This Week in Enterprise Tech 518 Transcript

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


Louis Maresca (00:00:00):
On This Week in Enterprise Tech, we have Mr. Brian Chi, Mr. Cruz Franklin joining me on the show today. Now the EU has it out for Apple to comply with some device standards and they aren't happy about it. Plus, business process automation is a challenging task. Well, today we have Danny Shayman, machine learning product manager of nru, and he's gonna talk us through the concept of AI to automate your processes and the concept of responsible and explainable ai. Definitely shouldn't miss it. Quiet on the set

VO (00:00:28):
Podcasts you love from people you trust. This is, TWiT

Louis Maresca (00:00:41):
This is this week in Enterprise Tech, episode five 18, recording November 4th, 2022. Skynet star linked to ai. This episode of this week in Enterprise Tech is brought to you by ON Logic. On Logic is helping innovators around the world solve their most complex technology challenges using on logic industrial computers, which are engineered for reliability, even in environments that would challenge or destroy traditional computer hardware. Learn more and find out about on Logic's 30 day risk-free hardware trial by visiting on and by Cisco orchestrated by the experts at cdw. When you need to get more out of your technology, Cisco makes hybrid work possible. CDW makes it powerful. Learn more at

Welcome to TWiT this week in enterprise tech, this show that is dedicated to you, the enterprise professional, the IT pro in that geek just wants to know how this world's connected. I'm your host, Lewis Meka, your guy through the big world of the enterprise, but I can't guide you by myself. I need to bring in the professionals and the experts on their very own and very busy senior analyst at md. He's the man that has the pulse of the enterprise and security. He's Mr. Curtis Franklin. Curtis make your fair. Orlando is here. How are things going there?

Curt Franklin (00:02:06):
Uh, maker Fair Orlando is indeed here. You can tell from the very subtle colored shirt that I'm wearing, I'm on the crew for, uh, maker Effects. Our Maker Fair, uh, maker Effects is the group that produces it. Uh, we've been there for two and a half days already getting ready for the fair. And, uh, boy, lots of exciting stuff going on. Uh, lots of battling robots, uh, will by this evening have a place to go into battle. Uh, we have all kinds of cause players and, uh, arcade players where you've got the vintage computer folks, we, it's just amazing. And, uh, it's wonderful being in the middle of that much creativity, let loose upon the world. So, uh, I'm gonna have fun this weekend and come out of it terribly exhausted, but fully recharged.

Louis Maresca (00:03:03):
Indeed, you had me at battling robots, so I'm, uh, I'm sold on that one. So definitely looking forward to some of the coverage there, <laugh>. Fantastic. Well, speaking of looking forward, we also have looking forward to Mr. Brian Chi. He's back on the show as well. He's net architect to skyfire. He is network expert, security professional, all around Tech Geek, and he is also very busy at Maker Fair, cheaper. How's it going on your side?

Brian Chee (00:03:24):
I'm actually having a lot of fun. Um, one of the things that I'm actually really looking forward to is Steam roller printing. Uh, one of our sponsors is a general construction company and they have donated the use of one of their steam rollers. So Kurt and his lovely wife Carol did some laser cutting on, um, some wood printing sheets, I guess you'd call 'em. Mm-hmm. <affirmative>. And, um, people are gonna be able to go and ink them up, put a sheet of paper on it, and then have a steam roller go right over it and they can take the prints home with them. That's gonna be a lot of fun. And, uh, I've also been having a lot of fun working with the, uh, central Florida fairgrounds folks on helping them to upgrade their network. So if all goes well next year, I'll just be able to walk in and ask them, could you restore the Maker Faire profile for me? And I'm done. So maybe next year I could exhibit something I've created. Ought to be fun.

Louis Maresca (00:04:27):
Yeah, sounds like a lot of fun. Well, I, uh, don't envy you guys on the sleep side of things, but I do envy you, uh, for seeing all the school stuff, so enjoy the show. Well, speaking of busy, it's been a busy week in enterprise. It would definitely should get started. The hardware market is definitely a hard one. Device companies are always under scrutiny to provide better and more compatible hardware. Apple is no exception here. Now we're gonna discuss a recent move by the EU to Apple to actually have them comply with more standards, plus artificial intelligence sparks new markets, and it gives you the ability to do things at scale that you never dreamed of. Now, automating tasks is definitely one of those things. Well, today we have Danny Shayman, he's machine learning product manager of in Rule, and he is gonna take us through the concept of AI to actually automate your business processes and the concept of responsible and explainable ai.

So definitely shouldn't miss that. Definitely stick around lots of stuff to talk about, but first, like we always do, we have some enterprise tech news to talk about. So let's go ahead and jump into this week's news. Blips, it's hard not to talk about the latest news regarding Twitter that has just evolved over the last two days. Let's go over it. There's some speculation. Uh, over the past several weeks, whether Elon Musk would execute a plan to cut costs involving mass layoffs at Twitter, he began his reign at Twitter by firing several exists before November 1st. Since then, there was no news of additional layoffs, especially since the California Warn Act requires an ample notice of large layoffs. It was thought that the rest of the layoffs were just speculation. Unfortunately, Thursday evening, employees of Twitter notice that they could no longer log into the devices or their Slack application.

They then receive an email from corporate saying they would be announced announcing layoffs early Friday morning over email. They either will stay or they'll go. They're also closed, all the corporate buildings and ask that people go home. In fact, the remaining people in the buildings were escorted out until further notice. Now, several people close to the process say the layoffs will be deep and broad and hitting almost half the organization, nearly 3,700 people. Now, people were confused about the closing of the corporate buildings ahead of time as part of the process. Now I've, I've been through layoffs before. I can say that corporations do that to actually minimize things like corporate espionage and other issues related to disgruntled employees as part of that are impacted by the layoffs. It's unfortunate it doesn't feel great, and I can tell you it doesn't feel great to everyone in the bills business.

Now, when you're escorted off the president's premises and you can't have access to corporate resources, it really creates a stressful situation for everyone. Now, starting late Thursday and into Friday, several Twitter employees impacted by the layoffs took so to to social media to talk about it. And one, once the email actually came out, those postings increased by a large factor. There was a large spectrum of feelings from gratitude to actually disgruntled. In fact, some may be looking for legal retribution after the event. Now, people were not impact, those were that not impacted by all of this. Also received an email from corporate thanking them for their service and letting them know that the change was coming to the corporate, uh, processes and Twitter itself, and that the buildings would be closed until Florida notice. On a personal note, Twitter was not making money. In fact, they were losing money. Deep cuts. So early on, on with a new regime doesn't look like tactical. It looks desperate. Now, reviewing all the business users at a large platform in a service company like Twitter takes time and it takes care. Personally, I don't think they gave enough time and attention to the process and the people. I feel that people were people who that were impacted today. I feel for them and I wish them good luck. And the message to them is that this is not your end. This is just your beginning.

Curt Franklin (00:08:02):
Well, just in case you're someone who still has romantic ideas about the sort of people engaged in cyber crime, I bring you this story from dark reading based on reporting by the LA Times, showing just what fine, ideal driven pirates. These folks really are. Los Angeles Times purports that Cambodia suffering economically because of the pandemic, has allowed Chinese mobsters to set up huge cyber crime operations using up to 100,000 people trafficked from across Asia. And they're doing it without consequence because of the revenue it generates for the country. The workers are lured by promises of good jobs, but when they arrive, their passports were seized and they're put to work in modern day sweat shops running cyber crime campaigns. The Cambodian government admits to the a hundred thousand workers, but says that they're all there voluntarily to earn good livings in legitimate technology. Meanwhile, stories from those rescued from the criminal organizations include tales of beatings and torture for failing to meet quotas and of being sold and passed around from gang to gang punishments for failing to meet quotas in cyber crime include such lovely actions as forced pushups and squats being tased beaten, deprived of food locked up in dark rooms, or worse.

Now, international law enforcement is working to put an end to this with the Cambodian government's cooperation with the criminals makes it tough. In the meantime, think about someone trapped in this scheme. The next time you're inclined towards lenient attitudes for those convicted of so-called victimless white collar cyber crime.

Brian Chee (00:09:48):
So while this story started with ours, Technica, um, this has been a really big problem. The bottom line is that cable com subic cable companies hate laying deep ocean cable and the cost skyrockets as it gets deeper. Not to mention repairing a cable that has been broken either from natural causes like underwater earthquakes, underwater volcanoes, or manmade idiots that are anchoring someplace they're not supposed to. Well, the OA cable observatory that I worked on at the University of Hawaii actually used to be the at and t Hall four cable was laid in the late sixties, early seventies. It got picked up by a US Navy cable ship, but we had to do it before the cable got too deep. Cuz as it gets deeper, it gets more expensive. And the US Navy wasn't willing to give us carte blanche. Well ship anchors dragging across cables as they reach shallow waters is the main reason, or yeah, one of the biggest reasons why marine authorities the world over have large areas of shallow waters marked as a no anchor zone.

So as waters get shallower, so does the possibility of a ship dropping anchor someplace that could hook cables at various shallow water choke points like in the Mediterranean, just off Egypt. Well, the article goes on to highlight that upwards of a hundred cables annually get damaged or cut by various problems like anchors, underwater avalanches, or purposely cut by underwater demolition teams with the bad guys. Oh, by the way, this is also why cable landing sites are rarely marked on maps to make it harder for people to play, you know, bad games on cutting cables on purpose, or threatening to say, blow it up or drop something on it, you know, like an acme safe, uh, to destroy key cables and trying to blackmail people. Well anyway, so some ask, oh, why don't you just use satellites? Uh, you can't. Those are harder to kill. Well, I ask, what kind of delay do you get when your signal has to go 10,000 miles for a round trip to orbit versus less than a thousand miles on a terrestrial cable? Remember, a latency matters. The point is that a huge number of our world's commercial transits underwater for underwater cables, and the RS technical folks mentioned that a full 17% of the world's current underwater trans cables transit, the shallow waters off the Egyptian coast

Louis Maresca (00:12:48):
Supply chain attacks have been in the news a lot over the last several years, and more so of the infamous log for Jay's set of attacks. Now, these type of attacks are complex, they're hard to follow and have brought impact across the industry. Threat actors are using it as a way to add ambiguity and complexity where organizations are weak in their processes. Now, according to this bleeping article, the Bleeding Computer article, another threat is eminent threat actors are using the compromised infrastructure of an undisclosed media company to deploy the SOC Ish's JavaScript malware framework, also known as fake updates on the websites of hundreds of newspapers across the us. Now, the threat actor behind this supply chain attack tracked by Proofpoint as TA 5 69 as injected malicious code into benign JavaScript files that it's loaded by the news outlets websites and the BBL malicious JavaScript file is used to install ish, which will infect those who visit the compromised websites with malware payloads camouflage as ache browser updates delivered as zip archives.

So you can get an example of, or Firefox to update with. They actually are alerted of these updates. Now, in total, the malware has been installed on sites belonging to more than 250 US news outlets, some of them being major ones according to the security researcher. Now, what does this mean for your organization? Well, first, ensure that the updates being installed are regulated by policies that you put down, and only updates that are approved should get processed. Good, good, good rule. Follow there. If it's not possible, your first line of defense is user education That will ensure that people who can go and identify weird behavior like this when they visit new sites, don't install the updates. It also means that supply chain attacks will continue to be a threat until we find a way to cut the attack factor off of the knees.

Now, let's hope recent moves by organizations to be more strict on consuming updates and dependencies will at least slow down the threat a little bit. Well, folks, that does it for the blips. Next up we have the News bites. But before we get to the News bites, we do have to thank a really great sponsor of this week, enterprise Tech, and that's on logic. Now, you're probably more than just a few feet away from a personal computing device that's really changed your life. I know I am right behind me here, that there's an entire hidden world of computing that's revolutionizing sustainable agriculture, bringing smart cities to life, and increasing manufacturing efficiency to improve the quality of our lives. And that's what you'll find on Logic's. Distinctive orange industrial and embedded computers. Non logic is the first choice in industrial computing for innovators worldwide who need computing power that can survive and thrive where traditional hardware might actually fail.

Now on logic designs and creates computing solutions that could fit in the palm of your hand on powering everything from advanced robotics and AI to manufacturing automation, digital media solutions and smart agricultural technologies engineered specifically for reliability. Many on logic computers are passively cooled and ventless protecting internal components from dust and other airborne contaminants. Now the removal of the fans also results in a completely solid state device to protect against shock and vibration. Now, other design features and specialized components protect systems from extreme temperatures and interference as well. And on Logic Systems are extensively tested to operate reliably wherever they're needed, whether they're monitoring a remote mining operation or riding on medical cart at a hospital. The team at On Logic truly cares about creating right fit solutions, tailored specifically to solve your unique technology challenges Now on Logic Partners with leading software companies like AWS to enable rapid evaluation and deployment of edge computing solutions.

Now, for example, on Logic's line of aws iot, Greengrass compatible computers has been vetted by AWS bringing you piece of mine that it will work right out of the box. Now, if you need a computing solution that can be easily configured to your particular needs, supported by industry experts who are just a phone call away, website chat or email away, and it's delivered quickly, the team at On Logic is ready to help you get started and learn more about on Logic's 30 day risk free hardware trial visit on That's on And we thank On Logic for their support of this week and enterprise Tech. Well folks, it's now time for the bites. Now, proprietary IO and Power Ports have been the name of the game for a long time for device manufacturers. However, you, you may have been, uh, seeing the broad impact that the, that the market has had, uh, I'm sorry that Apple's had on the market and most of the eyes are on them.

Apple has had their proprietary lightning connector port for let's say over 10 years now, and they've made a lot of profit from it, from, from those custom specs that are out there and cables that they're selling. Although organizations and people grumble about it, you know, they're, they buy it anyways, <laugh>, they don't, don't push back. Well until now, European Union strikes again and this time they're requiring that specific devices have a more standardized port like usbc. As our Technica article points it out, the EU put the final approval into the common charger law that will require all devices within specific categories, smartphones included to standardize on us b c. The most devices have already gone that way, including apple's, MacBook, and iPad lines. But the iPhone line is stuck with Apple's proprietary lightning connector. My guess is because they don't wanna go and retool all of their manufacturing processes for the phone, especially this part of the, uh, how the economy is doing and especially all the costs of, uh, of updating their, their manufacturing plants.

Not only does that open the door to more options for organizations, consumers in the eu, some believe that this will actually be less e-waste because of it. Apple doesn't seem to agree here. Now my point of view is that they go to usbc. It actually opens the door to a number of large number of third party solutions, uh, for charging and for data, which could actually can cause some unforeseen consequences when supporting the devices. However, standards are good and ensuring that consumers and organizations have options allows for broader ecosystems. And it also brings down the cost. You wanna bring my cohost in here because I think, I'm sure that they have several different points of view here. What do you guys think? Is the decision here gonna impact the market in any way? What do you think critters

Curt Franklin (00:19:25):
Well, in terms of moving the needle on sales one way or the other? I don't think so. Uh, it, it might boost sales of, um, uh, lightning to USB transition cables, but, but that's about it. Um, I'm not a fan of highly prescriptive legislation and rule making. Uh, what, what if the EU had existed and had mandated that we all do DB 25 for our charging ports, we would now be dealing with phones that are the size of cigarette cases. Um, I, I just don't know that it's a good thing to do. Our standards good. Sure. Uh, should standards be mandatory? I'm not convinced, at least not at this level. Um, so I, I don't see that it has any tremendously harmful effects, but I'm not a big fan of going about it this way.

Louis Maresca (00:20:26):
Right. What about you cheaper? It seems strange that Apple will have to go retool themselves to do all this work. You think they'll just do it for the EU and, and restrict it only there and still sell the remaining backlog of inventory of lightning stuff to the rest of the world?

Brian Chee (00:20:42):
Uh, let's, let's take a look at that first statement, retooling. Now, I don't know if you know if you've ever been involved with a device that has a radio in it. Getting it FCC certified or EU certified, uh, is difficult. And anytime you change any perforations of your shielded case, it's almost black magic to retune the RF envelope. Um, someone near and dear to our good friend Padre actually works on the Apple RF envelope. And he said sometimes just moving a hole in an RF cage just a fraction of an inch, you know, we're talking about moving it sub millimeters will be all that's necessary to get it to pass or fail certain tests. Now, as far as some of the other comments in this article E-waste, well if all of a sudden the, the most popular phone on earth, and I believe it still is, changes its interface, that means all the lightning devices that people have, and a lot of us have, you know, I've got a ton of them are now becoming e-waste.

It's gonna be a flood. You know, maybe in the long run a USBC would be great. Yeah, there you go. Um, I happen to really like usbc, but the lightning's been around. It is a herma, um, cable, meaning you can plug it in either way, just like the usbc I think my only, uh, complaint about the lightning is it doesn't have enough ground lines. So powering high powered devices is actually kind of hard. So USBC is better there. But let, let's tell a really short version of this story. Mandating an actual interface has a lot of interesting ramifications. Having been with the United States government and working in some interesting environments, there were lots of cases where the government mandated certain things. Let's put it this way, we still use paper tape on some devices in the, in the classified world. And that's because of someone mandating that paper tape must be supported and it is still supported 30 years later. Is that what's gonna happen to usbc? I say it's great to mandate that we go to a common interface someday, but I think the EU is being silly in trying to name usbc cuz maybe someday soon there's gonna be a US B, D or E or F and if they've mandated s bbc, the rest of the world will just wave at the European Union and say, nanny, nanny boo boo. Right?

Louis Maresca (00:24:11):
Right. The US before is coming along. I know, I noticed, I've noticed a lot of, uh, smaller companies outta Taiwan and so on building out these solutions. So we'll see. Let's see how this evolves. I think that one puts that one to bed. Thank you guys. I wanna talk about the next one cause this next one's pretty interesting as well. Starling is known for satellite internet for people and organizations, but drones, what's going on there? Cheaper.

Brian Chee (00:24:34):
All right, let's tell you another story. <laugh>, there's been lots and lots of things written about the United States ums. Um, reconnaissance drone will actually drone fleet that they're flown out of the desert in the middle of the United States, and they b go over a satellite into the drone. And one of the challenges is the delay between when you flip your wrist on the joystick and the Air Airlines on the drone actually move. Well, and that's with dedicated satellites that are custom built just for the military, just to fly drones. There might be other things on there, but we're talking about billions, maybe trillions of dollars spent on the US drone fleet way, way out of the price range for the commercial world. Cause gee, sometimes you need to do surveys in the deep desert or over the Arctic or in the middle of the ocean over, say, a new island coming up where there are no cellular connections, there are no terrestrial ways of connecting to some sort of radio.

But starlink opens up a really interesting possibility. So while it's probably not gonna be that, you know, little consumer quadcopter that you see there, it might look something more like that quadcopter where it will have some specialized radios on it so it can take advantage of the ultra low latency of starlink in Leo low earth orbit, and allows you to go and with relatively low delay, be able to remote view. Because, you know, there's a lot of times where a drone will just be mowing the lawn doing GPS waypoints, and you might go right past something really, really interesting, maybe that big gold deposit in the Arctic. Well, if someone's actually flying it and looking at the instrument, they can react in real time and go back and go, wow, that's really cool. I I we would've missed that and we would've lost all those advantages.

Well, so an Ontario company called R Rs, um, had a, they basically had a successful integration in starlink into their drone system, and it allows them to do some really interesting things. Well, one of the main issues facing commercial drone operations is communications outside of urban areas where maintaining sufficient internet connectivity may be challenging according to r rs. To overcome this challenge, the company has been looking at integrating satellite modems into drone system. But according to R rs, the equipment is costly and only supply it limited bandwidth, while starlink, on the other hand, uses its own satellite dishes to receive high speed internet with download rates as highest 350 megabits per second. Now, one of the things they don't mention is the low latency because they're in low aortic, which is the name of the game with drones. Well, lots of interesting things. Now this is gonna be one of those good thing, bad thing, you know, two sides of the coin.

If it's a commercial system and anyone can go and buy it, that means, gee, the entry level for military drones, say for both good guys and bad guys gets a lot lower. And as you can see from the conflict, the, let's actually call it a war in the Ukraine, the Ukrainians are using commercial drones and being able to do some really, really interesting things and really frustrating the Russian Army. Well, if they can fly them using the SpaceX Starling system, gee, that's gonna be interesting. Hmm. So some of the other things, you know, obviously is surveying, um, whether you're looking for, you know, natural resources, you know, gas oil and so forth. Um, I think I'm gonna stop yy and I'm gonna go and toss this over to Kurt and start talking about we've, you and I have both talked to people like oil companies and the cost of surveying is astronomical. What do you think, you know, having real time control of drones and real time surveying of really remote areas, I gotta imagine corporate America or corporate world is really interested in this. What do you think?

Curt Franklin (00:29:44):
I think the world, the corporate world will certainly be interested because this should allow for a much better surveying of natural resources. Whether we're talking about timber, uh, water, doing, uh, various research, looking for rare earth minerals, um, all kinds of things. So no question that there's gonna be a great deal of interest from the corporate world. I think there will also be a great deal of, um, interest, as you pointed out from the research world, um, everything from, uh, natural resources to archeology. Uh, imagine being able to fly lidar, uh, emitting drones, uh, over areas that have never been fully surveyed before. Now there are lots of good reasons for this. Now are there some potential misuses? Sure, there are. Uh, but I think that on the whole, the real possibilities of technology like this, uh, far exceed any dangers that we might see.

You know, it's only been 15 years or so that we were able to say that less than half the people on Earth had made a phone call, you know, much less been able to get access to the internet. Uh, and we're seeing huge leaps in connectivity and what's possible, not only in areas where the infrastructure is currently well developed, but in, in areas where the infrastructure is going to be space based. And that has the potential to change an awful lot of things in a big hurry. Uh, I'm excited by this and I think that, uh, if I were engaged in certain types of research, uh, off the beaten track, I would be even more excited.

Brian Chee (00:31:55):
Super cool. So Lou, what do you think, you know, you're, you're kind of in, you know, the DevOps, you know, your feed are in the DevOps world. What do you think, you know, we're gonna be talking to a guest about ai. Gee, what do you think? You think maybe we're gonna be able to start seeing survey automated surveying that has an AI driven driving this to go and look for things and then, so that way you don't have people going crossey staring at a screen for hours on end, you know, what do you think? You think that's farfetched? I don't think so. I think,

Louis Maresca (00:32:38):
You know, know, we already see a lot of Google's recent ai uh, investments on doing things like looking for, you know, watching wildfires or, um, you know, these types of things. I can almost apply these types of, um, uh, these types of, uh, things to, to, to the drone side of things as well, uh, where, you know, you can now have more of a visibility into it by, you know, having devices like drones be able to survey areas for things like wildfires, for things like flooding, for, um, you know, for, uh, better access to oil drilling and oil drilling sites. Um, and we still have, uh, you know, still have connectivity to those, those drones. I think there's, there's, there's a, you know, 1,000,001 applications that can happen here. And I think, you know, by looking at what a lot of companies are already doing in the AI space, um, this will just extend their extremities, their arms out even farther. Cause now they have physical access to those sites to gain even more data, to produce more insights. So I, I, I feel like this is just the tip of the iceberg for that. Um, and we'll, we'll probably see companies like go, like the Googles of the world, the Microsofts of the world doing things like this once they have, uh, that physical access.

Brian Chee (00:33:58):
Awesome. You know, that's just awesome. I'm looking forward to it. I really am. There's all kinds of possibilities when communications gets cheap and ubiquitous. But you know, we're gonna have to see, you know, the star, the star SpaceX. People need to go and play nice with the, uh, you know, the developers. The antennas need to get smaller and a little less like sales, um, for it to work nicely on a drone. And, um, you know, maybe just, maybe we will kick off a brand new industry anyway. We can all dream. But, you know, I think it's time to go to a guest, but don't we have an ad first?

Louis Maresca (00:34:51):
It is indeed is is indeed. Well, yes. Before we get to our guest, we do have to think another great sponsor of This Week in Enterprise Tech. And that's Cisco orchestrated by the experts at C T W. The helpful people at CDW understand that hybrid work continues to evolve and that your organization must evolve with it to succeed. Now, with so many remote collaboration options, you need a strong and consistent network to empower your workforce and keep them together. Consider a Cisco hybrid work solution designed and managed by CDW experts to deliver the same quality network experience to all of your offices, even your satellite ones, connecting your team from pretty much anywhere. Now because Cisco networking keeps things flowing smoothly and securely with embedded security, security compliance, and multifactor authentication that protects collaboration among your spread out team. Now with real time visibility into distributed application security, user and service performance, you get better line of sight into how network is operating and how better to grow your organization.

And Cisco networking levels, the playing field, providing access to flexible, high-end collaborative experiences that create an inclusive work environment. When you need to get more out of your technology, Cisco makes hybrid work possible. CDW makes it powerful. Learn more at And we thank Cisco Orchestra by CDW for their support of this week and enterprise tech. Well folks, my favorite part of the show, we actually get to, we're gonna guess to drop some knowledge on the TW riot today. We have Danny Shayman, he's Machine learning product manager of in Rule. Welcome to the show, Danny.

Danny Shayman (00:36:38):
Hey, good afternoon.

Louis Maresca (00:36:41):
So, you know, before we get into the thick of it, we have a lot of interesting topics here, um, and, and lot of stuff to talk about for our audience. But before we get to that, we do wanna talk about your origin story. Our, our audience loves to hear people's journey through tech. What, what was your journey through tech? What brought you to in real?

Danny Shayman (00:36:58):
Sure. So I probably have a bit of a nonstandard, uh, career path compared to probably many of your guests. Uh, I spent, uh, seven years as a equities trader on the floor of the Chicago Stock Exchange, uh, for going back to school at that point. And, uh, finishing green bachelor's degree, political science. Um, so yeah, uh, preceding that, uh, I grew up, you know, very fortunate to have parents who were both small business owners that used computers. Um, they brought computers into home very early. I had a, a 3 86 that I was, you know, cranking away on a free color monitor. And, uh, you know, grew from there was always deeply engrossed in technology. Um, started cutting my chops, programming, mostly doing like, uh, video game mods and the like as a kid. Um, graduated high school, went off to study computer engineering. And, uh, about a year into I realized that spending my life, uh, writing code in front of a computer was probably not for me.

Uh, left and, uh, pursued a career in karate. Um, was training to be a member of the US National Karate Team at that point and teaching karate. And, uh, ended up stumbling my way into finance from there. Uh, and spent seven years there. Finished my degree up, uh, at Northwestern at mostly through night classes. And, uh, graduated with a degree in political science. And, uh, coming out of that got a position doing technical writing for a machine learning startup called SIM machines. Uh, and they had a, a focus in explainable machine learning. They have, uh, a proprietary similarity search engine and, uh, ended up getting, uh, offered a job as a solution architect with them. Grew from there and then, uh, in rural technology. Acquired some machines about a year and a half ago and joined Enroll.

Louis Maresca (00:39:07):
It's amazing. That's an amazing story. I love that. I love people who are, you know, start out on one side and go to the other side. Uh, you know, and so, so that's amazing. That's an amazing story. Thank you for that. So I wanna talk a little bit about process automation, cuz we talk a lot about this on the show. We talk about robotic process automation, business process automation, and you know, kind of the difference between them. Now we see a lot of, uh, you know, on the side of rpa, the robotic process, salvation, just, you know, building, using machine learning and AI to, to help you with, you know, smaller tasks, repetitive tasks. But business process automation is a lot harder because it requires a lot more data points to go figure out how to automate an entire business process across, you know, multiple different, um, parts of your organization, that kind of thing. What, what, what are you seeing in this space? So how, what is rule doing that is different from the rest?

Danny Shayman (00:40:01):
Sure. So we've really joined three technologies that are usually thought of as being somewhat distinct, which is business process automation, decision automation and machine learning or, and in our case specifically explainable machine learning. And on the business process automation side, you know, as you mentioned, one of the big difficulties is that companies have a vast number of discreet disparate systems that are often not networked with each other. Um, you know, no role in APIs or anything like that. Um, and so business process automation in many ways gives you the, the tool to, to connect those systems to create the interfaces that you need to move between them. Um, create tracking systems to, to measure, uh, behaviors moving within and between systems, um, and, and letting you consolidate, uh, your processes into really like a single system versus having, uh, you know, hey, we're doing customer onboarding. Step one is, I remember which Excel file we keep our, you know, employee account numbers in, so I can go and update that. Like, that's right <laugh>, that's the process. People

Louis Maresca (00:41:20):

Danny Shayman (00:41:20):
That <laugh> nominal amount of organizations and it's just like, no, like you can standardize most of what you do on a regular basis and, and you should. And so we provide tools, uh, to enable companies to do that.

Louis Maresca (00:41:33):
Now it's interesting because we see a lot of organizations, obviously some organizations are moving a lot of their software and services into one place where they have a lot of end to end solutions. Like I see, obviously I work at Microsoft, I see a lot of organizations moving to the M 365 world cuz they can live and have their data all and all their processes all in one spot. And that's kind of the, you know, the marketing way of, of me saying, Hey, that makes it easy for you to build out process automation there. But you know, you, you work with a lot of organizations that don't live in one cloud or don't have data in one spot, or don't have software in one spot. How, how is, how is Nru helping those organizations?

Danny Shayman (00:42:09):
Yeah, so you, you end up with a wide variety of disparate systems. Um, as you were saying, the, there's very much a need to establish connectivity into a central point. So you end up with systems that are basically routers for you, um, and, you know, can bring those, uh, together to consolidate information, to, to bring your data together to make it accessible. There's a, uh, say a relatively new concept called the data lake house that has very much been, been taking off. And that's, you know, the idea that wherever your data lives, you have some mechanism to consolidate it to one place so that you can tend to it and digest it and, and leverage it. Right?

Louis Maresca (00:42:57):
So we talk a lot about, you know, the use of AI and machine learning in this side side of the things. Um, you know, what, what are you guys using as, as like the data points for this and how are you kind of utilizing, uh, ai, ML and AI to, to, to kind of make this easier to, to have better goals for organizations and, and especially in process automation side of things?

Danny Shayman (00:43:21):
Yeah, so we actually provide a, what I call it a data agnostic machine learning tool. Um, so I'm the, I'm the product manager for tool that we call the ML Studio. It is, uh, an application within the enroll ecosystem for build, building, uh, deploying, maintaining, uh, explainable machine learning models. Uh, the data that you wanna drive into an ML use case, uh, oh, sorry. The data that you wanna drive into an ML model is extremely use case specific. Uh, we do a lot of work in fraud detection, for example. Um, we've also got major deployments in, in silica drug discovery. Um, and, you know, obviously the, the data that you'd wanna drive one versus the other is entirely different. Um, so, so it's a, you know, bring your own data and, uh, you know, then we provide the tools for data manipulation and, uh, application from there.

Louis Maresca (00:44:24):
Right now, a lot of organizations, obviously, they, they think about, okay, well this, this company's using machine learning ai, you know, are they using it in a responsible manner? I mean, obviously we, we have, we've heard a lot of things in, in the news around, uh, in fact, I I just read a survey, uh, that came from your, it was a survey commissioned by Forester that talked about this. You know, what are organizations worried about there and how's enroll kind of focusing on that?

Danny Shayman (00:44:50):
Yeah, so machine learning, you have these systems that are self optimizing functions. Basically it's you say, here's a bunch of data, here's the thing that I care about. What if this data is important for reaching this outcome? And then you say, you know, machine learning system, like go wild, learn the, the, you know, figure out the absolute best way to do that. Um, now potential issues with that are, you know, the, the machine learning model sees the data that you provided it as the entire universe. It doesn't know that there's a single person outside of the data set that you've provided. It doesn't know what gravity is. It doesn't know anything about anything about, except for what you're telling it. Now, if what you're telling it is biased in some way, like, you know, almost every human touched process, um, your machine learning model is going to learn that bias and is gonna amplify that bias. Um, um, typically in society, we agree that having bias against certain groups of people is, uh, undesirable. And we've got regulations and laws against that. Um, and also, uh, because it's, you know, negatively perceived, there's major reputational risk, um, in releasing a model that is harmfully biased towards people, especially protected classes.

Louis Maresca (00:46:20):
Right, right. I, you know, we, we were kind of talking about this, uh, before the show and, um, you know, we, we talked a little bit about the Twitter layoffs that are happening. Um, and it seems like, you know, you know, the whole concept of ethical, um, you know, uh, the ethics process of machine learning, uh, and ai, uh, Twitter was directly impacted by this. How, how do you see this kind of going, uh, on that side? Uh, do you see this as a, uh, as a terrible thing?

Danny Shayman (00:46:52):
Yes. Um, well, as, as one takeaway from, uh, from this podcast, if anybody, if any of your listeners walk out of here and they're like, wow, we've got this whole AI practice and we haven't considered the ethics of it at all, and you're like, really need to go and bring an AI ethicist on to team, uh, Twitter did, uh, my understanding is that they, that they fired their entire AI ethics team, which had been probably the best in corporate in, in, you know, in corporate tech. Um, so that's, uh, you know, they're, they're available for work. Um, but I, I think it's a really strong indication on the need for regulation. Um, you know, this follows Google's ethical AI team, implo last year after, uh, they refused to allow a researcher to publish, uh, their work because it was minorly critical. Um, but, you know, I, I think it really speaks to the fact that we, we can't count on these companies that are spending millions to billions of dollars, uh, on AI practice to, uh, to, to police themselves. Um, there's, there's a very real need for governments to step in and, uh, put some, put some guidelines in place.

Louis Maresca (00:48:18):
Right, right. Well, I do wanna bring my cohosts back in cuz they're chomping at the bit here and behind the scenes in the chat. Uh, I wanna bring you guys back in, Curtis bringing it back in first.

Curt Franklin (00:48:31):
Sure. Uh, happy to do so. And one of the things that I'm really interested in is your take on the blueprint for the AI Bill of rights that, uh, the White House has put forward. Um, first of all, do you think we need some sort of AI bill of rights and is that the sort of thing that could be effective if it was all voluntary? Or are we going to need something that has some enforcement teeth behind it?

Danny Shayman (00:49:01):
Yeah, so as you mentioned, the a the AI is really, it's not even the AI Bill of Rights, it was like a blueprint for the AI bill of Rights, right. Um, and that's cause it, it, it doesn't have any real teeth to it right now. It was, you know, a a an executive branch proposed, uh, initiative that, you know, many executive branch agencies are going to set out to follow. There's nothing really holding their feet to the fire. Um, it's got a, a bunch of recommendations that, you know, that they say corporate, you know, corporations should follow as well, of course, without legislation to back it. Uh, no, no real teeth, uh, for it there either. Um, but that said, a, a rights oriented framework that lays out the source of protections that people, uh, should expect, uh, when they interact with technologies like this, uh, provides, you know, it provides a nice framework to see where the, you know, what sorts of legislation could be passed, uh, to have a, a positive effect.

And, uh, I don't think that any of the proposals in the, uh, blueprint were overly, uh, honors and, uh, yeah, I, I I anticipate that we will start more, you know, actual legislation, uh, in this area. Uh, we, we've been seeing that starting in the EU with the AI Act. Uh, we're seeing a little bit in the US like, uh, uh, this city of New York has legislation protecting, uh, or rather mandating, uh, harmful bias audits in machine learning models that are being used for employment decisions. And, uh, yeah, I expect that to continue to grow.

Curt Franklin (00:51:01):
Well, you know, I, I tend to agree with you on all those points, and I'd like to loop back as well to talk about something that you, you touched on. You know, we have seen a number of instances where AI that was based on a limited model had negative repercussions for members of underrepresented groups. Uh, and it seems to me that in addition to that, that if you have limited models, if you base your, your, uh, machine learning or ai uh, processing on models that come from one point of view, you really are limiting yourself into what that system can do for you. Uh, so how important do you think it is that we somehow get more diversity in the teams that are building these models? Uh, because it can, you know, it can be difficult to, for a team to go outside its own, uh, experience in figuring out how to build a more complex and more complete model. How, first of all, do you think it's an issue? And second, is it an issue worth addressing within the industry?

Danny Shayman (00:52:40):
Yeah, so there's, there's a couple pieces there that I'd like to address. Um, the first one is around the, the scale of the models. Um, you know, you mentioned the, the limited point of view of the models, the, the relatively limited data. You know, you can, you can, you're never gonna have a AI model that has a perfect understanding of the world, right? Like there's, you know, you need some deterministic recording of, you know, every model, you know, every atom in the universe. Like it's, you can't do it. Um, there's always gonna be something that's gonna go wrong. Um, and so, you know, what you end up having are these ML models that are, you know, very narrowly optimized and tailored for very specific problems. They're trained on very, you know, relatively narrow data as you mentioned. Um, and that can result in, uh, either cases where your MO model is going to be deployed, uh, in situations where it doesn't have good data, because that's relatively limited.

Um, and that can be very hard to understand as, uh, the creator of a model, the, the range of use cases down the, you know, down the line to where the model's gonna get used in. Um, and so to, to that extent or, or to, you know, to address that there need to be some very limited, uh, very real limitations that model creators put in place in terms of where and how models can be used to ensure that they're only actually being used for the, the cases that they were trained for. Um, you need observability into how they're being used so that you can ensure that they're not being used, uh, harmfully. Um, but then the other side of that is you need to ensure that you're collecting data that's accurate, that's relevant, um, that's kept up to date. Um, and that can be extremely difficult as well.

Um, you end up, if you look at a, a lot of these like generative language models that have been coming out, uh, the, the burs and the, and the like, uh, if you look at the releases for those, all of those companies say, Hey, we trained this on some massive corpus of data that wasn't a specially clean, and we know that there are harmful biases in this model, and we're leaving it up to you. Like, sorry. And I don't think that's sustainable. I'm surprised that these companies, legal teams are signing off on it. Uh, but that's, that's the current state.

Brian Chee (00:55:23):
Okay. I'd like to get you to speculate a little. One of the conclusions that I, you know, we've kind of come to over the years that we've been watching the no code, low code, um, you know, uh, revolution, I guess you can call it. And the, the trick is a lot of that's being driven because of frustration with the entire DevOps process. You know, we not be able to get it done fast enough. Well, getting AI into the business automation feels like this might be also being driven by some of those same market pressures. And I'd like you to speculate, you know, we're, it feels like we're running headlong into this, what do pe what should people really be doing? What are the milestones? What are the stop signs that people should really consider before they jump off this cliff?

Danny Shayman (00:56:32):
So I'll preface this by saying that in rule is a low code platform. So I, I have a bit of a horse in this race. Um, but yeah, uh, uh, a good chunk of the advantage of machine learning is that you can tell the system to go and optimize on its own. You don't need to sit there and have, and, and write a rules based system to go, you know, and gradually try and improve it over time and spend months building it up only for your, what you're modeling to change over time as well. Um, so from that perspective, machine learning can be a massive time saver, uh, in, in getting effective models out the door. Uh, at the same time we're seeing, and you know, this is again, the, the work that I do. Um, we're, we're seeing, uh, a lot of effort going into reducing the complexity and effort required to actually build a, a machine learning model, um, to again, kind of do this DevOps Enron around needing to actually get developers, or in this case, you know, data scientists involved, uh, in the actual model creation. Um, so, so, you know, our, our intention is to give users, regardless of whether they're, you know, PhD data scientists who we have as users, or people like me who have bachelor's degree in political science, um, the ability to, to build the models that they need for themselves and not need to go and wait for an opening under data science team so that they can, you know, kick off a probably, you know, some, some new internal POC around that.

Brian Chee (00:58:14):
Well, okay, now I'm a small to medium business. I'm thinking this might help my bottom line. What kinds of homework should I be doing? You know, what kinds of things should I be looking up, quantifying, cataloging before we give a company, like enroll a ring?

Danny Shayman (00:58:45):
So I think there's kind of two ways to look at this problem. Um, almost every machine learning project and every machine learning problem, uh, that I've run into has come back to data avail, data availability. Um, and so if you're a company that's thinks that they need to, to get into this game and haven't been there, um, I think step one is to look at the data that they have that they've collected and consider what that could be used, uh, to drive in terms of models. The flip side of that is to say, Hey, I have, you know, uh, a problem with a business problem. I don't have a good way to address it. Machine learning seems like it should be, you know, machine learning or automation, uh, seem like they could address this for me and then, you know, sit down and work out what data they need to to drive that. Um, but, but to me it really boils down to data availability there.

Louis Maresca (00:59:53):
Well, it's amazing how time flies when you're having fun. Danny, it's great to have you on the show. We're running a little low on time, but we do wanna maybe give the folks a home some more information about in real work, can they go to learn more and maybe how their organization get, get started,

Danny Shayman (01:00:07):
Uh, in Absolutely. Uh, if you need, uh, to automate your business, if you need to make better decisions, better decisions faster, uh, give us a ring.

Louis Maresca (01:00:20):
Thanks again for being here. Well, folks, you've done it again, you sat through another hour, the best enterprise and IT podcast in the universe, definitely tune your pod catcher to twit. I wanna thank everyone who makes this show possible, especially to my coho, starting with their own Mr. Brian Chi cheaper. What's going on for you in the coming weeks? Where can people find you and where can people find you at Makerfair?

Brian Chee (01:00:45):
Well, I'm actually gonna be running around like a chicken without a head maker fair fixing things. I just found out that, uh, one of the, um, new policies of the fairgrounds is to run a captive portal to get you to go and say yes, or we agreed to, you know, such and such, you know, terms, it breaks iot left and right. So we're work gotta work on turning that off, at least for the fair. But, you know, I, I'm gonna throw out some, you know, I like complaining. You know, there's all kinds of things I talk about that I've run into, and I like to share my experiences and a lot of the sharing I do on Twitter. Uh, for Twitter, I am a D V N E T l A advanced Net Lab, and we'd love to hear your ideas. We'd love to hear your comments. Um, you could hear me ranting, sometimes I rant at my friends. You know, lots of fun. You, you hear me ranting at Padre every once in a while, and even though Padre's in Rome, but you know what, you've, you the viewer have sent me all kinds of interesting ideas. Some of the ideas have come via email. I'm sheer spelled C H E E B e R You're also welcome to send email to TWiT at TWiT tv and that'll hit all the hosts. Love to hear your ideas, love to hear your comments.

Louis Maresca (01:02:14):
Thank you cheaper for being here. Well, we also have to thank our very own Mr. Curtis Franklin, Curtis work. What's going on for you in the coming week? Where could people find you and where could people find you at Maker Fair?

Curt Franklin (01:02:28):
Well, after I get finished with things like steam roller printing and welcoming people to the Maker Effects booth, that's the maker space and the, um, foundation that produce Maker fair, uh, I get back to work next week, uh, working on various research into the human side of security, starting to look at, uh, some risk issues in quantifying that risk and addition, I've got some research going on, on the use of AI in security, uh, and, and a bunch of other things. So follow me on Twitter at KG four gwa, uh, and come on over to dark where you'll find, uh, the things that I write and the things that my colleagues write. You know, one of the things we're very proud of at, uh, the and cybersecurity practice is that we put together an incredible team during a pandemic. Uh, not at all easy to do, but, uh, we've got a whale of a group of individuals over there. So, uh, please come on over and check us out.

Louis Maresca (01:03:41):
Thank you, Curtis, for being here. Well, folks, we also have to thank you as well. You're the person who drops in each every week to watch and to listen to our show, to get your enterprise goodness. And we wanna make it easy for you to watch and listen and catch up on your enterprise and IT news. Go to our show page right now. That's right, Twitter tweet twitter <laugh> There you'll find all the amazing back episodes, the show notes, goho information, guest information, of course, the links of the stories that we do during the show, but more importantly, next to those videos, you'll get those helpful. Subscribe and download the links that's right, right there. Show I get your audio version, your video version of your choice. Listen on any one of your podcast applications or any one of your devices cuz we're on all of 'em.

So definitely check it out and subscribe and support the show. Plus you may have also heard that's right, we also have Club TWiT as well. It's a member's only ad free podcast service with a bonus TWiT plus feed that you can't get anywhere else. And it's only $7 a month. And then you'll get a lot of great things about, you know, there's lots of great things that come with Club twit. One of them is the exclusive access to the members only Discord server chat with the host, the producers. You can have separate side discussions and all the great channels that are on there. Plus they also have special events. Lots of fun stuff there. So definitely Chuck check out Club twi, join Club Twi, be part of the movement. Go to TWIT tv slash club TWiT. Now, club Twit also offers corporate group plans as well.

It's a great way for you to give access to, to our ad Free Tech podcast. And the plans start with five members at a discounted rate of $6 each per month. You can add as many seats as you like for your organization and this is a great way for your IT department, your developers, your sales team, your tech teams to stay up to date with access to all of our podcasts. And just like the regular membership, you can get access to the Discord server as well and get that TWiT plus bonus feed as well. So definitely check out club twi, TWiT. And after you subscribe, please impress your friend friends, your family members, and your coworkers with the gift of Twit. Cuz we talk a lot about some fun tech topics on this show and I guarantee they will find it fun and interesting as well.

So definitely have them subscribe and support the show as well. Holiday time. So definitely give the gift of Twight. And after you subscribe and you're available, 1:30 PM Pacific Time. On Fridays we do this show He check that out. Come see how the pizza vein, the all the behind the scenes, all the fun stuff, all the banter that we do before and after the show as well. So come check out the live But of course you gotta jump into the chatroom live as well. IRC Twitter tv, we have all the returning characters in there as well as a lot of new characters each and every week. We get a lot of tech topics in there, lots of great questions, just really great people in there. So definitely check out the chatroom as well if you can watch the show Live Now I want you to also hit me up on slash lu.

Man. There I post, post all my enterprise tidbits. Also, you can hit me up on LinkedIn. At LinkedIn at Louis Moka. I'm the, I'm the Louis, only Lewis Moka on there other than my dad. But you know, hey, you know, he doesn't post at all. <laugh> and of course direct message me, whether it's on Twitter or LinkedIn. LinkedIn and let, let's have some good, uh, conversations and, you know, gimme some show ideas, gimme some, uh, you know, topics to talk about on the show. Uh, maybe just talk about tech in general. So definitely hit me up there. If you want to know what I do my during my normal work week at Microsoft, definitely check out There it is. It shows you the latest and greatest ways for you to customize your office experience to make it more productive for you, for you. Whether you wanna create macros for your Excel documents that are cross platform, that run in power automate.

You can do that with, with the latest and greatest office scripts. That's on there. A lot of great stuff. So definitely check it out to help your organization be more productive. I also wanna thank everyone who makes this show possible, especially to Leo and Lisa. They continue to support this weekend at Enterprise Tech each and every week. And we couldn't do the show without them. So thank you for all their support over the years. Also, thank you to all the engineers and staff at twit. We definitely couldn't do the show without them. Of course. Thank you to Mr. Brian Chi. He's not only our co-host but he's also our Kyles producer as well. He does all the bookings and the plannings for the show and we really could do the show without show without him. So thank you cheaper for all your support over the years. Of course, before we sign out, sign out. We also have to thank the editor that's gonna edit the show after the fact. You're gonna remove all of my mistakes. So thank you for all your support and plus I thank our editor, our TD for today, our a technical director, Mr. An Pruitt. He's the talented Aunt Pruitt. He does an amazing show called Hands on Photography, which I watch each and every week. And you know, I'm always curious what's on the next episode, what's going on this, uh, this week and hands on photography an

Ant Pruitt  (01:08:29):
Hey Mr. Lou, thank you for the support, sir. Uh, this week I dive into a question that I've gotten a gazillion times over the last couple of weeks and that's photo restoration cuz I think people trying to get some gifts ready for the holiday season. And I'll be honest with you, I tried to talk people out of doing it, <laugh>, it's a lot, it's a lot of hard work and it ain't for the faint of heart, but check it out TWiT tv slash h for hands on photography.

Louis Maresca (01:09:00):
That's amazing. Well, you know, you know, restoration's also great for a lot of old photos too, so I'm, I'm definitely interested in that. So I will definitely check that out. Thank you, Ann. Well, until next time, I'm Lewis Mareka, just reminding you wanna know what's going on in the enterprise. Just keep quiet.

Mikah Sargent (01:09:18):
If you are looking for a midweek update on the week's tech news, I gotta tell you, you gotta check out Tech News Weekly. See, it's all kind of built in there with the title. You get to learn about the news in tech that matters. Every Thursday, Jason Howell and I talk to the people making and breaking the tech news, get their insights and their interesting stories. It's a great show to check out TWiT tv slash tnw.

All Transcripts posts