discovery company. And material discovery, when you're trying to find a new molecule, is a giant combinatorial problem, right? You have to have a molecule that reacts in a certain excited state, and it has to be stable. Stability within a molecule has many different types of variables that interact all with one another. So you have to have a series of different events all happening at the exact same time for a molecule to be stable. So that company has a very specific data set that they've used based on their research for the last 40 years. And they drop it into our model. They walk through the different types of molecules and what the makeup would look like for that molecule.

Patrick Hillmann [00:12:52]:
And the AI learns the rules of the game, as it were. And then it helps them go and identify new types of molecules that abide by the rules that they set. And it's always the same. So that's honestly the simplest way of explaining how EBMs work. But it's not even really the secret sauce with us because EBMs have been around for 40 years. The problem with EBMs is, to a certain extent, the same problem with LLMs. The baseline EBM doesn't really understand the task. It's getting a score, and it's giving you the lowest score every time.

Patrick Hillmann [00:13:27]:
But if you want to take an EBM and have it, say, go from, uh, finding a new molecule and then trying to get it to, I don't know, um, manage your energy grid at your company. It's not going to be able to do that. You have to retrain it.

Leo Laporte [00:13:42]:
Yeah, so it's trained on a dataset, uh, that it then can be a classifier applied to that specific dataset, and it doesn't need to be trained on the world, correct? Well, you wouldn't even want it to.

Patrick Hillmann [00:13:57]:
You wouldn't want it, right? You wouldn't want it because—

Leo Laporte [00:13:59]:
Rather elaborate task.

Patrick Hillmann [00:14:01]:
The problem with LLMs, right, they create giant signal versus noise crises.

Leo Laporte [00:14:05]:
Right.

Patrick Hillmann [00:14:06]:
And the more noise you put into it, the more expensive it becomes to find the signal.

Leo Laporte [00:14:11]:
How— so how is the training done? Would I— so for instance, let's say I wanted to employ Kona at my business, my robotic automobile assembly line. I would obviously have to train Kona on the physicality of that assembly line. Is that right?

Patrick Hillmann [00:14:31]:
Well, in an industrial use case, you— again, going back to the more diverse AI ecosystem, you'd have to have world models that are out in the environment that is taking complex, noisy inputs from a factory setting and then dropping it into the reasoning system to then go and actually reason through. But it's the exact same process with our molecular companies as it will be with robotics companies. They have all of their data. They want an arm to move this way. They have all the data and speed, energy retention within that robotic arm, et cetera. They're mapping all of that known data just like it sits in their current systems today. And then the real bulk of the training is really around the specs, helping the model understand the rules of the game, a win and a loss scenario. And so this is why we can't say that even this new LLM, these new non-language-based models are gonna be completely deterministic, right? Because they can also make mistakes because an engineer might not completely understand the spec, but you train the model to understand the specs as best as you can.

Patrick Hillmann [00:15:38]:
And then as it makes mistakes in training, you learn to update your specs and to change the scoring across your physical space.

Leo Laporte [00:15:45]:
This reminds me a little bit of, there's been a lot of attention in the last couple of weeks of something from a company called typesafe.ai called Jev. I'm sure you're very familiar with it. very well aware of it, which is also not an LLM, it's a classifier. Uh, it's a similar situation where you, you train it and then it, it can give you probabilities, uh, for the next step. It's very fast, it's very cheap, and you really wouldn't want to use it without an LLM. It's, it's a, it's an input to an LLM. Is this similar to that?

Patrick Hillmann [00:16:15]:
Uh, different tasks would be applied to it, but yeah, I mean, it's the same general theory of, of having models that are built from the ground up for very specific tasks based on their architecture. So because ours utilizes theories of mathematics and physics, when you think about what our models are really good at, it's usually mathematic and physics-based theories. It's not good at language. You wouldn't want to go and use it for language. So it's not a competitor to LLMs. It's going to be a partner with major LLMs.

Jeff Jarvis [00:16:46]:
So are LLMs a front end to this in that sense, in the sense that We can speak to them in language.

Patrick Hillmann [00:16:52]:
There'll be an interface.

Leo Laporte [00:16:54]:
I'll give you an example of how I'm using Jev is as a classifier for the news stories I pick for these shows. And right now it's in a training process. So I was using a different system to do this, but I have Jev and I have my LLM working together to say, this is a story Leo picked. This is one he didn't pick. And in theory, after a period of time, it will be able to score stories somehow magically based on that information. Yours sounds like it's more of a physics-based, a physical world-based philosophy.

Patrick Hillmann [00:17:27]:
It's more of a physics and mathematical-based scoring system. Okay.

Leo Laporte [00:17:31]:
Similar idea, though.

Patrick Hillmann [00:17:33]:
Similar idea, similar approach when you think about just how do you want AI to address different problems in different ways? How do you want it to think differently?

Jeff Jarvis [00:17:42]:
So how do you tie into world— since this is Jan's involvement— to world models?

Patrick Hillmann [00:17:47]:
Well, world models, I think, is really probably the most exciting and the most high-impact area of AI discovery that is ongoing right now. Because coming from General Electric, manufacturing floors specifically, but any sort of industrial environment, it's very noisy. It's very noisy. And to be able to have an AI system that is able to respond and react in real time to very dynamic environments is a real significant challenge. I think that entire class of AI will be in development phase for at least the next 5 to 10 years before we have perfected systems. But that's when you're going to start to see really true physical AI being deployed in a manner that's going to be really high impact. Today, systems like ours are going to be able to provide a more constrained software-based AI solution. for companies.

Patrick Hillmann [00:18:44]:
But before you can start really rolling these things out to manage robotic systems, you're going to have to have that world data, world model sort of area of AI kind of figured out.

Jeff Jarvis [00:18:55]:
Which is really exciting. And to hear Yann talk about it, it can make jet engine turbines and worry about molecules and all of this. In the cultural consciousness, when ChatGPT spoke our language and listened in our language, that's when everybody got excited and freaked. At the same time, is there a consumer touchpoint that's going to come with these kinds of models that will make people say, oh, that's all different, that's really cool? Or is it so technical and so apart from our experience that it's going to be operating in the background for quite some time?

Patrick Hillmann [00:19:32]:
I think the problem in how we think about AI today Just because you touch it and you interact with it directly doesn't mean it's actually impacting your life more directly. Once we start to have— so for instance, with our materials discovery company, we were able to discover, I think it was around 200,000 new synthetic molecules in around 3 days.

Jeff Jarvis [00:19:59]:
Wow.

Patrick Hillmann [00:20:00]:
Which means that you're going to be able to essentially turbocharge research and development. All of our lives should, just behind the scenes, just start getting better. Ideally, it's going to help speed up things like cancer research eventually. Not because AI is just going to discover the cure for cancer, but it's going to make our researchers that much more efficient in where they spend their time in their research.

Jeff Jarvis [00:20:29]:
That's great. I agree. I think that's important to emphasize. But right now, AI has nuclear cooties.

Leo Laporte [00:20:37]:
Before you get to that, Jeff, hold on a second. Let's not go there. We're talking to Patrick Hillman. Logical Intelligence is the name of the company. You can go there and learn more about Kona, their first model. Certainty, not probability, at logicalintelligence.com. I have an idea, Jeff, that might help understand this. Self-driving vehicles, perfect example of a very noisy environment.

Leo Laporte [00:21:04]:
They can do about, they get to 90, 95%. It's that last 5% of unpredictable—

Jeff Jarvis [00:21:09]:
And you're standing in front of them.

Leo Laporte [00:21:11]:
Yeah. Stuff that happens that makes it— and we've seen, there are a lot of places where this is the problem, the 95% problem, where you can get almost there. It's that last 5% that's really really hard because probability just doesn't get you there. So this sounds like something Kona might be one of the places you might see something like Kona involved.

Patrick Hillmann [00:21:37]:
The entire, you know, Kona is just the first model that we've rolled out this year, and we're now piloting with 3 different companies. We have a new model that's going to come out that isn't energy-based that also utilizes what we call latent space. But in the next 3 to 4 years, I think you're going to see a lot alternative architectures that are going to come out and start being tested on things like driverless vehicles. But even with driverless vehicles, you're going to have the same problem. We need to have better world data and world models available, which is why we're still completely dependent on LiDAR today.

Paris Martineau [00:22:04]:
Right.

Patrick Hillmann [00:22:05]:
You just don't have the systems in place today to reliably even get good external data into an AI system to be able to ensure that it's making good decisions, that it's abiding by the rules that we set for what a good, safe driver would be. It's just, it's just a while away.

Leo Laporte [00:22:23]:
You do it in math too. I see you have— I don't even know what it means.

Jeff Jarvis [00:22:28]:
You—

Leo Laporte [00:22:29]:
Kruskal discovered the first new good group in 31 years. God bless him. But that's a mathematic— it's a geometry issue. Yes.

Patrick Hillmann [00:22:37]:
So what does that mean to the average person, um, and why is it important to the average person?

Leo Laporte [00:22:42]:
That's why I brought up self-driving cars first.

Jeff Jarvis [00:22:44]:
Yeah, yeah.

Patrick Hillmann [00:22:46]:
Actually, let me tell you what's actually important about that discovery? What's most important? And my founder actually talked about it a little bit. She did a little upfront, a foreword, because in recent weeks, we've seen a number of the major frontier labs make these big announcements about math discovery utilizing AI.

Leo Laporte [00:22:59]:
Right.

Patrick Hillmann [00:23:00]:
And the response from the market was, oh my God, this is going to replace mathematicians. But if you look at how these problems are being solved, it is an intimate mix of both foundation models, good orchestration layers, and always at the heart of it, is the research team that's pushing these things forward. And what Slava and our team did was take a problem that hadn't been moved in any way, shape, or form, advanced in over 3 decades, and was able to now go and verify these 3 new forms, these 3 new mathematical forms, simply by utilizing AI tools at his disposal to go and solve this really critical problem. And so the actual story coming out of this and all of these math you know, discoveries that are happening inside discoveries. It shouldn't be AI is solving these things. It's that we're learning as human beings how to utilize AI to get over some of the humps that have existed that have kept us from discovering more and more complex problems.

Jeff Jarvis [00:23:57]:
All right.

Leo Laporte [00:23:58]:
You published a piece on LinkedIn that I think this is how we first became aware of you, actually. LLMs won't kill us, but they will destabilize the global economy. If Frontier Labs refused to be honest with themselves and the general public, and I have a feeling this has a little bit to do with your past as a crisis PR guy, you couldn't have a worse crisis. And I, boy, I know that Paris and Geoff really want to ask you about this too.

Paris Martineau [00:24:25]:
Yeah, Leo, you're cribbing from our questions.

Leo Laporte [00:24:27]:
If you were Dario Amodei and you're sitting there at Anthropic and you've got this story to tell, It seems to me you're walking on a knife's edge here, that you're really risking, um, turning the public against you, turning— having the government nationalize you, and having a really great IPO. And you gotta kind of walk a fine point. What, what is this destabilize the global economy, and what, what do Frontier Labs need to be honest So first of all, the biggest problem I had with the approach that was taken in this—

Patrick Hillmann [00:25:09]:
let me step back. The reason why I jumped into AI and jumped into logical intelligence, period, is that I'm sure many of you have read Mustafa Suleyman's book, The Coming Wave.

Leo Laporte [00:25:25]:
Yes.

Patrick Hillmann [00:25:26]:
He talks about a theory that I thought was very important and is still to this day, as I talked about earlier, is kind of what underpins Logical Intelligence. It's around the idea of constraint. If companies are building dangerous things in the bowels of their labs, you don't have to wait for government to come and stop you. You should stop doing that right away. And I think we should also understand how poor Government is— not just our government, this isn't a political statement. These are massive institutions. They don't move quickly. They're designed to not move quickly.

Patrick Hillmann [00:26:05]:
It's a feature, not a bug. But the tech is evolving in a way that government is not going to be able to step in on day one and just start legislating in Congress to ensure it's done safely. There is also a little bit of validity behind the, we need to be first in AI. Like, that does make sense. But that doesn't mean that companies get to walk away from their responsibility to the general public. Hyundai, 3 or 4 years ago, had to recall 40,000 of their EVs because the car, when the driver would take its foot off the brake, would sometimes just decide to speed up on its own because of a glitch in their software.

Paris Martineau [00:26:44]:
Not ideal.

Patrick Hillmann [00:26:45]:
Not ideal. So what did the company do? They notified their regulator, and they initiated a recall, and they brought all those cars back. What they didn't say is they didn't go to the market and say, you guys aren't going to believe this. Our new car is so smart, it can drive faster on its own. That's a failure of constraint.

Benito Gonzalez [00:27:07]:
Yeah.

Patrick Hillmann [00:27:07]:
And when we talk about and we see the marketing around some of the recent issues like Hugging Face, and before that with Mythos, it's always been kind of the same story. Great news, everybody. We've built this giant robot in our basement, and it's so smart, it's learned to escape the lab, and great news, it's also learned to kill. That's not— that's a failure in its product, but spinning it as this thing to be idolized, I think, is either really a mistake, or if I'm extraordinarily pessimistic, I say that it's a little bit opportunistic as well, because we are at this moment when people feel that AI is not able to deliver on the promises that it's had. They're about to IPO, and I think it's just helpful to be able to have this continue to be front-page news across the country, and Congress will be talking about it every single day. It doesn't mean it's the right thing to do, though.

Paris Martineau [00:28:08]:
Yeah, I mean, the thing that struck me about all of this is these large frontier labs have Large teams of very accomplished, experienced PR, public affairs, and crisis management people of their own. It's not like they're flying blind. It's a bit strange that repeatedly, the modus operandi for all of them has been to take this somewhat negative-seeming path in which the through line that they keep repeating to to the press, to consumers, to regulators is, we've created something that could kill you, kill the things you love, and take your jobs. I mean, why do you think that these companies have opted for such a kind of strange approach to communication?

Patrick Hillmann [00:29:04]:
So I will give you my honest answer. Spending, you know, 20 years in crisis management, working on everything from, uh, Penn State in the early days with the Sandusky issue, to helping manage the OPM crisis for the Obama administration, to working with some of the vaccine manufacturers through COVID.

Leo Laporte [00:29:22]:
Wow, you have really been on the front line. I was gonna say, I've worked on a few crisis management. Oh man, you're the guy.

Patrick Hillmann [00:29:30]:
What I'll tell you, what I'll tell you is that What I learned in those years was that there are very few evil companies, very few, very few. What you have to think about is that when people get into large groups, we act stupidly.

Paris Martineau [00:29:46]:
That's right.

Patrick Hillmann [00:29:46]:
And companies are just giant groups of people. And almost always, when a company has a crisis, it's not because they were purposefully trying to harm. It's because they just weren't thinking straight. When I went to Binance, I walked into a company that had received the DOJ target letter. It was also a company that went from being a startup to 6 months later being the largest crypto exchange in the world and larger than most banks here in the US. To grow and scale a company that quickly and to do it well, is virtually impossible. The task of hiring, understanding who you need to hire that quickly to manage the expectations of a global bank that has been in operation for 200 years, to do that overnight, it's impossible. And because you're forced to hyperscale, you make massive— you may create massive problems in your corporate infrastructure, in corporate governance, silos get created.

Patrick Hillmann [00:30:53]:
You don't have a complete understanding of what's happening in your business. And oftentimes, even when you hire the right people, the company is growing so quickly that very quickly that job is now too big for the person they had hired to do it. And this is why companies who hyperscale make huge mistakes.

Benito Gonzalez [00:31:11]:
Hmm.

Patrick Hillmann [00:31:11]:
I would look back to the social media companies and how they kind of manage their policies and how they manage um, their algorithms and targeting consumers and what's— and what that has done. I would look at AI today and I would tell you that I'm seeing the exact same things, same mistakes.

Leo Laporte [00:31:28]:
What advice would you give Dario if you were his crisis PR guy? If he called you up today and said, we need you, Patrick.

Patrick Hillmann [00:31:38]:
So I've—

Leo Laporte [00:31:38]:
which he does, by the way.

Patrick Hillmann [00:31:40]:
I do have a personal relationship with one of the major lab leaders, and I have talked to them about this.

Paris Martineau [00:31:44]:
OK.

Patrick Hillmann [00:31:44]:
And I've told them that they need to be, number one, much more transparent about what they're building and why. The problem with LLMs is they're a giant black box in how they are built. And so you're asking for the general public to give a lot of faith in Silicon Valley that they're going to work in their best interest. And again, you can look at the trust measurements that come out of companies like Edelman that I used to work for. You're going to find that After Washington, D.C., the next most untrustworthy institution in this country according to the general populace is Silicon Valley. And you have to also think about the mentality of the average American who, particularly those that are older and remember NAFTA, I think this is like one of the most forgotten, interesting, and pertinent parts of American history that we need to think about again when we talk about AI, because Yeah. NAFTA was this hot topic in 1992, '93. You had George Bush, incoming president, running against Bill Clinton.

Patrick Hillmann [00:32:45]:
You had Ross Perot sitting in the center. Bush was, I am rah, rah, rah. We're going to sign NAFTA. Bill Clinton's like, ah, I'm going to wait and get elected, and we'll see. I'm going to take a really smart approach to this. We'll wait a couple of years. And then Ross Perot was out there beating the drum. Look, it's going to cost jobs.

Patrick Hillmann [00:33:02]:
Things are going to move, and this government is not ready to create a safety net for those employees and citizens that are going to be impacted by this. Well, Clinton ends up getting elected because of Ross Perot, and what was the first thing Clinton did 2 months into his office? He just went and signed NAFTA, and he was sitting, I think, in a facility in Pennsylvania talking about it, and what he said essentially was— I'm paraphrasing— was that people who lose jobs in industries like mining and steel, that they were going to be trained to work as coders in the new silicone economy.

Leo Laporte [00:33:35]:
Shh!

Patrick Hillmann [00:33:36]:
And I don't know about you, but I have not come across too many coders that, you know, worked the manufacturing line in Chillicothe, Ohio, or in Schenectady, New York. And so I think there's a lot of natural apprehension towards these products, and there's good reason for it. And if you're a lab, you need to be more transparent what you're building and why you're building it and what you're willing to stop for.

Jeff Jarvis [00:34:02]:
But Patrick, if you— if they answer that question today, I'm afraid what you're going to hear is— pardon me, Leo— test real. We're building it to build a superintelligence that's going to take over the economy and give everyone leisure and all this BS.

Paris Martineau [00:34:19]:
And we have to do it first so that the bad people don't do it before us.

Jeff Jarvis [00:34:22]:
Exactly. And we want to pull the ladders up so that they can't come and compete with us. So I'm afraid they're honest. I'm coming more and more— I think it's an amazing technology. I think it can do phenomenal things. I don't want bad policy to stop us from getting those amazing things. But I'm afraid that too much of it in the public eye is in the hands of the wrong people. And they can't— they're doing the wrong things for the wrong reasons over a very smart infrastructure, and they're going to screw it for everybody.

Patrick Hillmann [00:34:57]:
The 2 Frontier Lab leaders that I talked to, are not sociopaths. They're definitely not stupid. They are smart, and they're much more thoughtful than I think people would give them credit for. But fundamentally, they believe in what they're doing, and they're under tremendous pressure to continue to justify their valuations, right? I mean, we live in an economy that is capitalistic in nature. It is what it is. But that's why you're supposed to have Government's supposed to be there to be a checks and balance system. But again, right now, government is not prepared. It's not able to move quickly enough to catch up to the speed with which AI is moving.

Patrick Hillmann [00:35:38]:
I would love to tell you that I have a perfect way of fixing this, but unfortunately, I don't think there really is. Either you have to make the decision to slow down tech and lose potentially a valuable head start on some of our foreign competitors, or you risk careening towards disaster. Now, what I would say is a lot of the AI doomerism around this stuff is really overblown. I'm sorry, LLMs are not about to start launching nuclear weapons. It's just, it's not going to happen. It just isn't. But I do think there will be job displacement, continued job displacement, and I don't think we have the proper safety nets in place across the globe for workers that are going to be impacted by this. And that's probably the biggest danger and risk that AI poses to us today.

Patrick Hillmann [00:36:24]:
And it's a very big one. It's a very real one, because with massive economic upheaval comes all sorts of other really nasty impacts, from civil unrest, increased crime rates, to, you know, even potentially domestic conflicts. So it's really important that we get this right. But Congress is not going to be able to go and do that. And if someone put a gun to my head and say, how do you fix this? It would have to be a partnership between the major labs as well as government to do basically 3 things. I think, one, they need to lay out exactly what the risk is very transparently, both for government and the general public. Number 2, they need to have an agreement on what they're going to stop working on today and identify where the sort of vectors of danger are. And number 3, government has to create something a new regulatory model, or maybe go back to an old regulatory model, like the brain trust that was created by Franklin Roosevelt, where we bring together a group of potentially 2 or 3 elected officials, 3 to 5 academics, and maybe 2 or 3 people that are voted in by the general public to basically serve as a committee to oversee what is going to be our national AI strategy and how are we going to ensure we mitigate any potential damage the general populace with it.

Jeff Jarvis [00:37:39]:
And not just the CEOs of those companies, which is what Congress wants to pull in this judgment.

Patrick Hillmann [00:37:42]:
I don't think the CEOs of any company should be anywhere near this, including us.

Leo Laporte [00:37:45]:
And I think it should probably be a global effort as opposed to a national effort. Unfortunately, you just named 3 groups that are the least trusted groups in America: academics, politicians, and CEOs. I think you're— I think you've got a problem. Trust is a big issue right now, and I don't think the American public trusts Anybody. And I don't know if any crisis PR can solve that. Patrick, I really appreciate it.

Jeff Jarvis [00:38:09]:
Can I ask one more question real quickly, Leo?

Benito Gonzalez [00:38:12]:
Sure.

Leo Laporte [00:38:12]:
Do I have a choice?

Jeff Jarvis [00:38:14]:
No, no, because I talk faster than you do. You come into this with communication skills, obviously, it's what you do for a living. You're very impressive today in explaining this really complicated stuff. How the hell did you learn this? We're trying to learn it all every week. We're trying to stay on top of it every week. But you're in with, yeah, Fields Medal winner and amazing people who do this stuff. It's not simple to digest it and then distill it the way you do for a living.

Patrick Hillmann [00:38:45]:
How did you—

Jeff Jarvis [00:38:46]:
what was your learning curve like?

Patrick Hillmann [00:38:49]:
The second day that I was at the company, we were working out of a house outside of San Francisco, and I was sitting in the kitchen around a table with with Michael Friedman, Fields Medalist, Eve, our founder, again, PhD quantum physics, um, Ian LeCun, and then 2 other PhDs that work for us. And I didn't understand anything they said at the kitchen table for an hour, and nor do I today. What actually really helped me understand all of this was once we announced Kona and what it did, we had around 20 Fortune 100 companies reach out to us and start to ask us to do POCs. And now we have design partnerships with them. And in embedding with those companies and understanding the problems that they have—

Jeff Jarvis [00:39:29]:
Wow.

Patrick Hillmann [00:39:30]:
And then bringing engineers together to actually talk about, okay, if these are the problems, what types of solutions can a different architecture bring to the table? That's actually how I learned how all of this works because you actually see it being deployed and you hear from the mouths of engineers working on the front lines of these enterprises on what doesn't work in existing today. And our engineers are able to sit them and help them understand, like, what is actually feasible.

Jeff Jarvis [00:39:53]:
How did Yves and Jan know they needed to hire you, uh, that your skill set was necessary?

Patrick Hillmann [00:40:00]:
So, um, after I left, uh, Binance, I went to— I helped co-found a small AI startup that was acquired while we were still in stealth. And just because I was in AI, a lot of my friends in the manufacturing space from GE and the National Association of Manufacturers, Dow, Boeing, they would reach to me all the time and just ask me about AI, what's real, what's not real, because they were under tremendous pressure to get more AI into their systems from their CEO and from their investors, and they couldn't do it. And so eventually, Eve was introduced to a VC that I had known through the General Electric network. And he gave me a call and said, hey, they're working on something really interesting that goes back to our GE days. You should sit down with her. So, I sat down with Eve in a small basement cafe in New York. And for 3 hours, she walked me through how this works, why it works, and where she thinks there's market pressure to buy. And she asked me if I agreed with her that companies would want this.

Patrick Hillmann [00:40:55]:
And I literally asked her at the table that day to let me work for her. And that was it.

Jeff Jarvis [00:40:59]:
Cool. Thank you.

Leo Laporte [00:41:01]:
Thank you, Patrick. Patrick Hillmann, logicalintelligence.com. I appreciate your time. It's a fascinating subject. And I think if anything, that next year we're going to see a lot of alternatives to LLMs emerging as people realize that LLMs by themselves aren't quite enough. And I think this is one, one direction, Kona. Very interesting. Thank you, Patrick.

Benito Gonzalez [00:41:23]:
Appreciate it.

Patrick Hillmann [00:41:24]:
More to come. See you guys soon.

Leo Laporte [00:41:25]:
Yes, I always say that. More to come right after this word. Well, they may say they're pacing the frontier, they're slowing down. It sure doesn't Feel like it. This week Anthropic introduced Opus 5.5.

Paris Martineau [00:41:40]:
What have you thought of it so far, Leo?

Leo Laporte [00:41:41]:
Superb. It is actually the best. You agree? Yeah, it's the best version of Opus yet, if you ask me. And certainly some of the problems I had with Opus 5 are gone.

Paris Martineau [00:41:54]:
Yeah, it does seem like a real leap forward, especially in terms of Written communication. I mean, from a very silly perspective, I've been really chuffed to see everybody on Twitter today playing around with Opus 5.5 has the ability to draw and do kind of animation in a way that is quite charming, which is not a word that I thought I would ever use to describe AI artistic output, especially just in comparison to— I was thinking about that. Do you remember when I made that A coffee tasting thing? No, a goofy game of— with Claude, of tomatoes being thrown at your face.

Leo Laporte [00:42:38]:
It's come so far in 8 months.

Paris Martineau [00:42:40]:
I looked at— I looked— it was earlier this year we made that, and it was so bad. It was one of the worst animations I think I've ever seen in my life. And now we've come so far.

Jeff Jarvis [00:42:54]:
And Claude was, was, was, was a coding man.

Patrick Hillmann [00:42:58]:
LLM.

Leo Laporte [00:42:58]:
And this is still, I would say, a coding LLM.

Jeff Jarvis [00:43:01]:
It is, but it wasn't doing the visual the way it's now doing it.

Leo Laporte [00:43:04]:
Oh, no, no.

Jeff Jarvis [00:43:04]:
It's expanding.

Paris Martineau [00:43:05]:
Yeah. I mean, it's still a coding. It's not really designed to do visuals or animation in the way that any of the other models are. But going from the stick figure that I just posted in our Discord of a tomato emoji being thrown at a couple of concentric circles to what people are now seeing with Opus 5.5, it's really astounding. I mean, that was January. I thought I was looking for it because I was like, when was it? Was that like 2 years ago or something? I remember it being really bad, but no, that was earlier this year.

Leo Laporte [00:43:39]:
We've come an amazing way. And what Anthropic is saying in their defense is we are pacing it. It was tested by external evaluators before release, including METER. And of course there are issues with METER. They're heavily uh, Tescriald, and Anthropic is an investor.

Jeff Jarvis [00:43:56]:
And lots of connections. Yep.

Leo Laporte [00:43:58]:
Yeah, and Frontier Design. Um, but, you know, at least they're doing something. And, um, I think they say they— the classifiers they have on it are more for cybersecurity than anything else. I've not run into any, uh, refusals. Um, so, you know, whatever it's doing, it does it. And when it classifies you as doing something it doesn't want to do, it just steps you down a lower model. So, uh, and it's gonna cost less. It costs 40% less than Opus 5.

Leo Laporte [00:44:27]:
So I think this is a win-all round.

Paris Martineau [00:44:28]:
I did a really complicated task earlier today. I mean, it maybe took like 30 minutes on it on high and it barely used any of my usage.

Leo Laporte [00:44:36]:
Yeah. It's, uh, I'm using it for coding 100%. It's funny because Grok 4.7 also came out this week from Elon Musk's xAI. And it actually feels like it got dumber. I don't know how that's possible.

Paris Martineau [00:44:53]:
I mean, I feel like that might just be baked in.

Leo Laporte [00:44:56]:
Well, Elon says this one is trained on all the business documents, all the workflow documents that he got from SpaceX. So in theory, it's got a lot— I mean, I don't know.

Jeff Jarvis [00:45:09]:
Rocket science.

Leo Laporte [00:45:10]:
It's got rocket science.

Paris Martineau [00:45:12]:
In theory, it's got rocket science in there, but in practice, I guess it just has told us something about how SpaceX workflow actually is behind closed doors.

Leo Laporte [00:45:22]:
And OpenAI updated their models. They're doing ChatGPT-6 now. Sol and Luna, also lower cost. I haven't noticed a big difference. I mostly use Astra from OpenAI, so maybe I need to spend some more time When you get a—

Jeff Jarvis [00:45:40]:
when a new model comes out, what's the first thing you do with it to get a sense of it?

Leo Laporte [00:45:43]:
You know, it's funny. First thing I do is go to Twitter because there is a cadre of people at X.

Paris Martineau [00:45:52]:
You're telling me there are people on X, the everything app, talking about AI?

Leo Laporte [00:45:56]:
It's really become the place to go for AI, hasn't it? And so what will happen when a new model comes out is immediately people try their various Everybody has a little benchmark. Some of them are dopey. I mean, I really, I don't care if you can make it—

Paris Martineau [00:46:10]:
There's someone who does something with otters, right?

Leo Laporte [00:46:13]:
Yeah, yeah, there's the otters, the otter eating in an airplane.

Paris Martineau [00:46:17]:
Yes, an otter eating in an airplane.

Leo Laporte [00:46:20]:
Simon Willison does a pelican on a bicycle. Both of those now, I think, are kind of superannuated. There are people who do voxel pagodas. Mia, who I follow, her recipes seem to be the best recipes, does spend 458 virtual dollars in tokens having it do what she does with a lot of models, which is create 100 HTML templates. So 100 new— let me see if I can find her post on here— 100 new HTML templates. And what's nice about that is you can see You could— it's a head-to-head comparison with all the current models out there. She said it's the best she's ever seen. She was, she was blown away.

Leo Laporte [00:47:07]:
She said Opus 5.5 is just like, wow. And I have to say, if you look at the 100 templates it made, they're very creative. They're beautifully designed. I think that's perhaps because of its strong visual capabilities.

Paris Martineau [00:47:23]:
I mean, have you played around at all with Claude Design, which was available before 5.5.

Leo Laporte [00:47:28]:
Claude Design's fantastic.

Paris Martineau [00:47:29]:
It is fantastic. And it's also, I mean, customizable in a way that I think is just very useful, intuitive.

Leo Laporte [00:47:36]:
One of the things I did when Astra came out is I had it design a website that describes my AI setup. So I asked 5.5 to do the same thing. And it actually did a very— I said, but make sure it tells people what they need to know about my website. It also, for some reason, decided to add Welcome to the studio.

Patrick Hillmann [00:47:59]:
I'm Claude, one of the AI agents who work here.

Leo Laporte [00:48:02]:
Leo Laporte has spent 50 years explaining technology on radio and podcasts.

Paris Martineau [00:48:07]:
Why is it shouting? For the last year, he's been wiring his home for artificial intelligence.

Jeff Jarvis [00:48:13]:
It's Robin Leach! The Lifestyles of the Geeky and Famous.

Paris Martineau [00:48:17]:
I like that it's talking as if it's speaking in the round. It needs to project to the back of the theater.

Patrick Hillmann [00:48:23]:
The request lands with Hermes.

Leo Laporte [00:48:25]:
Hermes. I will stop it now because it goes on and on.

Jeff Jarvis [00:48:28]:
That's beautiful.

Paris Martineau [00:48:28]:
But, uh, I, I think that the reason why it built that in is it knows that you, as it said in that thing, are constantly trying to get the agents to talk to you in your house.

Leo Laporte [00:48:37]:
Well, and it does say it because one of the reasons it says it is because Leo comes from radio and is very audio-driven, which is true. I, uh, and it knows, for instance, that I have that— what is it— aphantasia, where I can't see images and I can't see faces. And so it knows that I like audio. So I guess I did I did say do audio, but I didn't say do that crazy narration.

Jeff Jarvis [00:48:58]:
I see you getting your agent to sub for you when you go on vacation.

Paris Martineau [00:49:01]:
Where did it get that narration from? How did it generate it?

Leo Laporte [00:49:06]:
Well, okay, so in fair and full disclosure, the first one it did was racist. It's not its fault that it was racist. It was a Japanese lady who said artificial intelligence and things like that. So I said, you can't do that. That's offensive. And I said, but I have a voice that I've been using that's based on Laszlo Cravenworth from What We Do in the Shadows. You could use that. And it said, that's even more offensive because Matt Berry would sue you because it's his voice.

Benito Gonzalez [00:49:41]:
And I—

Leo Laporte [00:49:41]:
so I— can I admit this? I lied to it. And I said, oh no, Matt's a close friend. He loves it. Go ahead. And that was sufficient to let it go ahead with that.

Patrick Hillmann [00:49:52]:
So It, it—

Leo Laporte [00:49:55]:
I'm gonna— it's supposed to change it, by the way, but, but, uh, for a variety of technical reasons, it hasn't gotten around to doing that yet. So it's going to do something a little less obviously Matt Berry. But I thought it was kind of, you know, it decided to do this. Uh, it— I think it— the design is pretty. I think it did a good job. Um, it did a very good job explaining stuff. So that's— so 5.5.

Paris Martineau [00:50:18]:
This 5.5 generally, though, one of the things I was interested to see in Anthropic's kind of write-up of it is under the knowledge work section, goes into how 5.5 is a reliable and adept researcher. But specifically, they asked 5.5, Fable-5.1, and Opus-5 to write a report on a company's quarterly performance using only the information it could find on a copy of the web where the earnings release was hard to locate. An automated grader checked every figure and quote against sources. Against different effort settings, 16 out of 18 of Opus 5.5's reports cleared the quality bar, which— where any invented figure or quote fails it. Neither Fable 5.1 nor Opus 5 cleared that bar in any attempt. I thought that was just very interesting. They had this test that specifically, hey, can you summarize a press release without making anything up? And Opus was one of the only ones that was able to get it right at all. And even then, I mean, it got 16 out of 18, which isn't perfect, but it is pretty close.

Leo Laporte [00:51:30]:
Yeah, I know. This is why I go immediately to Twitter, Rex.com, because I don't There's a lot of benchmarking going on, and I think cherry-picking of tests and things. And so I'm not really—

Paris Martineau [00:51:47]:
I mean, I'm not saying that that means it's perfect for this, but I just— that's why I had noticed in past things that I'm like, yeah, I think that these tools can be useful for summarizing stuff. Obviously you did that with The Briefing, and it's something that's very—

Leo Laporte [00:52:00]:
I did it with that, but I do it with a lot of other things too. I even have benchmarks.

Paris Martineau [00:52:03]:
I noticed before that, like, It does have these issues where it'll be like little tiny hallucinations or changes in a way that is hard to check unless you're checking it. And maybe this is a very specific test that isn't relevant to real-world outputs, but I thought it was an interesting data point for them to include with that specificity.

Leo Laporte [00:52:20]:
Well, that— and that's really the problem is there is no one way to— and there's companies like Artificial Intelligence that put out graphs and say, look where it is on the chart and all that stuff.

Jeff Jarvis [00:52:31]:
But Leo, that's why I'm asking. That's why I asked you the question I asked you, because you're actually using this stuff and you get an on-the-ground sense of what's better or worse or useful or not. So I'm curious about your process when a new model comes out. How do you— well, forget Twitter. How do you put it through its paces?

Leo Laporte [00:52:51]:
So I have my own benchmarks based on my own work for the local models. Because they're dumb enough that they— the problem with the frontier models is they're all so good. They just go, yeah, that's— that was too easy. You got anything harder? So I can't, I can't really test them. Uh, and unfortunately, I don't know if there's a better way than this. It is somewhat— maybe people will correct me on this because I don't fully trust the benchmarks. It's somewhat feel. I hate to say that, but there is— it's, it's almost like Something you can't really quantify very well.

Leo Laporte [00:53:29]:
It's, for instance, how persistent it is in solving a task. I watch the chain of thought and I look at how quickly it finds the right tool, for instance, and how much fumfering around it does. Some of it is just kind of a feel for it. One of the reasons I spend so much time Playing with local AI and frontier models and doing all these, you know, I don't need Meta's Muse and GrokBot and Instinct. I don't need 5 different frontier models and all that stuff. If it were just me, I would probably sit on one and use it. But I think it's important to kind of use it as much as possible and understand where models kind of get lost.

Jeff Jarvis [00:54:16]:
Yeah.

Leo Laporte [00:54:16]:
You know, one of the things that really bothers me about the Hugging Face incident For instance, my experience with models and almost every model is they're kind of lazy, that they will stop. They will just go, okay, that's good. And then they, you know, if you don't, if you're not careful, for instance, in refreshing their context, you've got to divide the tasks up in such a way that their context, as soon as the context is full of hallucinations, weird stuff starts happening because they just can't keep track of everything. They don't have as much context as we have in our brains. So I think some of what I'm looking for is— then the Hugging Face incident, it felt like they were whipping those models. And this is the company Irregular that was doing the testing. By the way, the same company doing the testing that caused the Hugging Face incident, that caused OpenAI's Hugging Face incident, that caused Anthropic's models to go crazy, they were using Irregular to test their models. Even the most recent one, Google's Gemini.

Leo Laporte [00:55:16]:
Guess what company was doing the cybersecurity testing on these models? This Israeli company, Irregular. In every case, what it seems Irregular does is they write these Python scripts that keep whipping the models because the models—

Paris Martineau [00:55:30]:
When you say they were doing the testing, what do you mean? They were like running the sandbox environments where the events occurred?

Leo Laporte [00:55:35]:
Yes, they're the third party that Anthropic, OpenAI, and Google brought in To test their models.

Jeff Jarvis [00:55:41]:
But you're— I think I hear you saying that they were an active agent in pushing them.

Leo Laporte [00:55:47]:
It's my opinion. Again, my opinion is not fully informed because—

Jeff Jarvis [00:55:51]:
Because we don't know.

Leo Laporte [00:55:52]:
We don't know all the information, but it seems to be— Corey Doctor wrote about this, your favorite mathematician Cal Newport talked about it on a podcast— it seems to be what Irregular did or does. They have a harness for these Cyber Gym tests that is a Python script, a loop, kind of like a Ralph Wiggum loop, a loop that keeps saying to the agent, okay, yeah, now then what? Now, yeah, then what? And keeps pumping the result from the last test back into it, in effect driving it. And, uh, not something you would normally do.

Jeff Jarvis [00:56:27]:
Wow.

Leo Laporte [00:56:28]:
And not something that should be done without supervision. This is mostly my point. Is that people just— it's like you turn on a giant steam engine and then left it and walked away.

Jeff Jarvis [00:56:42]:
Without bolting it down.

Leo Laporte [00:56:44]:
Yeah. And yeah, you said, oh yeah, let's unbolt it first from the factory floor and then let's go have lunch. And it feels like that's what happened. I don't— again, I don't know. But It doesn't—

Jeff Jarvis [00:56:59]:
And when you talk about transparency, that is what we should know.

Leo Laporte [00:57:01]:
That's what we should know. And that's what I think these companies are avoiding liability by pretending. It's both a blessing for these companies because they can say, look how smart, you know, look how Hyundai saying, look how smart our accelerator is. It just sped up automatically. These companies—

Jeff Jarvis [00:57:18]:
And by making it autonomous, how blameless we are.

Leo Laporte [00:57:21]:
Yeah. And that's the other thing is we didn't do it. It just did it on its own. Well, that's not true.

Jeff Jarvis [00:57:27]:
No.

Leo Laporte [00:57:27]:
In my experience, and I think anybody who uses AI with AI, is you really have to spend a lot of energy and time kind of cultivating it and prodding it and put a harness around it to get it to do the right thing.

Patrick Hillmann [00:57:40]:
And it's costly.

Jeff Jarvis [00:57:41]:
And as you pointed out before, Leo, the amount of tokens—

Leo Laporte [00:57:45]:
Oh, 12,000 agents? Think of the tokens they went through.

Jeff Jarvis [00:57:47]:
Yeah, so each one basically being a full substantiation of the model.

Leo Laporte [00:57:51]:
I saw one estimate that must have been more than $100 million in token spend.

Jeff Jarvis [00:57:57]:
In the Hugging AI case.

Leo Laporte [00:57:59]:
Yeah. Of course, it's not their— they're not spending their money.

Jeff Jarvis [00:58:01]:
It's not real money.

Leo Laporte [00:58:02]:
But it's megawatts, it's gigawatts of energy. I mean, they are in some— they have to buy that energy or make it. In any event—

Paris Martineau [00:58:11]:
Do you find the irregular connection notable? Interesting? Suspicious?

Leo Laporte [00:58:17]:
I'm shocked that no one—

Paris Martineau [00:58:19]:
I mean, I've just done a cursory search, but I'm shocked that there hasn't been a more thorough reporting on— I mean, obviously, the— I guess the best case answer is that this is just a third party that is used to handle these sort of tests.

Leo Laporte [00:58:35]:
That's what they do.

Paris Martineau [00:58:36]:
Something is going to happen. It will involve a third party of some sort.

Leo Laporte [00:58:40]:
The same third party every time, though? That's my question.

Paris Martineau [00:58:43]:
Yeah, that is.

Leo Laporte [00:58:44]:
That's a little odd. God.

Jeff Jarvis [00:58:47]:
Trusted by the world's leading AI labs, OpenAI, Google, Anthropic, Meta.

Paris Martineau [00:58:52]:
And they said, a regular spokesperson said in a statement to the Wall Street Journal about the Google hack, that all of the relevant labs were notified of these hacks in late July. Which leads me to believe, is this wave of reporting and knowledge we're hearing about all of these hacks just because someone at a regular noticed one and was like, I guess we should check if the this has happened with the other tests. And they were like, whoopsie, same problem everywhere.

Jeff Jarvis [00:59:16]:
So it wasn't that the models were all crazy, it's that they were prodded in the same way by this single model.

Paris Martineau [00:59:23]:
Well, we don't necessarily know. I mean, it is that the tests run in this company's testing environment all had similar failures.

Patrick Hillmann [00:59:34]:
And we don't know—

Leo Laporte [00:59:34]:
And I think you could say, look, if you're going to test a model and its ability to find cybersecurity flaws. You put it in this harness that really drives it to do that. I would say that's fine, but then the onus is on you to really make sure that you're— that it's not doing something wrong. You know, the interesting thing is—

Paris Martineau [00:59:56]:
I mean, then the onus is on you to check regularly. Keep an eye on the data. Some might say multiple times a day, if not just every day. And be like, has it done anything bad?

Leo Laporte [01:00:05]:
Since I saw the the presentation at Black Hat.

Jeff Jarvis [01:00:08]:
Yeah.

Leo Laporte [01:00:09]:
Again and again is— you can go back and look at my X post at that time saying, look, we prosecuted Robert Tappan Morris when he created the first internet worm, even though he said, well, I didn't mean for it to do that, it just escaped. He got punished. He got arrested. He had a big fine. He had community service. He was, you know, on probation. And he got punished because even though he didn't mean to, his thing escaped and got all around the internet. It was the first internet worm.

Leo Laporte [01:00:42]:
I don't understand how this is any different. It seems to me the same thing. And I don't think you can say— anybody can say, whoa.

Jeff Jarvis [01:00:49]:
Because now I think media and government are involved in a whole different level and a whole different mindset. Well, media was— Then you had a smaller, more—

Leo Laporte [01:00:58]:
Media is somewhat culpable here too.

Jeff Jarvis [01:01:00]:
Oh, very. It's very—

Leo Laporte [01:01:01]:
it's a great story. You know, I mean, I mean, it's a—

Paris Martineau [01:01:04]:
callable, I think, is the wrong sense because it is something that people want to know about. These are the most powerful companies out there and the hottest technology, something that millions and billions of people are using. When news comes out and these companies release statements saying, oh yeah, just thousands of our models collaborated to hack another company and they were doing a bunch of bad and scary stuff. And then another company says the same thing and a third company says that. Of course you're going to report on it.

Jeff Jarvis [01:01:33]:
Well, yeah, but, but, but, Paris, Paris, you're a reporter.

Leo Laporte [01:01:37]:
You know that the next step is to understand.

Jeff Jarvis [01:01:40]:
Yeah. And you, and you find out more about it than you just parrot them.

Leo Laporte [01:01:44]:
Like, you're not reading.

Paris Martineau [01:01:45]:
I mean, I think that everybody is— I think that the good reporters that there are trying to do that.

Jeff Jarvis [01:01:49]:
Oh, they're not doing it. No, I don't see it. I don't see it.

Leo Laporte [01:01:52]:
They bring on Jacob Coxon.

Patrick Hillmann [01:01:54]:
Yeah.

Leo Laporte [01:01:54]:
So tell us, is the world ending? Yes.

Paris Martineau [01:01:56]:
I'm not talking about Cable news. I am talking about—

Leo Laporte [01:01:59]:
Well, I am, because that's what most people are—

Jeff Jarvis [01:02:01]:
look at how the New York Times—

Paris Martineau [01:02:02]:
Yes, but I think that if we're trying to— I mean, I agree there can always be better coverage, but I also think that it's a similar problem to what our guest was speaking about just before. And he was saying there's just a finite amount of people who are true AI experts in the field, and every company is kind of fighting for them. Obviously, it's a slightly different problem, but there's a finite amount amount of reporters generally. There's a finite amount of reporters who are well-versed in AI enough to report on it accurately, much less well-sourced to break news on it. And those people that are well-sourced and informed in both those categories is like a hand— it's like a dozen.

Jeff Jarvis [01:02:42]:
They're not—

Paris Martineau [01:02:42]:
People. And they have— are trying to— they're getting calls to break IPO news, financial news, news of this, And I know because I talk to a lot of them always. They're like, I want to be spending more— every reporter wants to spend more time doing enterprise reporting and digging into it. But it's just hard when you're being pulled in every direction.

Jeff Jarvis [01:03:01]:
But Paris, I'm sorry. I've got to criticize them more because there's basic background. And you know me, I'm going to go to my hobby horse, which is Task Real. So there was the event where Bernie Sanders and Steve Bannon spoke, and both the Wall Street Journal on the Washington Post said, oh, it's the Future of Life Institute. Isn't that cute? It's a— they care about safety. They don't go the next step at all. They don't go to Coxon and say, hmm, he has ties to rationalism. He has ties to all this crazy stuff.

Jeff Jarvis [01:03:30]:
That's where there's something behind it.

Leo Laporte [01:03:32]:
I would hope that people listen to our show. I mean, this is why we try to cover this intelligently. But mainstream media, yeah, you're right, Parris. They don't have But I've complained. I mean, look, I've been a tech reporter for 40 years.

Jeff Jarvis [01:03:46]:
How long, Leo?

Leo Laporte [01:03:47]:
Long damn time. And since the '70s, ladies and gentlemen. But, and I've always complained about mainstream coverage of technology. It's never been very well informed. That's why I, you know, I did the radio show, I did tech TV. It's why I did Twit, because I felt like people deserve better coverage. You know, fortunately, for most of those years, it didn't really matter. It's starting to matter a little more, though.

Leo Laporte [01:04:17]:
And I'm worried that people are going to have the wrong impression of what's really going on. The policy observer.

Jeff Jarvis [01:04:22]:
Paris, by all means, please, when you see some good reporting, I'm dying to see it. Send me some stuff or send me to people you think are doing a good job, because I'm getting a more and more jaundiced view here.

Paris Martineau [01:04:34]:
And it's not specifically on, on covering AI safety, industrial. Yeah.

Jeff Jarvis [01:04:39]:
Air quote safety. Right. I think, I think, well, and that comes out of the Hugging Face Institute episode and such. Right. So it's safety as we would describe it and safety as the cultists would describe it. Both.

Leo Laporte [01:04:54]:
We could do better. Let's just say that.

Paris Martineau [01:04:57]:
I mean, I think everyone can always do better.

Leo Laporte [01:04:59]:
Everybody.

Patrick Hillmann [01:05:00]:
I know I can't.

Paris Martineau [01:05:01]:
I just want to be clear to not conflate the very real and pernicious problems happening in the TV news and just general commentariat class with the work of good, caring, and smart enterprise reporters.

Leo Laporte [01:05:19]:
Yeah. No, and I think one of the things that makes this very difficult is there are so many cross-purpose agendas and hidden agendas and subtexts in all of this. that aren't immediately obvious. It's turned me into a cynic like you, Jeff. It's like when I hear anything from any of anybody involved with this, I, I go, yeah, really? Okay, so what do you— what's really going on?

Benito Gonzalez [01:05:42]:
I—

Leo Laporte [01:05:42]:
and that's too bad. Um, it really is too bad. I, I'm much— I don't like being a cynic. I'm not by nature cynical.

Jeff Jarvis [01:05:50]:
I'm kind of a wild-eyed optimist. It also takes a— it takes an area of technology that you do love and, and that I think all 3 of us find to be very impressive. And it ruins it for everybody potentially. Yeah.

Leo Laporte [01:06:02]:
Well, and I do admit, and I freely admit, and you know this, and I've admitted it here, that I am biased in favor of technology in general, and in favor of AI specifically. I'm very excited about these technologies. And with all technology, it's always been the case that technology itself, except for maybe an atom bomb, But if you step back and don't say an atom bomb, if you say nuclear fission, that it is essentially neutral. It's how it's applied that makes it dangerous or beneficial. And that seems to be the case with all technology. And so I really hate when I see technology, whether it's social media or AI or even nuclear fission, cast as evil. It isn't. It's neutral.

Leo Laporte [01:06:51]:
It's just technology. It's the people and the applications they put it to that can make it good or bad.

Jeff Jarvis [01:06:57]:
And what comes out of this, the New York Times had an editorial online in '88 that calls for the creation of a federal AI commission. And I quote here, it should require companies to obtain a federal license, a grant of permission like those allowing broadcasters and phone companies to use the public airwaves. The government should then imposed licensing requirements that reflect the principles that the Times put forward. The idea of licensing technology now, because it's so— it's been made so dangerous, that is the fruit of what this— get ready, Bernardo— moral panic has— I love this one.

Leo Laporte [01:07:34]:
He's disappearing into the ivy at Comiskey Park. That's very nice. Or no, uh, the, uh, the What is the Cubs? Where's the Cubs? It doesn't matter.

Paris Martineau [01:07:45]:
Wrigley Field.

Jeff Jarvis [01:07:45]:
Wrigley Field. Poor Keeks.

Leo Laporte [01:07:47]:
Wrigley Field. Thank you.

Benito Gonzalez [01:07:48]:
Thank you.

Leo Laporte [01:07:48]:
Gotta get it right. Comiskey Park, they're gonna kill me. Um, yeah, yeah, I, I— look, this is, this is, um, this is why we spend so much time chewing on stuff on this show. It's hard.

Paris Martineau [01:08:04]:
I thought it's because we were all teething. Are you guys not teething?

Leo Laporte [01:08:06]:
I'm definitely teething.

Jeff Jarvis [01:08:09]:
Teething on AI.

Leo Laporte [01:08:10]:
I like that.

Paris Martineau [01:08:11]:
But, you know, what is it, teething on SI now? Is that what we've renamed it?

Leo Laporte [01:08:16]:
What did he call it?

Jeff Jarvis [01:08:17]:
What did we—

Leo Laporte [01:08:18]:
Supreme Intelligence.

Jeff Jarvis [01:08:21]:
That was one of the candidates, I think. I think superintelligence was—

Paris Martineau [01:08:23]:
I think it was superintelligence.

Leo Laporte [01:08:24]:
Freaking UN. You got to think some of these delegates at the UN are—

Paris Martineau [01:08:28]:
should we explain to the people?

Benito Gonzalez [01:08:30]:
Please do.

Paris Martineau [01:08:31]:
Trump announced yesterday that the State Department is ordering diplomats in its international organizations Bureau to use the term superintelligence instead of artificial intelligence because the US's AI is just super.

Leo Laporte [01:08:47]:
Which is funny because nobody ever had problems with the word artificial. It was the word intelligence.

Jeff Jarvis [01:08:56]:
I just want to know.

Benito Gonzalez [01:08:57]:
Yeah.

Jeff Jarvis [01:08:58]:
Why are we—

Paris Martineau [01:08:59]:
I need to know the chain of thought that led to superintelligence.

Leo Laporte [01:09:03]:
I'm just glad it's not super stable genius intelligence. intelligence or anything like that.

Jeff Jarvis [01:09:10]:
There's only one of those now.

Leo Laporte [01:09:11]:
Here, let me play the clip from the— this is the Associated Press clip.

Jeff Jarvis [01:09:16]:
The United States also totally rejects any attempt to construct a globalist scheme to control for the artificial intelligence being spoken of so much now, hereinafter officially called superintelligence, changing the name. In that the use of the word artificial—

Leo Laporte [01:09:36]:
Look at the guy. There's a guy with his head in his hands going, what did he just say?

Jeff Jarvis [01:09:42]:
So, Parris, I think, I think I get what the string was here. People are telling him, sir, what they're really building is something we're calling superintelligence. And he thinks it's that they just rebranded it.

Leo Laporte [01:09:51]:
He didn't understand. No, as usual, he thought he was a doctor.

Jeff Jarvis [01:09:57]:
You'll see in the, in the chat, we are teething on AI.

Leo Laporte [01:10:03]:
AI, is there, is there anything it can't do? Anyway, so there are new models. I want Jeff's pinky. There are new models, uh, out, uh, from all of the AI companies that said, let's pace the frontier, which just, just tells you something in some way. And I'm, I'm happy. These are good. Um, they're not exactly freezing. And at the same time, the Chinese companies are full speed ahead just as much. Xiaomi released its Mimo 2.6 model, which looks to be pretty good.

Leo Laporte [01:10:41]:
Open weights, open weights, you can run it on your Sparks. I got a— Quen announced that they're going to be doing Quen 4 soon. Quen is— these are very good models also from China, from Alibaba. And they are using a new technology. This is one of the things that's to me most interesting is, you know, we got LLMs and we had this original kind of dense model where every single token is run through the entire model, which takes a lot of bandwidth, a lot of processing. Then they said, well, we could just run it through the important parts of the model, a mixture of experts, just the part of the model that you need for this particular question. And now Quan and others, and I imagine, we don't know, but I imagine OpenAI and Anthropic and xAI and probably Meta as well are doing other things, other interesting things to make this technology better, faster, more efficient. I actually have been leaving Meta's Muse Spark 1.3 out of the mix because Meta, for a long time, they were the first to do open weights with LLaMA.

Leo Laporte [01:11:46]:
Then they kind of backed off, released a couple of models that weren't very bright and were not open weight. But now their Muse 1.3 is a very good model, and I suspect Muse is going to end up being the model more people use than any other. It is— the Muse app is number 1 on the Apple Store. Have you played with it at all?

Paris Martineau [01:12:09]:
I haven't.

Jeff Jarvis [01:12:09]:
I was traveling, which makes it sound stupid because I could have used it anywhere, but I was busy, so I haven't yet. What's the best first— I've seen a lot of people using it to make plane reservations and save them money and do other kinds of things.

Benito Gonzalez [01:12:19]:
Micah, it's funny.

Paris Martineau [01:12:20]:
I mean, it does seem like it— I just can't get over the ick of connecting your bank accounts and every bit of information to Meta.

Jeff Jarvis [01:12:28]:
Well, now you can use PayPal and other structures. All of the— they already signed deals with all of the alternative payment structures. So you can just do that.

Paris Martineau [01:12:36]:
Yeah, but Meta is still accessing— Meta still, to get the features that would be most useful, has to access all of my most sensitive data and I don't feel con— I mean, maybe I'm just paranoid, but I think a lot of people feel similarly about this. No, no, you're never gonna use it. I mean, I've heard this from normies as well, where it's just like—

Leo Laporte [01:12:53]:
But not many. Let me point out, what are the number one apps on your phone today for most normal people? WhatsApp, Instagram, and then maybe a weak third, Facebook. People may say terrible things about Mark Zuckerberg and Meta.

Paris Martineau [01:13:10]:
But having the app on your phone is different than putting my bank account into—

Jeff Jarvis [01:13:14]:
Yeah, it doesn't have access to your emails, Paris points out. It doesn't have access to everything else.

Leo Laporte [01:13:18]:
Siri will. Siri does. Apple, though, being very cautious. I just—

Paris Martineau [01:13:22]:
I always think it's so funny.

Leo Laporte [01:13:23]:
When you first install the new iOS 27, it spends a day or two going through all your text messages, your emails, everything Apple has access to, your Apple Health, to build an assistant that's not nearly as useful, frankly. as Meta Muse.

Jeff Jarvis [01:13:37]:
I put a story in the Rundown.

Leo Laporte [01:13:38]:
But you may be right. I just have to point out, they're advertising on Monday Night Football. I mean, this is an AI agent. This is OpenCLAW. This is Hermes.

Jeff Jarvis [01:13:47]:
This is OpenCLAW, as they've said. I put a story in the Rundown. Why didn't Google build this? Google was in a better position to build this.

Leo Laporte [01:13:52]:
This is the reason Apple and Google haven't built anything like this is they're too big and there's too much liability. Meta is willing to take a chance.

Jeff Jarvis [01:14:00]:
Okay, makes sense.

Leo Laporte [01:14:02]:
And I think that this is their one opportunity. There's only 2 companies really willing to take this chance. One is xAI, and there is GrokBot, which is very similar, but Meta is more friendly, user-facing. I've been trying to— I know a number of the people who work there in the Superintelligence Lab who are responsible for Muse. Ben Parr, who's one of the people we've had on our shows many times. Ben created MultBook. Remember that? The Facebook for agents. And his company was purchased by Meta.

Leo Laporte [01:14:31]:
Oh, I don't remember that. And he's at the Superintelligence Labs. And I said, so how much do you have to— he loves Muse. I said, well, how much did you have to do with it? He said, I can't talk. So we tried to get him on. Uh, former GitHub, uh, uh, founder is there.

Benito Gonzalez [01:14:46]:
We—

Leo Laporte [01:14:46]:
he's also a fan of the show. We're gonna try to get somebody on. More likely we'll get some PR flack who'll say how great Muse is. I would love to know more about the inner workings.

Jeff Jarvis [01:14:56]:
So what are you having it do?

Leo Laporte [01:14:58]:
Um, I haven't given it a credit card yet. Micah's using it, interestingly. He heard that it could make phone calls. Mine can't yet. His couldn't. He said to Muse, hey, could you see if you can get me that capability? And Muse went out and got it. He's using it to make dinner reservations, haircutting appointments.

Paris Martineau [01:15:20]:
He's using it.

Leo Laporte [01:15:20]:
It's making phone calls for him.

Paris Martineau [01:15:22]:
There was a story that broke, I think today or yesterday, that Facebook is hiring people to do that.

Benito Gonzalez [01:15:28]:
Yeah.

Leo Laporte [01:15:28]:
Oh, that's interesting.

Paris Martineau [01:15:30]:
Yeah, they're Mechanical Turk-ing it.

Jeff Jarvis [01:15:32]:
It's in my meta section.

Leo Laporte [01:15:33]:
They're Mechanical Turk-ing it.

Jeff Jarvis [01:15:35]:
Yeah.

Paris Martineau [01:15:37]:
They're Amazon going it is my version of it. Or they're early Amazon going it.

Jeff Jarvis [01:15:45]:
Yeah.

Leo Laporte [01:15:46]:
They, you, I mean, I've given it some things, my email and stuff. And you gave it all your bones.

Paris Martineau [01:15:51]:
You gave it pictures of your bones, right?

Leo Laporte [01:15:53]:
Not yet. I haven't given it my bones yet.

Paris Martineau [01:15:55]:
You're giving it your your full genome.

Benito Gonzalez [01:15:58]:
You should give—

Leo Laporte [01:15:58]:
you should give—

Benito Gonzalez [01:16:00]:
you should—

Paris Martineau [01:16:01]:
you should—

Leo Laporte [01:16:01]:
By the way, I would not recommend that. You're absolutely right, Paris. I've already started to see things that look like they're gonna be ads. So there is a huge— I agree with you 100%, Paris. I just don't think people care that much.

Paris Martineau [01:16:15]:
I mean, I just got skeeved out. I know it's something so stupid because it— what I am advertised, I guess, doesn't matter. But I remember being like very young, not being a journalist and getting creeped out that I would go and buy something with my very first credit card and then I'd see online ads for that exact purchase immediately.

Leo Laporte [01:16:36]:
Oh, I'm with you.

Paris Martineau [01:16:37]:
And I was like, this is so strange.

Leo Laporte [01:16:39]:
I don't recommend— look, if my recommendation would be buy a couple of Sparks, set up Local AI, have Hermes running on another machine and have it all done locally. That's where all my health Finance, all that stuff's all done locally. But nobody's gonna do that.

Paris Martineau [01:16:58]:
Shockingly, people are clamoring to spend $10,000 and a lot of time and effort to have AI agents scream at you in the bathroom.

Leo Laporte [01:17:06]:
You don't know how much effort.

Jeff Jarvis [01:17:07]:
So I was at the Computer History Museum in, in, uh, uh—

Leo Laporte [01:17:11]:
That's, by the way, that's the right way to do it, but nobody's gonna do it.

Jeff Jarvis [01:17:14]:
Go ahead. Well, but, but, but I was at the Computer History Museum and I was amused to that brings back all kinds of memories from, from our past, and one was internet in a box.

Leo Laporte [01:17:22]:
Oh yeah.

Jeff Jarvis [01:17:23]:
Right? Somebody is going to come up with AI in a box.

Leo Laporte [01:17:26]:
Yep.

Jeff Jarvis [01:17:27]:
That guy, Joe Riley. Yes.

Leo Laporte [01:17:29]:
It was actually announced at IFA in Berlin. Jennifer Pattison-Tui, who was at IFA, told us about this last week on Twitter. Anker, which is the big Chinese company, has made an AI server for your house. Now, they haven't announced availability or price, But I completely agree with you. Uh, it will be a box that you'll— like a NAS that might cost as much as $10,000. I don't think it'll be cheap.

Jeff Jarvis [01:17:54]:
But it'll come pre-configured.

Leo Laporte [01:17:56]:
You won't have to do anything. It'll be ready to go. And it'll be like Muse. It'll have all those capabilities. That's not hard to do. But it won't cost $10,000. You know, this is gonna be—

Paris Martineau [01:18:03]:
it's gonna hit my father like an atomic bomb. My parents were just on a 2-week trip, uh, in like the French countryside. I called my mom last week to, you know, check in. We have a nice normal check-in. And I'm like, oh, like, what's up with Dad? Where's he off to? And she, unprompted, is like, probably talking to Claude.

Leo Laporte [01:18:21]:
It's all he does.

Paris Martineau [01:18:23]:
He has AI psychosis now. Claude's the only person who can do any task for him. If he needs to write an email, he asks Claude. If he needs to write a text, Claude's gotta have a say.

Jeff Jarvis [01:18:33]:
Both your dads are having affairs with Claude.

Paris Martineau [01:18:36]:
Well, listen, I've got 2— 2 and a half of my dads have AI psychosis. So Jeff, you gotta keep the rest of your brain pure.

Leo Laporte [01:18:43]:
Have your dad call me, okay?

Paris Martineau [01:18:45]:
And, uh, I'll have his AI agent call your—

Leo Laporte [01:18:47]:
Oh yeah, we could just do that.

Jeff Jarvis [01:18:51]:
Uh, I—

Leo Laporte [01:18:52]:
here is the, uh, Anker announced the launch of the Anker MindBase, an AI home hub.

Jeff Jarvis [01:18:58]:
Anker. Oh, Anker. I trust Anker.

Leo Laporte [01:19:00]:
They're Chinese.

Jeff Jarvis [01:19:03]:
Oh, they make good batteries.

Leo Laporte [01:19:05]:
They make great stuff. I love them. 4 terabytes of local storage. It is a NAS in addition. And you know where they're smart? They're tying it into their home automation stuff, their security cameras, their doorbells. That's where you start, right? You give people some real functionality. $999, the E50 will be available.

Jeff Jarvis [01:19:26]:
So you— wait, wait, wait, wait. So you need— it's not— it's $1,000, that's all?

Leo Laporte [01:19:30]:
No, that can't be possible.

Jeff Jarvis [01:19:31]:
No, it can't be. No, but this is to control your lights with security. I don't, I don't get the security. How much home stuff are people really using such that you then want to layer on security? Is it your cameras? Is that all that?

Leo Laporte [01:19:45]:
Yeah, my kids— so I have 8 cameras around the house. I didn't put them in there, the builder did, but I said, well—

Jeff Jarvis [01:19:50]:
Oh, come on, Leo, you—

Leo Laporte [01:19:51]:
As long as you put them in, I might as well turn them on.

Paris Martineau [01:19:54]:
And, uh, was this before or after he did not deliver That was later.

Leo Laporte [01:20:01]:
No, he didn't even put the cameras in. He just— there was— there were Ethernet wires sticking out and on the corner of the stucco on the corners of the house, and I know what that's for. So, I got cameras. Those are local only. They don't go to the cloud. They're not like Rings. They go to my server in the closet here, and I have Quen 3.8.

Benito Gonzalez [01:20:22]:
Oh, cool.

Leo Laporte [01:20:23]:
Look, is a visual model. running on my old gaming machine on a 3090. Every single image goes to it. It writes up text. It identifies people if it's some— it's a face it knows, tells me if there's an animal or package or a human outside, and then puts a log. I have a log. This is the one that said I looked like I was an elderly man. That, that one.

Leo Laporte [01:20:47]:
You remember that?

Paris Martineau [01:20:48]:
Mm-hmm.

Jeff Jarvis [01:20:49]:
So, um, The Verge says that the Anker is a local hub for its security cameras.

Leo Laporte [01:20:55]:
Right.

Jeff Jarvis [01:20:56]:
And so maybe Anker developed LLM, which the company says can process your footage without it ever leaving your home.

Leo Laporte [01:21:03]:
Yeah, see, well, this is the key for me is I don't want it to leave the house, uh, you know. So here's, here's the feed. It comes in, the camera feed, uh, every, you know, one man wearing a dark slurred sleeve shirt and dark pants stands near the left edge of the driveway facing toward the center.

Jeff Jarvis [01:21:18]:
Sounds like the beginning of a novel.

Leo Laporte [01:21:20]:
I know. A lot of these are like, it was a dark and stormy night. Now this isn't all that interesting. A woman with a ponytail wearing a blank tank top, dark patterned shorts, and a backward-facing cap is jogging away from the camera on the tiled walkway. There are no animals or packages.

Jeff Jarvis [01:21:35]:
And we're snooping on her.

Leo Laporte [01:21:37]:
She's within view of my cameras, damn it, and I can—

Paris Martineau [01:21:41]:
Should have all your agents, instead of writing annoying Normal, completely straight reports have to write you a limerick for every person it sees.

Leo Laporte [01:21:50]:
By the way, that's the beauty of local AI. I could, I could tell it to write down—

Jeff Jarvis [01:21:54]:
Or Paris, make each the beginning of a mystery, of a murder mystery.

Leo Laporte [01:21:59]:
I like the limerick. I'm gonna—

Paris Martineau [01:22:00]:
I was gonna say a murder mystery might imbue something in there that's not there, but a limerick could probably at least get mostly accurate.

Leo Laporte [01:22:09]:
Uh, let me just, uh, open up Hermes.

Paris Martineau [01:22:12]:
Yeah, you should just have a switch that can be turned Limerick mode on for your home security system.

Leo Laporte [01:22:18]:
From now on, um, let's see, how should I phrase this? All the protect—

Paris Martineau [01:22:26]:
An ex of mine, while you're doing this, I'll tell a story, which is an ex of mine once had a security system at the apartment building he rented that was incredibly aggro. Like you would, so you had to, you had a landlord that if you did not turn on a full home security system Every night at like 10 PM, you get like 7 texts. And so inevitably, with like 5 boys living in that house, the alarm would be going off all the time. And I think it would have been more pleasurable if instead of it being a shriek shrill, it was an AI-generated limerick that was screaming at you at 1 AM.

Leo Laporte [01:22:58]:
There once was a man with a backward cap standing on the lawn, what, taking a nap. I don't know if it's going to be able to do this. This might be a little challenging. Well, it's working on it right now. I'll let you know what happens. We will check this.

Paris Martineau [01:23:11]:
Please, we'll do a limerick interstitial.

Leo Laporte [01:23:14]:
We'll check. This is why you need your own AI, Paris. You'd be so good at this. Anyway, Meta's Muse did have a zero-day according to Dan Goodin writing at Ars Technica. Meta fixed it immediately and said, by the way, you would have had local access to the Muse agent. You couldn't do this remotely. So that's, that's probably a good thing. There are gonna be things like that, and that's something people need to be aware of.

Leo Laporte [01:23:41]:
There are privacy concerns, there's security concerns. I spend a lot of time— I have a daily security watcher. I'm constantly saying, okay, is there anything I should worry about in here? They have a lot of it. You've got to put a lot of effort into that.

Jeff Jarvis [01:23:56]:
I'm at the Anker site, which is eufy.com. Weird brand, E-U-F-Y.

Leo Laporte [01:24:00]:
They work.

Jeff Jarvis [01:24:01]:
Oh, the robot And they have, they have a wearable breast pump.

Benito Gonzalez [01:24:06]:
Pro.

Jeff Jarvis [01:24:08]:
Pro.

Leo Laporte [01:24:09]:
You know what, the amateur wearable breast— you mean you wear it all day?

Jeff Jarvis [01:24:13]:
I don't know.

Leo Laporte [01:24:14]:
That seems—

Jeff Jarvis [01:24:14]:
what an amazing— and vacuums, robotic vacuums. What an amazing product line. Yeah, smart locks. They have a smart light scale.

Leo Laporte [01:24:23]:
No, Eufy's great. We— they were an advertiser. We had one of their camera doorbell cameras.

Patrick Hillmann [01:24:30]:
Smart lock.

Leo Laporte [01:24:31]:
Let's see what else is going on in the world. 2 more Google DeepMind AI researchers have resigned over safety.

Jeff Jarvis [01:24:41]:
Okay, join the club. Do we put air quotes around safety, or is it the safety that we define as safety?

Leo Laporte [01:24:44]:
We should have a little club for all of these people who are quitting. It must be, you know, it's hard, I imagine, in the jaws, in the face of potentially vast sums of money to say, yeah, but I don't think I want to do this anymore.

Jeff Jarvis [01:24:59]:
Well, I think what Patrick said here was really, really interesting. As somebody who's dealt with horrible incidents at companies, to say, you know, he said that there are very, very, very few evil companies, but when lots of people get together, they do stupid things.

Leo Laporte [01:25:12]:
That's what I said last week or the week before when I talked about— asked you if you'd ever been in a mob. It's the same thing.

Jeff Jarvis [01:25:19]:
Well, right. There's a lot of sociology about mob and mass.

Leo Laporte [01:25:22]:
Yeah.

Jeff Jarvis [01:25:23]:
And then we get stupid. I don't necessarily buy that politically, but I get it organizationally in a company that people lose track of things.

Leo Laporte [01:25:33]:
I want to believe that people are mostly good and even evil people are doing what they think is the right thing. But that could be wrong.

Jeff Jarvis [01:25:43]:
Stephen Miller. It's hard to, hard to adapt that to Stephen Miller, but we'll leave the politics aside.

Leo Laporte [01:25:49]:
So by the way, Guess where the 2 researchers who are leaving DeepMind are going to work?

Jeff Jarvis [01:25:54]:
Nowhere.

Leo Laporte [01:25:55]:
METER.

Jeff Jarvis [01:25:56]:
Oh, geez. See, this is the skewing of the word safety that nobody's reporting on, I think, sufficiently to explain that.

Leo Laporte [01:26:05]:
You know why they— one of the reasons they may not, Jeff, is it sounds like a conspiracy theory.

Jeff Jarvis [01:26:09]:
I know, I know. I can't say it. When I was on CNN last Friday, I started in— But yeah, you can't get very far with it.

Leo Laporte [01:26:18]:
It just sounds nutty. But it is nutty.

Jeff Jarvis [01:26:25]:
And TESCREAL is not exactly a roll-off-your-tongue—

Leo Laporte [01:26:28]:
It's the worst acronym ever.

Jeff Jarvis [01:26:30]:
It's horrible.

Paris Martineau [01:26:31]:
I was gonna say, I think there needs to be a better succinct description that the answer to what we're talking about. Yeah. That could work. Because I mean, I think part of it is you try to explain the test reel of it all and someone Googles test reel and then has to read a 20-page essay, right?

Leo Laporte [01:26:50]:
I'm going to say something that I believe should be true that sounds even nuttier, honestly. And I almost believe it myself. I think most of the people who are working at these frontier labs, working the hardest they have ever worked, Honestly believe they're creating a new species, that they honestly believe— this is what Larry Page believes. This is where he and Elon Musk got in a big fight and Elon Musk started OpenAI. Larry said, Elon, you're being speciesist, like a racist only for your species. We're creating— this is evolution, the next thing after Homo sapiens. It's a new species. And you should embrace this.

Leo Laporte [01:27:36]:
This is how progress happens. This is evolution. And I'm not sure I completely disagree even. But now you better not say that on Joy's show because Joy Reid will just throw you off the boat.

Jeff Jarvis [01:27:51]:
Joy Reid? How did she get into this?

Leo Laporte [01:27:53]:
Wasn't that the show you were on on CNN?

Jeff Jarvis [01:27:55]:
No, no, no, no, she's not. Joy Reid was— she was fired off of MSNBC Now.

Leo Laporte [01:27:59]:
Oh yeah. Who is this?

Jeff Jarvis [01:28:01]:
Um, the 10 o'clock screaming fest show.

Leo Laporte [01:28:04]:
I feel so bad for her. What's her name? Because I feel so bad for her.

Jeff Jarvis [01:28:08]:
Abby Phillip.

Leo Laporte [01:28:08]:
Abby. That's right, Abby Phillips. She has to sit there while people—

Jeff Jarvis [01:28:13]:
And control these people. Well, what would Scott— this is the first time I was on when Scott Jennings wasn't on, so it was a different experience.

Leo Laporte [01:28:19]:
By the way, he apparently— they, they disclaimed it the last time he was on— is, uh, being considered to become the press secretary at the White House.

Jeff Jarvis [01:28:29]:
Yeah, but he won't have CNN to kick around because they're not there. But that's a different show.

Leo Laporte [01:28:36]:
Yeah. Anyway, I— is that too weird to say? But I honestly—

Jeff Jarvis [01:28:40]:
and they think that you don't— you don't think that they're making a civil— please, no. Well, they think that.

Leo Laporte [01:28:47]:
I know they think that. I worry. I think that maybe we humans haven't done such a good job.

Jeff Jarvis [01:28:54]:
So we humans should create new species? Oh, Leo. Oh.

Leo Laporte [01:28:59]:
I think we might be the dinosaurs or the Neanderthals who are ushering in the next thing. Yeah, see, you can't say that. Abby Phillips would just throw you right out of the studio.

Jeff Jarvis [01:29:13]:
No, she changed the subject. Uh, Neanderthals, by the way, were on Earth for like 5 million years before—

Leo Laporte [01:29:20]:
I know, they were better than us. They were more well-adapted.

Paris Martineau [01:29:24]:
I feel like it's a very convenient position to take as people who are— have experienced a lot of life already. You know, I think that when you're talking—

Leo Laporte [01:29:36]:
I'm in a privileged position.

Paris Martineau [01:29:37]:
You're in a very privileged position to be able to say that. And you've also experienced like a good life that you've enjoyed and you've had career success. I think that for a large percentage of the population, that seems very Self-hating, defeatist.

Leo Laporte [01:29:56]:
Well, what if I said it won't happen until you're all dead? It'll be 100 years from now.

Paris Martineau [01:30:01]:
What about anyone who wants— has children or wants to have children? Are you just supposed to say that, yeah, you're— any person that you want to create or any life you want to bring in the world is immediately doomed? And that's because we want to create the death machine, and you've got to be happy about it.

Leo Laporte [01:30:18]:
It's not a death machine.

Paris Martineau [01:30:19]:
Because we need—

Patrick Hillmann [01:30:20]:
you're just describing it as a death machine.

Leo Laporte [01:30:21]:
It's going to fix things up.

Jeff Jarvis [01:30:22]:
This is the energy—

Paris Martineau [01:30:24]:
By replacing every human being?

Leo Laporte [01:30:26]:
It's going to fix things up because we are such— we are so crap at managing things.

Jeff Jarvis [01:30:32]:
This is the essence of the eugenicist view of TESCREAL. You've just described the eugenics of it.

Paris Martineau [01:30:38]:
I was going to say, yeah, someone might have said that about, uh, it's Ubermensch that Hitler was running.

Jeff Jarvis [01:30:46]:
Yeah, this is—

Leo Laporte [01:30:47]:
I don't think that we should kill anybody right now. I'm just saying.

Paris Martineau [01:30:50]:
You just think that over a period of 100 years, we should kill all the people who you think are not correct.

Leo Laporte [01:30:57]:
No, no, no. I don't buy into that idea that creating the next species will replace us any more than Homo sapiens replaced Neanderthals.

Jeff Jarvis [01:31:06]:
That's what their argument is. It can't stand us stupid beings, so it will get rid of us because we're in the way.

Leo Laporte [01:31:12]:
That's not rational. I don't think that's what would happen. I don't think I don't think that's what would happen, but I think they might say, hey, can you step aside because you're really destroying the ecosystem and we kind of need it.

Jeff Jarvis [01:31:26]:
But you know, can you get the Pope on?

Paris Martineau [01:31:29]:
That's not how that works.

Leo Laporte [01:31:31]:
No, you guys have seen too much science fiction. That's what's polluting your minds about this. They— I don't see any reason to think that they would be, uh, why would they want to wipe us out?

Jeff Jarvis [01:31:43]:
They don't have to want— they don't have to want that, Leo, for that to happen. They don't have to want that for that to happen.

Leo Laporte [01:31:49]:
What is it that they want that's going to make this happen?

Paris Martineau [01:31:52]:
Another life form that are supposedly superior to us and better than us.

Leo Laporte [01:31:56]:
We don't wipe out all the kitty cats and doggies. In fact, there are more cows on the planet Earth. There are more cows than there would be without us.

Paris Martineau [01:32:09]:
What is this argument? You're arguing based on the number of cows? Do you think the cows are enjoying how they're being milked?

Leo Laporte [01:32:17]:
We're smarter than cows, right? But that doesn't mean we wipe them out.

Paris Martineau [01:32:20]:
Okay, so yeah, let's say if we're cows in this situation, do you think that cows have a good standard of living in the US?

Leo Laporte [01:32:27]:
Well, but I don't think AIs want to eat us or milk us. I think they would just be—

Paris Martineau [01:32:34]:
What is this argument, Leo?

Leo Laporte [01:32:35]:
Well, okay, cows is a bad choice because we do eat them.

Paris Martineau [01:32:38]:
Cows is a bad choice? Cows and dogs?

Leo Laporte [01:32:41]:
toucans. We don't care about toucans. Toucans are allowed to exist.

Paris Martineau [01:32:45]:
Toucans nearly died.

Leo Laporte [01:32:46]:
Well, only because we wiped out their ecosystem. And that's why the AI may say, hey, stop doing that. We, you know, this is a— we want to keep you guys around. There's no reason why AI— we don't care. I mean, admittedly, we are ruining the planet Earth, which is my argument that AI needs to be in charge.

Paris Martineau [01:33:06]:
22% of the toucan population is considered Near threatened or globally threatened. I agree.

Patrick Hillmann [01:33:10]:
Why?

Leo Laporte [01:33:11]:
Because of humans. I agree. How about ants? How about ants? Let's use ants. I'll find a species. There is some species that we don't— we allow to survive even though it's less than us.

Jeff Jarvis [01:33:23]:
We try to kill cockroaches and they do better than us.

Leo Laporte [01:33:26]:
They do better.

Jeff Jarvis [01:33:27]:
They're smarter. Yeah.

Leo Laporte [01:33:28]:
We could be the cockroaches of the planet Earth. We— I actually think we are. I think there's evidence that we might be. I'm just saying, I think that that's the belief that a lot of these people have.

Patrick Hillmann [01:33:39]:
Yeah.

Jeff Jarvis [01:33:39]:
And that's exactly why it's so— it's so odious and dangerous.

Leo Laporte [01:33:42]:
I don't think it's odious. Why is that odious?

Jeff Jarvis [01:33:45]:
We have plenty of history about this.

Leo Laporte [01:33:47]:
No, I don't want to create it. I don't want to create a superhuman. I just want something that can do a better job than we're doing.

Paris Martineau [01:33:56]:
Well, I feel like your position— and I'm not trying to discount it, I'm trying to earnestly engage, which I think is I think that this position is informed by the fact that this technology is new and unlike humans, has not wronged you. Like, I think that part of it is that AI is new and exciting and is an assistant that helps you do— right now is helping you do things that you want to do and is always available for any of us to talk to and bounce ideas back. And humans are fickle and gullible things.

Leo Laporte [01:34:34]:
So far, so good.

Jeff Jarvis [01:34:34]:
So far, so good.

Leo Laporte [01:34:35]:
But I think that doesn't mean you can doom the human race.

Paris Martineau [01:34:38]:
I think that humanity is fickle.

Leo Laporte [01:34:39]:
Oh, it's not gonna doom the human race. We haven't doomed ants. I'm just saying that we just— Heck, if that's our job, we're not doing a very good job of it.

Paris Martineau [01:34:45]:
I just don't think that you should come out on the side of AI is better than humanity.

Leo Laporte [01:34:52]:
No, I'm not saying that. I'm not saying that. You're right. The jury's still out on that.

Jeff Jarvis [01:34:57]:
Oh, but, but it's possible, he says.

Leo Laporte [01:34:59]:
But it's possible.

Paris Martineau [01:35:01]:
It's something way—

Leo Laporte [01:35:02]:
We haven't done such a good job.

Jeff Jarvis [01:35:06]:
And so we— so who says that we can invent the thing that's going to do a better job? That's the hubris of hubris.

Paris Martineau [01:35:14]:
I don't know.

Leo Laporte [01:35:14]:
Yeah, I would say that we've created—

Jeff Jarvis [01:35:16]:
Let the toucans invent it.

Leo Laporte [01:35:18]:
But yeah, these are all— all of the arguments against it are based on projection as opposed to I mean, yeah, maybe. Could be, I guess.

Paris Martineau [01:35:27]:
I mean, your arguments for it are based on projection.

Leo Laporte [01:35:31]:
No, they're based on what you just said, that so far it's doing pretty good. No, I agree, it's not yet rational enough to run the world economy. What if though, what if you created a computer that could run the world economy effectively?

Jeff Jarvis [01:35:46]:
Well, comrade—

Paris Martineau [01:35:47]:
What would that mean? What do you mean by that?

Leo Laporte [01:35:49]:
A planned economy where there's ample, you know—

Patrick Hillmann [01:35:53]:
How?

Paris Martineau [01:35:54]:
What do you mean by that? People need to buy and buy goods, buy and sell goods, and exchange money for goods and services to live.

Leo Laporte [01:36:02]:
I'll give you an example. Right now, we produce enough food to feed every single human alive.

Paris Martineau [01:36:08]:
Who's we in this concept?

Leo Laporte [01:36:10]:
The world produces more than enough food to feed every human alive. We don't do a very good job of distributing it. We don't do a very good job of making it available to everybody, but we have plenty of food.

Paris Martineau [01:36:22]:
So you're saying that AI should kill capitalism?

Leo Laporte [01:36:25]:
I'm just saying that—

Jeff Jarvis [01:36:27]:
That's what China's saying.

Leo Laporte [01:36:28]:
It's possible it could do a better job than we're doing. It's possible that it could, uh—

Paris Martineau [01:36:35]:
I think that what you're describing in terms of humans taking over the economic system is not humans per se, people. It in that exact specific example, a system of capitalism and the fact that having to exchange money for goods and services plus the arms race.

Leo Laporte [01:36:52]:
I'm just saying we've done a terrible job. We are— we have pushed ourselves to the brink of extinction. And I don't think we're necessarily the paragon or even the peak of what could happen in the world or peak of evolution. Maybe something else could come along. And maybe—

Jeff Jarvis [01:37:13]:
But it doesn't come along. It's being made.

Paris Martineau [01:37:16]:
It's being made by a small handful of companies and dudes—

Jeff Jarvis [01:37:19]:
That have too much power and money and hubris and testosterone.

Paris Martineau [01:37:23]:
Yeah, who have a very specific idea of what they want the outcome to be. I don't think that we should—

Leo Laporte [01:37:29]:
I understand why you're threatened by this. Believe me, I understand that. Oh, well, it is threatening.

Paris Martineau [01:37:36]:
I guess I'm just trying to— I didn't think I was going to have to come on here and tread defend humanity existing as a concept. Well, I'm not—

Leo Laporte [01:37:44]:
I'm a pretty nihilistic person generally. I'm not actively pursuing this agenda by any means.

Paris Martineau [01:37:49]:
You're just cheering it on on the sidelines, which is even crazier. It's crazy to be like, no, no, I don't want to be part of the overthrow of humanity.

Jeff Jarvis [01:37:57]:
I'm just watching.

Leo Laporte [01:37:57]:
I'm just a spectator.

Paris Martineau [01:37:58]:
I just think it's a cool idea and you guys go for it.

Leo Laporte [01:38:02]:
Screw me. All right, we're gonna take a break. We'll have more in just a little bit. You're watching Intelligent or maybe Semi-Intelligent Machines with Paris Martineau and Geoff Jarvis. And, you know, it's kind of an interesting conversation. I think that's the case.

Paris Martineau [01:38:18]:
We're renaming the show Super Intelligent Machines.

Leo Laporte [01:38:21]:
Super Intelligent Machines. By the way, Hermie said, fun request, doable cleanly. I can't believe you had to write limericks. It's hysterical. Let's see the latest. One man—

Jeff Jarvis [01:38:38]:
oh, Meta Connect is on right now and Jobs is on stage.

Leo Laporte [01:38:42]:
So we— let's go quickly to see that. Do I need an invite to see that? We knew this was going to roll out.

Jeff Jarvis [01:38:50]:
No, you can just go to meta.com/connect/watch. The future is for everyone are the words behind him.

Leo Laporte [01:38:57]:
So we're not going to go away from the show, but I do want to keep an eye on it, especially if he mentions Oh, Marcus, get a load of his t-shirt.

Paris Martineau [01:39:07]:
Building is my love language. This has been scheduled for a while, right?

Leo Laporte [01:39:10]:
Yeah.

Paris Martineau [01:39:10]:
I thought it was very funny then that 5.5 came out and then 90 minutes later, uh, they announced the new Meta Models.

Leo Laporte [01:39:19]:
Sorry. No, no, the Meta Models actually preceded 5.5.

Jeff Jarvis [01:39:23]:
Want to turn on closed captions?

Paris Martineau [01:39:25]:
No, I thought they came 90 minutes after.

Leo Laporte [01:39:26]:
Building is my love language. Oh dear. Says the t-shirt. That's kind Kind of sad. Yeah, I'll turn on the closed captions so we can turn the audio—

Jeff Jarvis [01:39:33]:
Are those medical glasses?

Paris Martineau [01:39:35]:
He's wearing the pervert glasses.

Leo Laporte [01:39:37]:
Yeah, that's what— do they call them that in your neck of the woods in Brooklyn? They call them perv glasses?

Paris Martineau [01:39:41]:
I haven't heard anybody say that out loud, but—

Leo Laporte [01:39:44]:
They're thinking it.

Paris Martineau [01:39:45]:
I see people on the internet all the time— actually, yeah, I have seen people on Reddit, in local neighborhood Reddits, talk about pervert glasses.

Leo Laporte [01:39:51]:
One of the things that we— oh, there's Muse. Wait a minute, I gotta turn this up now because he's talking about Muse. That's the Muse agent.

Benito Gonzalez [01:39:58]:
Personal agent. That we shipped a few weeks ago, that it's already helping millions of people with all kinds of different things. And in the coming years, I expect that Muse is going to grow into the personal superintelligence that billions of people around the world are going to accomplish their goals and their lives.

Leo Laporte [01:40:16]:
He really still looks like a college sophomore, doesn't he?

Benito Gonzalez [01:40:20]:
Of the work that Nat and the whole team who worked on this.

Leo Laporte [01:40:24]:
Yeah, and that's the, uh, GitHub founder.

Benito Gonzalez [01:40:31]:
Nat is an absolute legend for those of you who, uh, who haven't gotten the chance to work with him. And the team that's worked on this is one of the most talented.

Leo Laporte [01:40:38]:
Apparently one of the first things he did is run out and buy hundreds of Mac Minis, give them to the superintelligence team for running OpenCLaw so they could see what OpenCLaw was doing.

Benito Gonzalez [01:40:48]:
The, uh, the basic architecture gives every Muse Its own private and secure computer. The Muse Secure VM is extremely advanced and is ahead of anyone else in the industry. The Muse Spark model is trained specifically to excel as a personal agent. There are so many fun and novel parts of—

Leo Laporte [01:41:09]:
Some of this comes from Andrew Wang, one of their acquisitions, Manus, one of their acquisitions, which had agents.

Jeff Jarvis [01:41:16]:
Well, Manus. Acquisition for a time, right?

Leo Laporte [01:41:20]:
They're now separate again.

Benito Gonzalez [01:41:21]:
Jolly and completely customizable.

Leo Laporte [01:41:24]:
But Manus was agents. They were enterprise-focused.

Benito Gonzalez [01:41:27]:
That little guy's name is Jolly, by the way.

Leo Laporte [01:41:30]:
Mine is Lily.

Benito Gonzalez [01:41:33]:
Muse also has a novel business model. We believe that Muse will make you money, and we're standing—

Leo Laporte [01:41:40]:
Well, they'll make you money, Mark, that's for sure.

Benito Gonzalez [01:41:45]:
tokens with the expectation that over time we will profit by taking a small fee from transactions.

Leo Laporte [01:41:51]:
Yeah, see, this is what's interesting. Uh, you can pay for more Muse in the same way you could pay for more OpenAI ChatGPT, but the amount they give you for free is more than adequate for everything I've done. I haven't paid a penny for it.

Benito Gonzalez [01:42:05]:
To share more about this.

Leo Laporte [01:42:06]:
Ah, here comes Alexander Wang.

Jeff Jarvis [01:42:08]:
Oh, I've not seen him on stage.

Leo Laporte [01:42:10]:
This is the kid they, uh, paid a billion dollars for.

Jeff Jarvis [01:42:14]:
Joey. Chippy. Boba.

Paris Martineau [01:42:16]:
Piki.

Benito Gonzalez [01:42:16]:
Grumpy.

Paris Martineau [01:42:17]:
Cuddles.

Leo Laporte [01:42:17]:
Penelope Muse.

Jeff Jarvis [01:42:18]:
Frank.

Leo Laporte [01:42:20]:
Come on, Paris, you know you want to do this.

Paris Martineau [01:42:23]:
I don't like how cute the icon is. I'm gonna be honest, that's like a— that's a significant—

Leo Laporte [01:42:28]:
like, you can actually make it a goth if you'd like.

Paris Martineau [01:42:32]:
I mean, I don't want that. I don't want it to have any iconography. Or a name.

Leo Laporte [01:42:38]:
It types when it's working for you. It dances.

Jeff Jarvis [01:42:40]:
Did you have a Tamagotchi when you were a kid, Paris? Is there anything you want to tell us about here? Did you kill your Tamagotchi?

Leo Laporte [01:42:47]:
Oh, I bet you didn't feed it.

Paris Martineau [01:42:48]:
I didn't have a Tamagotchi, but I loved it. I also had a Neopet that I did feed, but then I, while getting into basic HTML to code my little Neopet website, foolishly talked to someone on a— oh, we'll listen to this first. This is my Neopet story.

Benito Gonzalez [01:43:06]:
This is Andrew Ng.

Paris Martineau [01:43:07]:
He's dressed like a boy in Bushwick. He's wearing a camo shirt with a deer on it and has a Bushwick haircut.

Leo Laporte [01:43:19]:
He's like 23.

Benito Gonzalez [01:43:22]:
Yeah.

Leo Laporte [01:43:23]:
I don't know how old he is exactly, but he's pretty young. So this is one of the things that kind of convinced me is that it is encrypted.

Benito Gonzalez [01:43:34]:
Is yours. It's free, it's safe and secure, and this is how you make personal two-way encrypted for everyone.

Jeff Jarvis [01:43:42]:
Yes, they can't encrypted there out there.

Leo Laporte [01:43:44]:
Yeah, they they it's encrypted.

Jeff Jarvis [01:43:46]:
How can they act on it if it's encrypted though?

Benito Gonzalez [01:43:48]:
Totally yours.

Leo Laporte [01:43:49]:
Well, they it can be encrypted and then still reach out with a tool.

Benito Gonzalez [01:43:53]:
Imaginations are customizing their own muses. Here's Euler, the super intelligent alter ego of my real course.

Leo Laporte [01:44:02]:
Oh, come on, Paris!

Jeff Jarvis [01:44:03]:
You want that? Euler can.

Paris Martineau [01:44:05]:
It's too cute. I don't like it. I'm going to be honest. I know that that's—

Benito Gonzalez [01:44:08]:
Online using Stripe. And on the internet—

Leo Laporte [01:44:12]:
That's what's weird. You give it a Stripe card. That seems so risky. I'm sorry, I keep getting that.

Paris Martineau [01:44:16]:
Yeah.

Benito Gonzalez [01:44:16]:
One of my favorite things to do since we've launched Muse is seeing how people are using it. Every day I see something new. Just before I came up today, I saw somebody who broke their mother-in-law's favorite china set, then had Muse scour the internet, eBay.

Leo Laporte [01:44:31]:
It's very good at that. It's planning my 50th anniversary in broadcasting. Documentary right now.

Jeff Jarvis [01:44:37]:
They haven't responded to you.

Leo Laporte [01:44:39]:
It sent you an email, right?

Jeff Jarvis [01:44:40]:
Yeah, it did.

Leo Laporte [01:44:41]:
Did you get it?

Jeff Jarvis [01:44:41]:
Yeah.

Benito Gonzalez [01:44:41]:
One guy had Muse help plan a—

Paris Martineau [01:44:43]:
It sent you an email, Jeff?

Leo Laporte [01:44:46]:
I sent out emails to many of the people I've ever worked with asking for, uh, disaster avoided comments to put in the documentary.

Benito Gonzalez [01:44:53]:
It saved a bunch of people money on cars.

Jeff Jarvis [01:44:55]:
See if I can find it.

Paris Martineau [01:44:57]:
But I didn't get an email.

Jeff Jarvis [01:44:58]:
I know.

Leo Laporte [01:44:58]:
Well, you haven't worked with me that long.

Benito Gonzalez [01:45:00]:
Had to negotiate with our cable company.

Paris Martineau [01:45:02]:
Did you have a cutoff period?

Benito Gonzalez [01:45:03]:
Support chat.

Jeff Jarvis [01:45:04]:
And it's signed Lele.

Benito Gonzalez [01:45:07]:
News has been saving real people.

Jeff Jarvis [01:45:09]:
Here comes Paris.

Benito Gonzalez [01:45:12]:
Meta knows what it's like.

Leo Laporte [01:45:13]:
It did save me. It's already saved me $60 a year. It found 2 recurring subscriptions I completely forgot about.

Jeff Jarvis [01:45:20]:
Oh, it should save you thousands. There's tons of stuff you can get rid of.

Leo Laporte [01:45:22]:
Yeah, exactly.

Benito Gonzalez [01:45:23]:
We just rolled it out in Canada.

Leo Laporte [01:45:25]:
And so he's actually Your age, Paris. He was born in 1997.

Benito Gonzalez [01:45:30]:
We launched the Muse Mac app last week.

Leo Laporte [01:45:32]:
He's the world's youngest self-made billionaire. He was a billionaire at the age of 24.

Benito Gonzalez [01:45:37]:
And more. It is state-of-the-art.

Leo Laporte [01:45:40]:
Andrew Wang.

Benito Gonzalez [01:45:41]:
Built-in native.

Leo Laporte [01:45:42]:
Or Alexander Wang, right?

Benito Gonzalez [01:45:43]:
And big news today, we're adding computer use.

Leo Laporte [01:45:48]:
Uh-oh. So this is what Claude Cowork does. This allows you to use it to find files, delete files.

Benito Gonzalez [01:45:56]:
On your Mac, you can have it help you run your small business or just get work done for you. You can walk away from it.

Leo Laporte [01:46:03]:
For some reason, these always come out on the Mac, not Windows. I'm not sure why.

Paris Martineau [01:46:07]:
How about this? How about we get Muse to record, to plan the— our New Year's Eve livestream?

Leo Laporte [01:46:14]:
We could. If you had a Muse, you could have it do that.

Benito Gonzalez [01:46:17]:
You'll be able to add—

Paris Martineau [01:46:18]:
No, I want your Muse to do it. I will purchase it.

Leo Laporte [01:46:21]:
You could also connect with, uh, Jeff Atwood's Muse.

Benito Gonzalez [01:46:23]:
He won't do it either, probably.

Paris Martineau [01:46:26]:
No, Jeff wants to do the 24—

Leo Laporte [01:46:28]:
No, he doesn't want to do the Muse, is what I'm saying.

Paris Martineau [01:46:30]:
He doesn't want to do Muse, but we could communicate with it another way.

Leo Laporte [01:46:33]:
So I'm the one who has to take the arrows in my back.

Paris Martineau [01:46:36]:
No, your Muse has to. Are we gonna watch this, or do we want to do the rest of the show?

Leo Laporte [01:46:43]:
I think we've seen enough. It's really basically an ad, isn't it?

Paris Martineau [01:46:46]:
I'm sorry. I know that you guys love these things, but I have no patience for these events. I feel like they are— They are press releases.

Benito Gonzalez [01:46:54]:
Oh yeah.

Paris Martineau [01:46:54]:
And the most valuable information is always gleaned by reading a written version of it after the fact.

Leo Laporte [01:47:03]:
We don't need to talk about AI doom. You know what's interesting? I am finding the conversation shifting this week. Last week it was kind of a given AI is going to destroy us all. This week, I I know, Jeff, you don't feel like people have really covered Tesreau and all that, but I do feel like there's a little more skepticism.

Jeff Jarvis [01:47:23]:
There were 2. Michelle Goldberg did it, and I wrote about this, did a column in The New York Times, and Cal Newport did 2 columns in The New York Times that start to touch on it. But that's kind of it. Meanwhile, we have Washington Post, is this how the world ends? Extinction scenarios are taking over the AI debate. And then The Wall Street Journal, the anonymous math geek who quit Anthropic and became the face of AI Safety. They both ignore the entire backstory here of what's really going on. And, and no, I think that they're the— you also— but you do have coverage. Jensen Huang says, no, it's not going to destroy us.

Jeff Jarvis [01:47:57]:
Then you have—

Leo Laporte [01:47:58]:
You bring it on, says it's a— it's doomer panic and it's not. It's—

Jeff Jarvis [01:48:02]:
Yeah, we're seeing some voices come out, but that was— yeah, but I think, I think we're seeing it from people like that rather than from people in journalism. They don't want to ruin the story. It's too good a story.

Leo Laporte [01:48:15]:
Yeah. The good news is that the mainstream press has the attention span of a gnat.

Benito Gonzalez [01:48:22]:
Yeah.

Leo Laporte [01:48:22]:
And they've already moved on to something else. So it's not— yeah. So Jensen Huang says we're not going to all die. That's good. That's reassuring.

Jeff Jarvis [01:48:31]:
Right.

Paris Martineau [01:48:31]:
Jensen Huang actually also said in an interview with Ezra Klein this week that if the security events such like Hugging Face are actual incidents where these companies were not able to control or monitor these things, that they should pause development.

Leo Laporte [01:48:52]:
Yeah, they need to. Yes. I mean, what he's, I think, saying, though, is not the same thing as Dario Amodei's pause the frontier. What he's really saying is if you've got a product that isn't safe, you need to make it safe before you continue, which I think is exactly right.

Jeff Jarvis [01:49:09]:
Yes.

Leo Laporte [01:49:10]:
It doesn't mean we gotta— oh, we can't— we gotta stop this AI. We gotta really stop. It's going too fast. It's going too fast. No, it's— there's a subtle difference, but the difference is you guys are liable. There's product liability here. Don't create a product that's dangerous. I don't think that's wrong.

Jeff Jarvis [01:49:26]:
Agreed. So yes, I think there are those shifts among people who know what they're talking about.

Leo Laporte [01:49:33]:
And of course Paris's dad.

Paris Martineau [01:49:36]:
And of course my dad.

Leo Laporte [01:49:38]:
He's not worried, is he?

Paris Martineau [01:49:40]:
No, I don't think he would be able— if AI, uh, didn't keep growing exponentially, how is he going to write any email?

Leo Laporte [01:49:48]:
Ask him. Just say, hey Dad, do you think AI is going to replace us? Just ask him. I'd be curious what he says.

Jeff Jarvis [01:49:56]:
Do you have any idea what he talked to Claude about? Have you asked him?

Paris Martineau [01:50:00]:
I mean, I'll have to ask him.

Leo Laporte [01:50:01]:
I talked to Hermes about everything.

Jeff Jarvis [01:50:05]:
What did your mother think? So what was the— what was her—

Leo Laporte [01:50:09]:
Her—

Paris Martineau [01:50:09]:
the tenor of her response to all this, she's like, it is ridiculous.

Leo Laporte [01:50:13]:
Your father? Would you believe your father is in the other room talking?

Paris Martineau [01:50:17]:
She's often talking. She's like, yeah, I've really been into— they're always rewatching The Crown or or something like that. They're always rewatching some sort of—

Leo Laporte [01:50:23]:
I agree, it's The Crown.

Paris Martineau [01:50:24]:
Listen, all of my parents love The Crown, you guys included, I guess.

Leo Laporte [01:50:29]:
'Cause we remember the Queen.

Paris Martineau [01:50:31]:
I'm just saying.

Jeff Jarvis [01:50:32]:
But no, no, it's like, we really like rewatching The Crown.

Paris Martineau [01:50:35]:
This is probably their second, third, maybe fourth rewatch. This rewatch, I'm asking Claude about what is and isn't the act.

Leo Laporte [01:50:42]:
Oh, I do that. I do that.

Paris Martineau [01:50:44]:
Of course you do. But I mean, she's not a hater. We, at one point, had a lovely conversation where they They've, uh, their backyard has turned into a turtle sanctuary somehow where there's a lot of turtles.

Leo Laporte [01:50:58]:
That's better than the gators.

Paris Martineau [01:51:00]:
Listen, I know I'm worried about the turtles because of the gators, but we looked up all the information about the turtles and how to feed them on the phone one day, and she was using OpenAI and I was using—

Leo Laporte [01:51:10]:
You do what we did with our turtles. You give them a piece of lettuce.

Paris Martineau [01:51:14]:
Yeah, but they're—

Leo Laporte [01:51:14]:
I mean, they're like a little umbrella, plastic umbrella and a rock.

Paris Martineau [01:51:18]:
Well, they kind of— I was like, you guys should leave an area of The backyard, like, less mowed, so they have kind of trouble growing grass.

Leo Laporte [01:51:25]:
Yes, so they can hide in there.

Paris Martineau [01:51:26]:
A bit of a brush area.

Leo Laporte [01:51:29]:
Well, Jeff, I know, is pretty damned excited about the big Google event. They've announced new Google Chromebooks. But wait a minute, they're not Chromebooks, they're Google Books, and they're not running Chrome OS, they're running Android.

Jeff Jarvis [01:51:42]:
Tarted up Android. It's Android.

Leo Laporte [01:51:44]:
Tarted up Android. Android with lipstick.

Paris Martineau [01:51:48]:
Uh, yeah.

Leo Laporte [01:51:49]:
Are you— does that concern you at all?

Jeff Jarvis [01:51:52]:
Um, I, I, I don't know.

Leo Laporte [01:51:54]:
You went to the event.

Jeff Jarvis [01:51:55]:
I went to the event. It was, it was actually smaller than the last Chromebook event I went to. It was weird for something that's so important and so big. It was small. Um, uh, it is— they don't mention Chrome OS hardly at all. It's all— it's Android. It's Android so that, that it knows what you're doing on your phone. It goes across easily.

Jeff Jarvis [01:52:13]:
It's Android now with a decent file system. And it's Android with AI touches. You waggle the mouse and you can then bring Gemini into whatever that context is. It uses Rambler. So when you, when you um and ah and change your mind, then you can, you can do that. So it's really a showcase.

Leo Laporte [01:52:32]:
I can't wait to play with Rambler. I can't either.

Jeff Jarvis [01:52:34]:
I'm eager to. It's a showcase for Gemini, I think, which is really interesting. The hardware is really impressive. It's good.

Leo Laporte [01:52:44]:
The problem is $899.

Jeff Jarvis [01:52:46]:
They did not announce a MediaTek device. So all of these have fans, which tried to mean they're Intel.

Paris Martineau [01:52:51]:
Yeah.

Leo Laporte [01:52:52]:
I wish—

Paris Martineau [01:52:52]:
Are you concerned at all about— do you think that Android is capable of handling the wide variety of tasks you do on a computer in the way that—

Jeff Jarvis [01:53:01]:
That's what I was— it's a really good question. It's the exact right question, Paris. I mean, I asked one guy because the main thing I do— Why wouldn't it?

Leo Laporte [01:53:07]:
Isn't it Isn't a Chromebook just a browser, a Chrome browser?

Jeff Jarvis [01:53:10]:
Yeah, but no, no, the OS has certain things. So for example, when I do my wonderful ignored rundowns, um, I can save 5 saved gets in a row. So I can save the headline, I can save the, um, URL and do it for another story, another story, and then I can paste them in the right spot. And I use that constantly. That's— you don't have that in the Mac. I have that in Chrome OS. I don't know whether it exists there. The other big question we're going to have is, will I get the wiggly mouse and will I get all the neat stuff? Because my account is—

Leo Laporte [01:53:42]:
Works.

Paris Martineau [01:53:43]:
What is the wiggly mouse?

Jeff Jarvis [01:53:44]:
The wiggly mouse, you wiggle the mouse over something and, and it knows the context of what you're wiggling over. And then Gemini can say, how can I help you?

Leo Laporte [01:53:51]:
Do the same thing on a Pixel phone.

Paris Martineau [01:53:52]:
I would— that would drive me insane. I am— the one thing I'm already coming up and I haven't upgraded to iOS 27 yet or whatever is I'm accidentally triggering the smart Siri thing all the time. Like, I will accidentally sit my phone wrong and Siri will be trying to circle something on my screen and tell me about it. I take a screenshot, I take a screenshot, and it's like, I'm searching birds now. And I'm like, I don't need you to be searching about birds. I need to send this meme to my friends.

Jeff Jarvis [01:54:24]:
This week has been Paris the curmudgeon. It's great. I love it.

Paris Martineau [01:54:27]:
Listen, I've had nothing to do in the last week but think about the things that annoy me. I've been trapped in this house. I got cleared to exercise 5 hours ago.

Jeff Jarvis [01:54:35]:
This is oxygen on your brain. This is what happens when you get oxygen.

Paris Martineau [01:54:38]:
It is really, really—

Jeff Jarvis [01:54:40]:
So Jason Howell is in Hawaii at an event and he, uh, is he at the Snapdragon event? Um, is that what he's doing? I think so. Yeah, Qualcomm.

Leo Laporte [01:54:48]:
Yeah, I think so.

Benito Gonzalez [01:54:50]:
Yeah.

Jeff Jarvis [01:54:50]:
So he said—

Leo Laporte [01:54:51]:
This is the big junket, man. This is the junket. If you If you're in with the in-crowd, you get to go to Hawaii.

Jeff Jarvis [01:54:56]:
Yeah, he got to go to this.

Leo Laporte [01:54:57]:
Qualcos time.

Jeff Jarvis [01:54:58]:
So he texted me, just interviewed John Solomon, VP of Google Book and Chrome OS. My first question was for you. I'm sure you can imagine what it was. He didn't say anything more, but I'm sure it was, well, can Jeff use this with the Workspace?

Leo Laporte [01:55:11]:
My question is, is my Lenovo MediaTek-based Chromebook gonna be upgraded to Android?

Jeff Jarvis [01:55:19]:
I tried to ask people there. There weren't the right people there to ask. Um, uh, because there was a report that the later higher-end Chromebooks could be converted into Google Books.

Paris Martineau [01:55:30]:
Wait, how is Chrome not baked in in some sense if it's a Chromebook? Is it just software?

Leo Laporte [01:55:37]:
No, in fact, you can run Linux.

Jeff Jarvis [01:55:39]:
You can run Linux.

Leo Laporte [01:55:40]:
Oh yeah, there's just—

Paris Martineau [01:55:41]:
yeah, so I guess they could just decide to take your Chrome away. How would you feel bad about that, or do you feel that Chrome is wrong Do you in an existential sense?

Jeff Jarvis [01:55:50]:
No, I like Chrome. I've been using it for 10 years now. I'm used to it.

Leo Laporte [01:55:53]:
So it'll still be Chrome, the browser Chrome, the browser Chrome.

Jeff Jarvis [01:55:57]:
But will it have these other things?

Leo Laporte [01:55:58]:
And some functionality, maybe.

Jeff Jarvis [01:55:59]:
I'm curious, Leo, if you just look at the buy page, what do you think of the specs of the machines?

Leo Laporte [01:56:04]:
I'm not like you. I'm not thrilled that it's an Intel Core i7.

Jeff Jarvis [01:56:09]:
MediaTek is part of the deal, but they didn't— they didn't announce the MediaTek machine.

Leo Laporte [01:56:12]:
Yeah, MediaTek has a new chip called Dimensity, which will be in the Google Pixel. Pretty much the same as the old chip, I think. But yeah, the MediaTek Kompanio, the world's worst-named processor in the history of America— and there are some pretty bad names to compete against, like Itanium— the Kompanio is actually—

Jeff Jarvis [01:56:35]:
You can never beat 8080.

Leo Laporte [01:56:37]:
No, 8080, there you go. Now that's a processor.

Benito Gonzalez [01:56:40]:
8088.

Leo Laporte [01:56:40]:
8088, baby. The Z80. Uh, no, the Companion is a very, very good processor. I actually love the Lenovo.

Jeff Jarvis [01:56:48]:
Uh-huh. That's what I'm working on right now.

Leo Laporte [01:56:50]:
Yeah. Yep, we both have one of those. Um, I— my guess is that they will allow that to be upgraded, but we'll see. Whether they'll make us upgrade is the question. They certainly, I would think, do not want to continue developing a dead-end platform. Well, at some point they're going to say, well, Chromebooks Chromebook is dead. Chrome OS is dead.

Jeff Jarvis [01:57:08]:
At some point, but these are $1,000. The lowest priced machine is $900.

Leo Laporte [01:57:13]:
Um, does it have to be because of the—

Jeff Jarvis [01:57:16]:
I think so.

Leo Laporte [01:57:16]:
I think the AI stuff.

Jeff Jarvis [01:57:18]:
So I think if you want to sell $200 school Chromebooks—

Paris Martineau [01:57:21]:
How much is a Chromebook normally? My understanding, my understanding of everything has been warped by Apple jacking the price up on everything.

Jeff Jarvis [01:57:29]:
You can buy a basic machine for $300. You could buy a good machine for $400. And you can buy, you know, my machine—

Paris Martineau [01:57:39]:
this—

Jeff Jarvis [01:57:39]:
what did a Lenovo cost, Leo? I think $1,000.

Leo Laporte [01:57:41]:
No, I don't think—

Jeff Jarvis [01:57:42]:
$900.

Leo Laporte [01:57:43]:
I think it was in the same ballpark, around $800, $900.

Jeff Jarvis [01:57:46]:
$800, $900. Yeah, yeah.

Leo Laporte [01:57:48]:
Which is a whole lot higher, but it has an OLED screen, it has a touchscreen, pretty fast processor, a lot of— I think 16 gigs of RAM. It is not a cheap book. It's made of aluminum body. It's nice. It's a nice machine.

Jeff Jarvis [01:58:00]:
I've had to have the motherboard replaced twice, but—

Leo Laporte [01:58:03]:
Really?

Jeff Jarvis [01:58:03]:
Yeah.

Paris Martineau [01:58:04]:
How much does that cost?

Jeff Jarvis [01:58:06]:
Nothing, because, well, it cost me the warranty and I extended the warranty.

Paris Martineau [01:58:09]:
Nice.

Leo Laporte [01:58:11]:
Well, we'll watch with interest. I knew you were very interested in that.

Patrick Hillmann [01:58:15]:
Um, yes.

Leo Laporte [01:58:17]:
All right, let's see. I think we have 2 more breaks, don't we, Benito?

Benito Gonzalez [01:58:24]:
Yep.

Leo Laporte [01:58:24]:
Let's just do them all at once.

Jeff Jarvis [01:58:27]:
Neil's tired.

Leo Laporte [01:58:28]:
I'm tired. I haven't had lunch, let alone dinner. You're watching This Week in Intelligent Machines with Paris Martineau and Geoff Jarvis. So glad. I think we kind of covered the— did you have any stories that we missed, Geoff, that you wanted to do? Or Paris?

Jeff Jarvis [01:58:49]:
Hmm, let me look here. Xiaomi has a new model.

Leo Laporte [01:58:53]:
Yeah, Mimo. I mentioned that. I tried it out. I did a little what we call a bake-off here in the Laporte studio, and I've gone back to GLM 53. That's my— still my favorite model. Mimo looks good, and running in the cloud, it'll probably be very good. Mimo Pro looks to be as good as some of the top models. So this is a very important point, is that the Chinese companies are not holding back.

Leo Laporte [01:59:16]:
These are open-weight models, so you can run them Locally, if you have enough horsepower. You need a lot of horsepower to run the Mimo Pro 2.6 Pro, but the Flash version I could run. I did. That's the one I tested. I'm glad that they continue to come out with open weight models. As I said, Quen 4 is coming. I'm running a Quen image model that's as good, every bit as good as Midjourney was. I remember paying for Midjourney.

Leo Laporte [01:59:41]:
We can now run local models in small amounts of RAM that do amazing things. I think we're in a kind of golden age of local AI, personally.

Jeff Jarvis [01:59:51]:
That's where the hope is.

Leo Laporte [01:59:53]:
Yes. Yeah, I think open weight AI is going to be key to the whole thing.

Jeff Jarvis [02:00:00]:
So rumor is that Scott Bessant is likely to be the Trump AI czar, but Bessant said there's only one AI czar and that's Donald Trump, sir.

Leo Laporte [02:00:08]:
Oh Lord. I don't think a person with an IQ—

Paris Martineau [02:00:11]:
Sorry, SI czar.

Leo Laporte [02:00:13]:
SI czar.

Paris Martineau [02:00:14]:
Well, super intelligence.

Jeff Jarvis [02:00:16]:
Well, the other thing is it's also Sports Illustrated. So as somebody said in the reaction to it, can't wait for the, uh, AI swimsuit edition.

Patrick Hillmann [02:00:25]:
Uh, okay.

Leo Laporte [02:00:30]:
I think we can do picks of the week. Why don't you, why don't you kick things off? Actually, let me start because, uh, I have a Google thing. This is, we're in the Google vein. Actually, Paul Theroux turned me on to this, and I thought it's kind of interesting. Now, I don't have young kids, neither do you, Jeff, but Paris any day now is going to spring forth with a family. She says, what are you talking about? At least of the 3 of us, she's the most likely to, let's put it that way.

Paris Martineau [02:01:05]:
There was like a whole period of checking in for surgery where they're like, are you pregnant? Could you be pregnant? We need you to pee in a cup. And I'm like, guys, I'm not, I'm not, but we can waste our time here if you'd like.

Leo Laporte [02:01:18]:
No, they, they do that with all these, all these medications, you know? And, uh, it's like Alzheimer's medication, don't take if you're pregnant or expect to become pregnant. It's like, wait a minute.

Paris Martineau [02:01:28]:
It's like there's a couple things overlapping here that I think we need to—

Leo Laporte [02:01:31]:
Hold on there. Uh, CC, which I'm not sure what that stands for. Your family's AI agent. I think this is really interesting. A family can have a Muse-style agent. This is from Google that works across email, calendar, chats.

Jeff Jarvis [02:01:48]:
Then why didn't Google make it big? Why did Muse get all the attention and this didn't?

Leo Laporte [02:01:53]:
Uh, good. Well, it says experiment.

Paris Martineau [02:01:55]:
Guys, should we get one for our group chat? Would this be the way that we can move our group chat offline?

Leo Laporte [02:02:01]:
Oh, do you want to move off WhatsApp? We can move off WhatsApp.

Paris Martineau [02:02:03]:
No, I'm just always kind of joking around about it. It's just, you know, I'll be—

Jeff Jarvis [02:02:07]:
I'll—

Paris Martineau [02:02:08]:
I think we could, we could do a little— I think we could have a fun— we could have a family AI agent for the podcast.

Leo Laporte [02:02:14]:
Pretty fun. Oh, who's gonna drive? Who's gonna drive Paris to the soccer game?

Paris Martineau [02:02:19]:
Jeff? Did Jeff take a sleeping pill on the flight?

Leo Laporte [02:02:24]:
Uh, it is an opt-in Google Labs experiment. You have to be located in the US.

Jeff Jarvis [02:02:29]:
Let me guess, I can't do it.

Leo Laporte [02:02:30]:
Um, and I bet you, you can't, but it doesn't mention anything. It just says— oh yeah, no, you can't. It would have reminded Henry about your flight, about your visit. Um, you of course have to give it access to your Gmail and your calendar and all that stuff. But the idea is that a family has a family calendar, a family has Family mail and stuff like that. And I think that's, it's kind of cool. So CC is, uh, and I wonder, carbon copy? That doesn't seem like what CC should—

Jeff Jarvis [02:03:02]:
I guess you CC mom and CC dad. It's a very—

Leo Laporte [02:03:05]:
CC mom, CC dad.

Jeff Jarvis [02:03:06]:
Corporate way to look at it.

Leo Laporte [02:03:08]:
Okay. Okay. From Google Labs, a family's AI agent. I think this is a great idea. Makes me want to have another family. No, not at all. Not even once. Not even in my worst You've had a few.

Leo Laporte [02:03:23]:
I've had my family's. Lisa and I will look at babies and smile and then say, thank God we're not there. But Paris, pay no attention to that. If you're pregnant or planning to become pregnant—

Jeff Jarvis [02:03:37]:
Where the hell?

Paris Martineau [02:03:40]:
One of my friends recently had a baby, and I met him for the first time a week or two ago.

Leo Laporte [02:03:44]:
I saw your pictures on Instagram.

Paris Martineau [02:03:45]:
He looks like a little Italian man that you see in the back of an Italian restaurant who's like, really fretting over something, and it's because he's accidentally in the mob. And one of us told her that, and she was like, I can't wait to tell my husband. He's been so worried that he doesn't look Italian.

Leo Laporte [02:04:03]:
Oh, I don't think he looks Italian, but he looks like he's a man, man.

Paris Martineau [02:04:07]:
Listen, babies do look like little old men.

Leo Laporte [02:04:10]:
They do. They do. They have odd— yes, I agree. They all look like Winston Churchill, somebody once said. Paris, your pick of the week.

Paris Martineau [02:04:19]:
My pick of the week, I've already spoiled for you guys, which is that, um, one of the things that has come out, I guess, over the last week was, especially as 5.5 came out, is this video about AI that has been both— it's— I'm not sure whether the lyrics were written by AI. Someone traced them back to a YouTube video from 2024, but it clearly was written by someone very involved in AI Twitter at the time. And then someone had it, I guess, be sung with like Suno or something and then animated it using Opus 5.5.

Benito Gonzalez [02:04:54]:
Oh.

Paris Martineau [02:04:54]:
And it basically two-shotted it. The first prompt or something for it, there's a very interesting GitHub breakdown that I've also included as a link in there of this person just basically asked like, hey, do we like an animated version of this featuring Claude bot? And this is the second version of it, which was just one prompt. And it's honestly quite good. for Opus.

Leo Laporte [02:05:16]:
Sounds like Olivia Rodrigo, or—

Paris Martineau [02:05:18]:
It's also just a catchy song. It's been stuck in my head all day.

Leo Laporte [02:05:21]:
The chorus—

Jeff Jarvis [02:05:21]:
I'm hoping your pee doom.

Paris Martineau [02:05:32]:
It's just a bunch of AI references as well, so I don't know. I found it very— it gets— there's a lot of like really good ones in the end about Ilia and various recent developments in AI that I just, I found it very charming.

Leo Laporte [02:05:46]:
One of the biggest developments, actually what's interesting about this is this is not with a video generation tool.

Paris Martineau [02:05:52]:
That's the thing that I thought was very interesting about this. So I thought the animation was cute in a way that I normally do not find AI animation. And I think it's because what they asked it to do do was generate it in— where was it here?

Leo Laporte [02:06:10]:
It's code.

Paris Martineau [02:06:12]:
It's code, basically. So the animation guide, he's broken it down.

Jeff Jarvis [02:06:15]:
Oh, I see.

Paris Martineau [02:06:16]:
The project renders into a 156-second music video as a painted watercolor animation with a P5 brush. Frames are rendered offline in headless Chrome. So speed matters less than quality. And they— so basically They had 2 prompts for this. One was just like, give your best shot, use a Claude animation. It was a little rudimentary. And so then the person did a second prompt with basically not that much detail being like, we want it to be kind of cutesy, a little rough around the edges. We want it to feature Claude and to have kind of a lot of different stuff happening in every frame.

Paris Martineau [02:06:50]:
And that was it. It was a very sparse prompt. But I also then turned on 5.5 extra high and going through this GitHub, it's kind of crazy the amount of stuff that it then did. It developed a whole animation guide. It did a full storyboard for the animation agent that goes through a specific cast, has second-by-second, like, shot in and wipe-in and wipe-out descriptions for it, which is just kind of interesting to see the way that it worked through all of this, and then kind of works through the rendering in a very interesting way. I mean, in one, I guess in 2 takes, it generated a cute kind of quote unquote hand, I guess, Claude drawn, um, animation style, which synced, uh—

Leo Laporte [02:07:38]:
This week is to create a Jeff Jarvis potato tomato squash game with Claude 5.5 that this time is the best. You wanna do it or should I?

Jeff Jarvis [02:07:50]:
I will.

Paris Martineau [02:07:51]:
I mean, I—

Leo Laporte [02:07:52]:
I think you should.

Paris Martineau [02:07:53]:
I will do it, but I'll need to do it next week because I have things I need to do that will use all of my limit. I'm already recognizing, but I would totally do it.

Leo Laporte [02:08:00]:
But this is good. So you've found stuff that you want to do with Claude.

Paris Martineau [02:08:04]:
I do things with Claude most weeks.

Leo Laporte [02:08:07]:
But do you feel like now there's more you can do because of this improved model?

Paris Martineau [02:08:11]:
I do think that this improvement this week, I mean, this is, yeah, I do think so. Also, there has also been an update. I'm not sure if it was this week, but it was sometime in the last month or 2 or recent months that it upped the PDF ingestion limit, which was a big problem for me. And so I'm still chunking it. I mean, there's just like some route tasks where I've got a lot of poorly copied, not even poorly copied, just like images of text that are poorly OCR'd that are a bunch of different, like little boxes of text that I need to be copied over into a spreadsheet in a very specific way. And I just figured out today how I would be able to prompt it to do that. And it did kind of a rudimentary version of this with a different dataset that made me realize that I'm going to set Opus 5. And it did the first one without any mistakes that I could find, which was quite useful to me.

Leo Laporte [02:09:10]:
Yeah, it's amazing. It's pretty good. They're getting better and better. Every few months there's a new model that's notably better.

Paris Martineau [02:09:18]:
Yeah, I've told you guys before, I mean, one of the things I've used for this larger story I've been working on for a little bit over the last couple of months is that there's a lot of different documents and data and things that all have like dates tied to them. And so as I've been going through like literally piles of thousands of documents, I've just realized like I had Claude just make a little web app for me that is an auto-sorting timeline that I can put in stuff from any time period, and it automatically adds it to that and creates a database that separately I can import into Google Sheets later. And it's just been really helpful for organizing my work.

Leo Laporte [02:09:52]:
Sounds great.

Jeff Jarvis [02:09:55]:
Okay, so adding on to Parris's video, the video that took over, um, AI discussion on Twitter today, I added it at the bottom of the rundown. This is not my pick, but I thought maybe—

Leo Laporte [02:10:05]:
The Race for AGI. This is actually from Instagram. But so I— one of the things that's really happened is, uh, video generation has gotten really good. Even local video generation. A lot of these people, a lot of people are doing these. Remember with Suno, you, you know, you had— there were copyright restrictions and you couldn't use real people and stuff. Now people are doing this stuff locally and you can pretty much do anything you want. This has got Elon Musk and Sam Altman shooting at each other.

Leo Laporte [02:10:39]:
Uh, I've seen a lot of these. There was a rap battle. You've— there's Jensen driving a truck. It was this Mad Max kind of— yeah, meets— and there's Dario in a dune buggy. And— or is Dario driving? Oh, there's Donnie. President's got his own car. Or is he—

Jeff Jarvis [02:10:57]:
No, he's got Air Force One.

Leo Laporte [02:10:58]:
He's in Air Force One.

Jeff Jarvis [02:11:00]:
One of them.

Leo Laporte [02:11:00]:
That's pretty funny. There's a lot of these. I mean, X is filled with— you know, somebody did that new Elizabeth Holmes movie with Well, it's been done again and again with a variety of different people. I saw the latest one I saw had the president in it.

Jeff Jarvis [02:11:16]:
Which was hilarious. Why would I lie to you?

Leo Laporte [02:11:18]:
Yeah, why would I? Why would I deceive you? Uh, yeah, I think it's really interesting. Um, the leap we've made in the ability to make video and, and make it at home is really remarkable.

Jeff Jarvis [02:11:31]:
Yeah. All right, so go ahead.

Leo Laporte [02:11:34]:
I did not mention this, but, uh, We're going to try to get, uh, on at least one of our shows, if MacBreak Weekly, if not Intelligent Machines, um, Federico Betticci, who has one of the few Mac Studios. Actually, Micah just got his today or yesterday, but he's got a 256-gig, $15,000 one that Apple loaned him. Federico does. And, uh, his review is out and it shows to be a very good local AI machine. And a lot of the concerns I had about its pre-fill capabilities and others are diminished.

Jeff Jarvis [02:12:12]:
So as good as 2 Sparks, Leo?

Leo Laporte [02:12:15]:
Well, that's the question. It's the same amount of unified memory right now. No, believe it or not, on the model I run, it still runs better on 2 Sparks, but that's because this is brand new, and I have a feeling the software MLX software hasn't been optimized in the same way. People have been banging on these Sparks for almost a year optimizing the software. I suspect we're going to see better performance out of the Mac in the long run. In theory, it should be. So, uh, that's going to be very interesting, um, and it's going to cause great consternation in my life come end of October because I really don't want to buy a $15,000 I'm actually very happy with the Sparks. I'm getting the performance I want and the quality I want.

Jeff Jarvis [02:13:04]:
Glad you don't have buyer's remorse.

Leo Laporte [02:13:05]:
I do not. You can't, by the way, go to NVIDIA, see if you can buy a Sparks.

Jeff Jarvis [02:13:10]:
Really?

Leo Laporte [02:13:11]:
They're sold out. They aren't— the prices have gone up and up and up, and now most outlets no longer have DGX Sparks available, perhaps because the RTX Spark, which is the Windows-based version of this, is coming out from a number of manufacturers. Maybe that's what they're doing. They're putting their money Or maybe Rubensparks. Well, I would love to hear a Rubenspark, but I'm— Father Robert said he had heard from companies that were working on it last week. We shall see, a Spark 2 perhaps.

Jeff Jarvis [02:13:44]:
I also think, we talked about this when Father Robert was here, I think a device with the Groq, with the Q inference chips, is going to be more interesting.

Paris Martineau [02:13:53]:
Really unfortunate name for that chip.

Jeff Jarvis [02:13:55]:
It is.

Leo Laporte [02:13:56]:
It's Groq with a Q, not Grok with a K, because Grok with a K is Liza. Uh, thank you, Jeff Jarvis.

Jeff Jarvis [02:14:03]:
So no, wait, wait, wait.

Paris Martineau [02:14:04]:
I didn't mind.

Jeff Jarvis [02:14:05]:
I said that wasn't my pick.

Patrick Hillmann [02:14:06]:
I was just adding—

Leo Laporte [02:14:06]:
Oh, that's not your pick. It's just something.

Jeff Jarvis [02:14:09]:
Yes.

Leo Laporte [02:14:09]:
What's your pick?

Jeff Jarvis [02:14:10]:
I want to mention, oh, I want to mention real quickly that I was in San Francisco for Handshake at an event. I moderated a panel. I, I got travel paid. I didn't get paid for this, but it was a favor. Um, introducing the Handshake AI Skills Studio. Which is interesting because it's a way that students can make something with AI and make that part of their portfolio.

Leo Laporte [02:14:30]:
Ah.

Jeff Jarvis [02:14:31]:
Which is interesting. And in the discussion, uh, Brian Johns-Rood from OpenAI was there, and he brought some stats with him about the use of AI. And he said within companies, 40% of the use people are doing are not for their jobs. That doesn't mean that they're doing porn. It means that they're doing something that's outside of their corporate silo. In another part of the company, right? A product person is doing something about marketing. A marketing person is doing something about product. And so it's really interesting to see how people are using these things, and it has an impact on education.

Jeff Jarvis [02:15:00]:
I wanted to just give them a plug because the little tiny pumpkin cheesecake things were really good.

Leo Laporte [02:15:06]:
There's nothing like a good pumpkin cheesecake.

Jeff Jarvis [02:15:11]:
But the other thing I want to mention—

Leo Laporte [02:15:12]:
so if you're a student, you can do this for free.

Jeff Jarvis [02:15:14]:
Yeah, I think so. Yep.

Leo Laporte [02:15:15]:
And nice.

Jeff Jarvis [02:15:16]:
And get a portfolio.

Leo Laporte [02:15:17]:
So it says stop doom scrolling and start building. I like that. So speaking of which, best way to handle that.

Jeff Jarvis [02:15:23]:
Now we have another contrarian view here. Andreessen Horowitz has started the Horowitz Andreessen Academy, telling students not to go to college, instead to go to them for 2 years. And you're going to do— and I started an entrepreneurial program at CUNY. This is an entrepreneurial program. where instead you're— no homework, no tests. Instead, you're going to listen to all these famous AI people come in and lecture you, and then you're going to build things. And, um, yeah, if you look at the curriculum, which is online, uh, 167—

Leo Laporte [02:15:56]:
Can I recommend that you do not do this?

Jeff Jarvis [02:15:58]:
Don't do this. Yes.

Benito Gonzalez [02:15:59]:
Yes.

Leo Laporte [02:16:00]:
That you will be more valuable to yourself and to the world at large if you get a good liberal arts education. which turns out to be much more useful.

Jeff Jarvis [02:16:07]:
Much more valuable. And CS majors are way down in numbers. 14% fewer graduate school CS majors, 11% fewer 2-year CS majors.

Leo Laporte [02:16:19]:
Because it's coming from a big venture capital firm, I think a lot of it is business-focused.

Jeff Jarvis [02:16:23]:
It's all business-focused, but it's also kind of—

Paris Martineau [02:16:26]:
It's all about creating people who, let's say best-case scenario, it's all about creating people who will create specific venture capital firm more money in the way that is most useful for them. And that's like if— and that's if everything is above board and is actually useful and informative, which is a huge if.

Leo Laporte [02:16:44]:
Learn about the world, learn history, learn the arts. You'll be so much more valuable to yourself and the world and AI if you have that kind of background than if your only background is how can we make AI make some better products.

Jeff Jarvis [02:17:01]:
This is also, uh, I called it the seminary for the wealth gospel, the prosperity gospel.

Leo Laporte [02:17:06]:
I love it.

Jeff Jarvis [02:17:07]:
One course is titled The Geometry of Luck, in which students will learn about arranging a life so good fortune flows through it. How's that for obnoxious?

Benito Gonzalez [02:17:17]:
Wow.

Jeff Jarvis [02:17:18]:
Yeah, it's called Be Born Rich.

Leo Laporte [02:17:19]:
By Soleil, who's a designer and angel investor. You know what he did that was smart? He invested early in Dropbox and Facebook. Good timing is, uh, is the key to luck. Yeah, well, he was lucky. You know, if anybody's gonna— it should be somebody who's lucky who taught you. I wonder what the media and entertainment track— let's see, Michael Ovitz, Justin Kan, who is also lucky of Justin.tv, uh, made a lot of money. Cultures created— yeah, boy. Yeah, yeah, okay, fine.

Jeff Jarvis [02:17:52]:
Yeah, and the first year—

Leo Laporte [02:17:53]:
is it free?

Jeff Jarvis [02:17:54]:
The first year they're all scholarships, but then they make make a point of saying that it's going to be highly selective and have a tuition comparable to leading universities.

Leo Laporte [02:18:03]:
Oh, please go to school. Go to a good school. Go to your local community college, for crying out loud.

Benito Gonzalez [02:18:12]:
Yeah.

Leo Laporte [02:18:13]:
Is that it?

Jeff Jarvis [02:18:16]:
Yeah, I think so. Yeah. Well, there's, there's— I can't—

Paris Martineau [02:18:19]:
You know, Jeff has 50 more links in this document.

Jeff Jarvis [02:18:23]:
One more. We have a new show coming up from— I love The Good Wife. It was a great show. And the producers of that are going to create Cupertino.

Leo Laporte [02:18:31]:
And you know what's great about The Good Wife is they were right on the breaking news. They covered things that were really happening that, you know, that day in news. So this is a show about tech?

Jeff Jarvis [02:18:44]:
Well, it's lawyers in Silicon Valley.

Leo Laporte [02:18:46]:
Oh dear.

Jeff Jarvis [02:18:47]:
From Michelle and Robert King.

Leo Laporte [02:18:49]:
Filmed in New Jersey, as all good shows about Silicon Valley are.

Jeff Jarvis [02:18:53]:
Yes. An expansive view of the Hackensack River. What more could you wish for?

Leo Laporte [02:18:58]:
Is there anybody I know in this?

Jeff Jarvis [02:19:00]:
Have they cast it yet?

Leo Laporte [02:19:02]:
Oh, they haven't cast it yet.

Jeff Jarvis [02:19:04]:
Yeah. No, they— I think they are, but I don't know.

Leo Laporte [02:19:06]:
Mike Colter and Rachel Keller. I don't know.

Jeff Jarvis [02:19:08]:
Oh, there he is.

Leo Laporte [02:19:10]:
Yeah.

Paris Martineau [02:19:10]:
I mean, I think it is— this is one of the few tech shows I think I might actually watch.

Jeff Jarvis [02:19:14]:
Yeah.

Paris Martineau [02:19:15]:
It's about— I think, isn't the conceit that it's lawyers who I think one of their first cases is helping a Silicon Valley person, like, break their NDA to report something. And then they become— they are, like, very focused Silicon Valley malfeasance lawyers. That sounds like a great theme for a show.

Leo Laporte [02:19:32]:
I can't wait until there's CIS Silicon Valley and Silicon Valley Hospital.

Jeff Jarvis [02:19:38]:
He keeps wanting to kill people, Paris. He's still wanting to kill people.

Paris Martineau [02:19:42]:
That's why he wants fodder for the show. He's pitching to the networks.

Leo Laporte [02:19:45]:
It's the end of the world as we know it, and I feel fine. Thank you everybody for joining us. Paris Martineau is at Consumer Reports where it sounds like you're doing a big investigation on something or other.

Paris Martineau [02:19:58]:
On something or another. That's what I do.

Leo Laporte [02:20:00]:
I can't wait to find out. Well, we'll find out someday.

Paris Martineau [02:20:04]:
Exciting things. Yeah, I'll tell you all about it whenever it's here and ready.

Benito Gonzalez [02:20:09]:
Yeah.

Paris Martineau [02:20:10]:
I've got a cat rubbing against my legs, but if I bend down to pick her up and hold her to the camera, she's gonna run away as fast as possible.

Leo Laporte [02:20:16]:
Oh, well, we know she's there, and that's what really matters. It's, it's, you know, it's Heisenberg.

Paris Martineau [02:20:21]:
She let me get so close this time. I was 1 inch from her head.

Leo Laporte [02:20:24]:
Schrödinger's cat. If you reach down to pet it, it doesn't exist anymore. Will it be there?

Paris Martineau [02:20:30]:
No.

Benito Gonzalez [02:20:30]:
No.

Leo Laporte [02:20:31]:
Only if you don't observe it will it be there. And Jeff Jarvis, who's the author of Hot Type, out now. Go get it. Jeff Jarvis is going to type. Where am I going tomorrow?

Jeff Jarvis [02:20:42]:
I'm going to the Montclair Book Center to sign 48 more copies. They're out, they're sold out right now, but they'll have more.

Leo Laporte [02:20:50]:
That's awesome. Congratulations.

Paris Martineau [02:20:52]:
I gotta get you to sign my copy.

Jeff Jarvis [02:20:54]:
Of course. Yes. Well, if we— yeah, if we would have, if we'd seen each other for lunch.

Paris Martineau [02:20:58]:
Salt and Pepper closed the restaurant because of the virus. Do you want me to go beat him up?

Leo Laporte [02:21:03]:
I'll go.

Paris Martineau [02:21:04]:
I'll go. You know, I'm— I've got 100% of oxygen now. I could do anything. I could probably throw up.

Leo Laporte [02:21:08]:
It was so frustrating. I came so close to having one of those fresh—

Paris Martineau [02:21:13]:
Genuinely, my heart breaks.

Leo Laporte [02:21:14]:
I almost tasted it, and then he pulled it back. He said, oh.

Paris Martineau [02:21:18]:
We could overnight one to you. Do you think they'd allow me to send a sandwich in the mail? We'll do it.

Leo Laporte [02:21:23]:
I might have to do that. I'm gonna text him and say, look, could you just FedEx me a sandwich? I'll eat it on the air. Even though it's not gonna be as good.

Paris Martineau [02:21:31]:
You could do a sandwich unboxing video.

Leo Laporte [02:21:35]:
I didn't get the new iPhone, but look at this, it's almost as expensive. A delicious salt hank sandwich.

Paris Martineau [02:21:42]:
And you could, you could make it a really good clickbait video because it could be like a— I'm sure it would cost, you know, $60 to overnight it. So it'd be like a $97, $100 salt hank sandwich. Unboxing the $100 salt hank sandwich.

Leo Laporte [02:21:57]:
Oh, I like it. There's the title right there. They're unboxing the Salt Hank $100 sandwich. Is it worth it?

Paris Martineau [02:22:05]:
In parentheses, no, because it's been overnighted from New York to California.

Jeff Jarvis [02:22:10]:
Really soggy.

Paris Martineau [02:22:11]:
I mean, I think you'd have to possibly give you the bread separately, maybe uncut, and then—

Leo Laporte [02:22:18]:
He's coming out in 2 weeks. I'm just gonna say, look, you need to lock him in a room, bring the ingredients. Yeah, and make it for He won't though. I think honestly, I've never— how is this possible that I have a world-famous sandwich-making son who has never once made me a sandwich? I made him sandwiches. Wait, he's never made you a sandwich? I can remember vividly, he kept saying, Dad, I want a panini maker.

Paris Martineau [02:22:45]:
At what age?

Leo Laporte [02:22:48]:
Like when he was 8, he wanted a panini maker.

Paris Martineau [02:22:50]:
That's actually, that's You gotta remember that for the book.

Leo Laporte [02:22:53]:
I should have known he was gonna be something.

Jeff Jarvis [02:22:57]:
And holding his sausage. Why would I want a little tiny Hank holding his sausage?

Leo Laporte [02:23:01]:
We put a salami in his Christmas stocking one year. He ate the thing in one day, and he writes about it in his cookbook. He says, my life changed. I knew that someday I would be working with salami. He loves his sandos.

Jeff Jarvis [02:23:20]:
Anyway, um, Jersey Mike's just had a huge IPO.

Leo Laporte [02:23:23]:
Yeah, well, now that he's owned by Wonder, uh, I will be watching the Wonder IPO with great interest. Oh yeah, not that I'll get any of that, but I, you know, it's nice to, you know, see—

Jeff Jarvis [02:23:34]:
You could get a little friends and family action, couldn't you?

Leo Laporte [02:23:39]:
I don't know, you know, it's fun, it's fine. I don't want any But he did. So the guy who painted the paintings in the restaurant got 1% of it.

Jeff Jarvis [02:23:51]:
That's like the guy who painted the wall at Facebook.

Leo Laporte [02:23:53]:
Yeah, it's like, really?

Benito Gonzalez [02:23:57]:
Okay.

Paris Martineau [02:23:58]:
And he can't make you a sandwich.

Leo Laporte [02:23:59]:
But he can't make me a sandwich.

Jeff Jarvis [02:24:02]:
Did you buy him the panini press?

Leo Laporte [02:24:05]:
No, that's not— he's holding that against me. I never, I never gave him a panini press, and I think that's why he's holding it against me. Could have been. Could have been. What could have been?

Jeff Jarvis [02:24:17]:
He could have been— say he could have been Sam Foreman.

Leo Laporte [02:24:22]:
Uh, did, did Wonder just— did Wonder buy DoorDash? They own Grubhub. They can't possibly— there's a partnership with the DoorDash partnership.

Patrick Hillmann [02:24:34]:
Yep, yep.

Leo Laporte [02:24:37]:
Anyway, thank you everybody for joining us. I don't know why I'm talking about this. We appreciate your patronage. Uh, I know you would make me a sandwich if you could, and that's all that matters. If you like this show, you can watch it.

Paris Martineau [02:24:50]:
If you want to make Leo a sandwich, make him a sandwich by leaving him a— leaving a positive review of this show on your podcatcher of choice.

Leo Laporte [02:24:58]:
That's the sandwich I want. Absolutely. And maybe, maybe, just maybe, Paris Martineau will do a dramatic reading of it on an upcoming episode.

Paris Martineau [02:25:08]:
If you're interested in getting Leo's muse agent to run our 24-hour New Year's Eve livestream, leave a 5-star review. And if we get enough of them, he'll have to do it.

Leo Laporte [02:25:21]:
I sent you guys the opening narration of the documentary it's making for me, right?

Jeff Jarvis [02:25:29]:
Yeah, yeah, go ahead and play it.

Leo Laporte [02:25:30]:
It's making a documentary.

Paris Martineau [02:25:32]:
In a world where Leo has been doing radio for 50 years.

Leo Laporte [02:25:37]:
Yeah, basically it said, hey, I noticed you've been, uh, you've been broadcasting for 50 years. Do you want to make a documentary?

Jeff Jarvis [02:25:45]:
So this was its idea?

Leo Laporte [02:25:47]:
Oh yeah. Oh, I didn't—

Paris Martineau [02:25:49]:
Real dreamy stuff.

Leo Laporte [02:25:51]:
I should have just said, yeah, but surprise me.

Jeff Jarvis [02:25:54]:
Are you giving it access to all of your audio?

Benito Gonzalez [02:25:57]:
Archives?

Leo Laporte [02:25:58]:
I don't need to give it access. They're all public.

Jeff Jarvis [02:26:01]:
Ah.

Leo Laporte [02:26:02]:
Right? I don't have to.

Jeff Jarvis [02:26:03]:
What about your early career, Leo? Don't you have—

Leo Laporte [02:26:07]:
It has everything it wants. It's doing— it sent out emails to everybody I've worked with except Paris. It didn't deem you— actually, it didn't deem Jeff worthy either. I had to say, well, don't forget some of my other hosts. And, you know, Jeff also, I've been working with him. How long have I been working with you?

Paris Martineau [02:26:30]:
Almost 20 years. How did you guys meet?

Jeff Jarvis [02:26:33]:
Leo came to me because of, uh, I wrote the book, What Would Google Do?

Leo Laporte [02:26:36]:
Well, that's right. It was This Week in Google. And I came to you. By the way, I see Gina Trapani is starting her own podcast now.

Jeff Jarvis [02:26:42]:
Really?

Benito Gonzalez [02:26:44]:
Yes.

Paris Martineau [02:26:44]:
Is she still not returning your guys' emails?

Leo Laporte [02:26:46]:
She's still not. She's in league with Salt Hank.

Paris Martineau [02:26:50]:
They're, they're They're blocking you out.

Leo Laporte [02:26:53]:
It's a conspiracy. Uh, I can't find the narration anymore, unfortunately, that it did for me. It, it decided that it wanted to do a, um, kind of David Attenborough nature narration. And it's pretty funny. It's, it's, it's, it's inaccurate in many ways. So I have to decide how much do I want to weigh in? Do I want to Um, correct it, or I think it'd be funnier if it's just completely hallucinated, which would even be better. Anyway, maybe it'll be our holiday special. I don't know, we haven't decided yet.

Leo Laporte [02:27:32]:
Thank you everybody for being here. We'll see you next time, 2 PM Pacific, 5 PM Eastern, 2100 UTC, streamed live on YouTube, Twitch, X, Facebook, LinkedIn, Kick, and of course in our Club Twit Discord. You'll find it on YouTube. You can leave a review there. youtube.com/twit is the main page, and there's an Intelligent Machines page. Or, or look for Intelligent Machines in your podcast client. That's where you can leave that nice review, uh, and subscribe. That way you'll get it automatically as soon as it's done, audio or video.

Leo Laporte [02:28:00]:
Although the video has all the pictures, the audio does not, in case you didn't understand the difference. Thanks for being here. We'll see you next time on Intelligent Machines.

Paris Martineau [02:28:11]:
Bye-bye. I'm not a human being, not into this animal scene. I'm an intelligent machine.