Techstrong Gang – January 30, 2025

Hey everyone. It's been a heck of a week here as we've seen the launch of Deepsea causing sort of a, a Sputnik mania, sweeping the industry, and even, uh, reaching out to the, uh, mother-in-law factor, as I like to say, when people outside our industry are wondering what's happening, we've also seen a lot of other changes including, uh, some really amazing, if you wanna say that, announcements from the US government, uh, shutting down cybersecurity protections among, well, lots and lots of other things. And wouldn't, you know it, we picked this week for AI Field Day.

So if you're wondering why we're all sitting together here in San Jose, we're gonna talk about that a little bit too. You're watching Textron Gang. Hey, everyone, I'm, uh, Steven Foskett, not Alan Shimel.

And I am leading Textron Gang today and tomorrow live from San Jose, California. We are here for AI Field Day, which is a production of the Tech Field Day Business unit of the Futurum Group. We're a sister company to Textron, so you've probably seen me on, uh, Textron Gang in the past, but we are joined today in person live by some of your favorite gang members, as well as a new gangster.

Uh, Gina Rosenthal. Gina is a very good friend of mine, somebody I have admired for a long time. And if you asked me who should we bring onto the Textron gang, I think Gina would've been the top of my list.

So, Gina, welcome to the gang. Thank You. Now, okay, that means I have free reign to just go crazy.

I'm excited. So, so tell us a little bit more about yourself. About myself.

Well, I am Gina Rosenthal, also Minks. Long story. I am living in Austin.

I have a product marketing agency where I do anything product marketing for, uh, technical B2B vendors. But, um, the story of AI is really important to me. I'm very interested in not, so I'm interested in technically how it works.

'cause some of that is pretty amazing. I'm more interested in how it affects people, how it impacts our language, how it impacts society. So I kind of have a different take to it.

Yeah, and that's the reason that I think that you bring such an incredible perspective to these conversations, because time and time again, you're a technologist who brings the human perspective into these things, and we can always count on you doing that at Field Day. We can always count on you doing that on our podcasts. And now, of course, here on the Gang, we are also joined by some familiar faces here on Textron Gang this week.

John Willis. Hey, It's great to be here. Yeah, excited.

Thank you for joining us here in San Jose, as well as, uh, on Textron Gang. Um, and then finally, John Swartz, a very familiar gang member. Hi Everybody.

Welcome to my hometown San Jose. I was born and raised in San Jose, um, and I'm excited to be part of Tech Field Day. I, i have heard a lot about it, and now it's time to experience it.

Excellent. Well, let's, um, let's kick things off before we get into the, the Heady and weighty news. Um, let's kick things off with just an introduction to what's going on and why we're here at AI Field Day.

So, AI Field Day is one of the Tech Field Day events that I produce with the Tech Field Day crew. Now, I've been doing this for 15 years. It's been my full-time job that whole time, and it is such an amazing thing that I've been able to have this be my job.

Essentially, I bring a dozen independent technical people together in, uh, some place, typically, uh, the Bay Area, but sometimes other places. And then companies come to us and present. Uh, we also have Indu industry organizations present.

Uh, we present to each other. We record podcasts. We record round table discussions, et cetera, over a couple of days.

All of these sessions are live streamed on Techstrong tv, along with the Techstrong websites as, as well as of course, the, uh, uh, gestalt it, uh, channels, which is the, the legacy, um, tech Field Day news site that's gonna be folded soon into the Tech strong family of websites. com and on LinkedIn. com, you'll see an archive of, uh, this event stretching back literally to 2009, including video recordings of presentations from companies that well came and went during the time.

Uh, Gina, you've been with me from the very start. Uh, Gina was actually there for Tech Field day number two in 2010. Um, you've seen this thing change.

Uh, tell us, uh, I guess tell our audience a little bit, why would they care about Tech Field Day? What is it for them? I think one of the things that's interesting about Tech Field Day is the, is the way you created it.

You know, at a time when we were all bloggers and we were all just trying to get the news out, and nobody would talk to us because we weren't analysts, we weren't media, you know, we were just bloggers who are those crazy people? Um, but if I'm right, you created it to have a way to get right to the companies so the companies could talk to the bloggers and tell them exactly about their products. But at the same time, you have these independent analysts being able to push back on what the company was saying, which is why I loved it, because I was at a, I was a blogger and a vendor, so I loved hearing those tough questions about how things work.

So if you're looking to figure out, you know, for example, we have VMware today, they're gonna talk to us about their AI offerings, and we'll definitely ask them the hard questions about, you know, how does, uh, how is VMware really doing ai? Or is it really ai? And I, I think that's important to know the reality versus the hype.

Yeah. And, and that really was the goal was to democratize it. I mean, you know, the two thousands and early 2000 tens, that was a, a time of demo, democratization of communication.

Um, you know, certainly the, there's a place for analysts and, and traditional press, uh, Yeah, I was gonna mention that. I got, you know, something, it's interesting when you said Gina, because I'm on the, I was on the opposite side, and I would be at these hermetically sealed events where I would have a controlled message feed into my head, uh, Microsoft, hello Microsoft, and this would go on. Um, there was really no give and take.

There was, there were some questions, but they were parried away usually, or we didn't have access to the executives themselves. And the really, the kind of tough exchanges that you really wanna have, that you only can have one-on-one after the press conference, they never took place, usually in front of everyone. So, exactly.

It was an extreme. So I'm looking forward to the other side and being kind of on the inside. Welcome To the dark side.

Yes. Or enlightened, really for me, perhaps Just for me. Um, you know, I, I started a network.

I created a company in 2014 that did, uh, SDN for Docker and wanted a gentleman on my co-founders was Brent Salisbury. And I knew great, very little about, I knew CIS admin, networking, I didn't know real networking like Inad Duke Event and Ali, and those were my partners. And like Field Date was like this thing, like, wow, all these brilliant networking people.

And over the years, I'd, I'd watch, you know, watch the shows. I watched Jerry Solomon, who's another one of my heroes at one of the earlier network field days. And I think when you called me finally I said, uh, I said, Joe, only.

I said, I feel like I've got my picture on the cover, the rolling snow. Well, honestly, what I said to myself is what took me so long. Yeah, there you go.

Um, I, I've been working with John, uh, longer than Tech Field Day for sure. Um, I think that we met, you know, in the early two thousands nonsense Cloudera stuff, Right? Yeah.

The cloud Oh, wow. Cloud camp and all that. Yeah, cloud camp.

Yeah. Yeah. And, and it was interesting because that was, I think, the dawn of some of the democratization of enterprise tech, uh, coverage and knowledge sharing.

And so much of that has blossomed in the 2020s. Now, to the extent that I think that the majority of people watching this, if you're an IT practitioner, if you're an IT manager, if even if you're a CIO, your first source of information is not through some hermetically steeled channel, whether it's analysts or media. Mm-hmm.

It's through Googling something, watching YouTube videos, listening to podcasts. It's all come to fruition to the point that traditional media and analysts have also gotten on the train. And, you know, we now work for, well, I work for, uh, you know, I brought Tech Field Day into futurum.

The reason I did that is because Futurum, um, is making a bold bet that the whole pay walled analyst thing isn't gonna win. That, that's something I've talked about with. So Daniel, I used to run into Daniel all the time at events.

I would always reach out to him for comments because I, I liked his insight and he told me about this model and how it was evolving, which got me interested in joining eventually. It's something there was that major, major wall between the two sides. And we can do both.

And it will be interesting to see, I mean, as we work together and, and especially, I'm gonna be doing more, uh, stuff with Mitch, for instance, and, uh, who's also Here with us, Who's also gonna be here tomorrow, Olivier Blanchard, who covers devices. We're gonna be doing more of this mixture. And I think it's good.

I mean, it kind of brings out the best in both sides rather than one side preaching to the other. Yeah. So I, I think also we're at this point again, where we have so much, um, talk about an emerging technology, and it's being driven by just extreme stream wish of what the technology is gonna be.

And there's so many technical questions, but the hype is so much that it's very hard to wind it back down and talk about, okay, this is a computer science evolution, so why aren't we talking about this as engineers? Why aren't we talking about how this actually works? And then why aren't we talking about how it's going to impact businesses and ultimately people?

And I think that was happening at the time backwards. We had the social media revolution, which what kicked it all off was the communications revolution. And, um, you needed something to ground us in reality and used this for good.

And, you know, it, it is good to have the press and the, the discipline of journalism and blogging brought together, because then you don't get the crazy, you know, diatribes that, you know, you've disciplined about it. You get that mixture and it's perfect. Yeah, we fell into, I mean, I would fall into the trap and I think my peers would, where we would instantly go to the, uh, Renta quote person.

There was a, the one analyst, yeah. There was like a handful of them that were all, I Think we know some of those people. Yes.

We, uh, yes, that's getting a little, but I, I'll I'll mention one by name. Who has, who's not part of, of our ecosystem. There's a guy at Syracuse named Robert Thompson.

He was the most quoted guy anywhere, but he could come up with a pithy comment about culture. New York Times consistently quoted this guy to the point where it was almost a one man industry, and he, he monetized it. And, and eventually, uh, there was a pullback on over quoting him because it was, in a sense, self-serving bill building into his brand and his space.

And I still think that exists. Uh, I want to think it doesn't exist as much. I think I see new, fresher names, and I also think the, uh, not the analysts, not just the analysts, but the people who cover this stuff have kind of progressed and gotten better at cutting through the bs.

Yeah. If You will. Well, I think the openness, like you talked about the democratization, right?

The same time that was going on there was Gene Kim, and we created what we call the Cambrian explosion of technology, where, you know, you had cloud, you had social media, and you had DevOps all happening around that same time. And from a technology place, there, again, you didn't have to go to the Gartner. People could sort of read the books, the publications, the things we were talking about that were somewhere between Pure Tech and the CIO And, and there.

So there is some pure tech. I, I, we have to kind of move on, I think, to the fair enough, to the important stories of the week. Um, I do wanna mention that, uh, as Gina mentioned, you know, we're gonna hear from VMware.

Uh, they're gonna talk about how they're gonna be privatizing ai, uh, check out Textron Gang tomorrow for a deeper coverage of all of these sessions. Uh, we're also hearing from member, which is a frequent field day company. I'm not sure if you all are familiar with them, but essentially their magic trick is snapshotting and cloning memory.

And they can do this, uh, to migrate live running servers. They can also do this, as you can imagine in AI training situations. They can, um, snap lock snapshot and clone and propagate, uh, running instances of training systems and, um, rollback, if there's a failure, that sort of thing.

It's pretty, pretty cool stuff. Uh, Kaza is essentially a, a service provider to the gods. Um, in terms of, um, really bringing incredible AI solutions to the enterprise.

Uh, they really rocked us at at, at AI Field Day last year at this time. Um, we're also bringing in, uh, industry groups. So we've got ML Commons coming in, uh, this morning.

They're gonna be talking about, um, what they're doing on the client side and with their storage benchmarks. Uh, David Cantor is a great friend of all of us, I think, and, and a great friend of the show. And, um, finally we're gonna hear some more of, uh, from the Futurum Group, which recently did a survey that you may have heard about here on the Gang of CEOs.

So we're gonna get a, a deeper look into some of the data from that survey. So all of that is gonna be live streamed actually, by the time you're watching that. It had been live streamed on Textron TV on Wednesday.

And we're gonna be live streaming again Thursday. All of these sessions will also be streamed again on Textron tv and posted to the Tech Field Day YouTube channel. Uh, let's take a break for a moment, and then we'll get right back in with deep seek and, uh, the undermining of the US government.

com is the leading resource for news analysis and education on challenges facing the cybersecurity industry. com covers all aspects of cybersecurity, including data security, DevSecOps, cloud security, application security, network security, security threats, and more. com has the largest selection of security content featuring breaking news, blog posts, podcasts, and more.

com to learn more. com. Home of Security Bloggers network.

Welcome back, everybody. We are living in a week as, uh, John Willis said to me last week, uh, there may not have been a week that has produced so much impactful tech news as this one. And here we are together in San Jose to discuss it.

The week isn't just about deep seek. A lot of the actions of the incoming US Government Administration are really impactful on tech and on the world, and the future of, well, everything we're doing here, we're gonna talk about that in the third segment for this, for this segment. I wanna push back a little bit on this deep seek mania.

How many of us have heard already that this is a Sputnik moment? This is the moment that the entire world learns about the, the promise of ai, that the whole AI table is upturned, that everything changes? Well, uh, although deep seek did achieve something pretty impressive, I'm not sure it's a Sputnik moment.

So, um, John, let me turn it over to you. Uh, jump in. Is this a Sputnik moment?

I don't think so. I mean, I, what I said to you, I think, yes, Monday was probably in my 40 years of doing this craziness. I've never seen one day, uh, be so explosive and so much hype and reality.

I mean, the glass, the one, the most interesting thing about deep seek narrative is the glass is like maybe a little less than half full. And they a little, I mean, there's some interesting tech that came out discussion. And so the, the geeks of the world are like fascinated and rightfully so, the hype, and I think you expressed, and I think you all can express more about some of the stuff that probably wasn't real about how many GPUs they used, did they, what actually have all that?

So there was like crazy hype. And I think the thing that scared me the most yesterday as I was trying to keep up on LinkedIn is I was watching security experts engage and just tell people to go to an API that's hosted in Kang Zoo, China. Right?

And, and, and, and in the actual, I didn't do it right in, in their, in their license, in their agreement, it said exactly what they were going to do with your data. Uh, and so I think probably when it's all accounted, you know, I, this, there's no scientific data from my point, but there'll be millions of people that expose critical information. Mm-hmm.

And then one last thing, the, the models themself. There's a lot of fraught in the data sources of models. We've learned a lot about how you can create inference injections, um, analytical exploits through the data itself.

So I think even the ones that were a little cautious that said, oh, just don't do the API, but go ahead and load the model. And by the way, you know, who had mo models that were not built from scratch on Monday, Amazon, the deep seek models they were running in Bedrock and Amazon and, and Google. And, and again, probably safer than going out to an API, but like, I wanted to just hear from the experts over the next couple of days, the people that really understand how to dissect this.

And so I, yesterday was just crazy. And again, it was half full because I do believe the open source and, and what was contributed and what we learned of how you can do it with somewhat limited resources was an amazing opportunity for compute network and storage. Yeah.

And I, you know, yeah, I was gonna say, sorry, Steven. There, there's kind this demystification of deep seek that's been going on, especially in the last 1224 hours. So there's a report out from Bloomberg that Microsoft and Open AI are investigating whether data output from open AI's technology was, was illegally obtained by a group linked to deep seeq.

Um, Microsoft's been looking at this since the fall. So they believe these individuals linked to deeps seek, and I have to be precise about this, we're exfiltrating a large amount of data using open AI's API. So there's that.

My, my Mike, can we just pause on that for a second? I know, I know, I know, I know where you're going. Isn't that hilarious?

How dare they, okay, go ahead. Okay, Here's the next, here's the next data point, because then again, this is a little bit of self preservation by Silicon Valley, right out of fear, you know, this, you know, they're kind of pushing back, but there's also a reality set because as we all know, hype goes from one extreme to another very quickly, especially in the AI age where everything's accelerated. So then we have David Sachs saying he has, or there is substantial evidence that deeps seek used open AI models to train its own.

And then semi analysis said there's proof that deep, deep seek spent more than $500 million on DPU over the course of its history. And then we have around Daniel Newman, our CEO, who, um, brings this point up. And I know John probably wants to push back maybe on this.

He said that when did we decide that we're going to just believe a paper that comes outta China? He thinks that it's not unlike China to potentially try and play a little bit of ops with Americans and the markets to see how we would react. So there's all this going on.

Um, it's, it's, it was inevitable. And also when we initially, when I heard about deeps seek, it sounded too fantastical to me. And it just is my, my sixth sense going off.

So I'm expecting to, to hear more of this. Now. I'm not dismissing completely what they did, as John said, I think they've built models that potentially are fantastic.

So, uh, the Tech is real, you know? Yes. I mean, you know, just to quickly, I mean that, we talked about this Steve last night, you know, just quantization, we've talked on text about quantization going from 32 bit floating points.

Like there's this laziness that happened in sort of general gen ai, which was like, we have all GPUs. Let's choose just using strategic or analytically when you use like eight bit floating point versus eight, that makes a world, there's papers before dc. Uh, the other thing I wanted to go to your point, which was really interesting, there's an article just today about somebody who created a virus because anthropic and, and Oh, nice.

Probably open AI have been ignoring robots. Text and robot text. Text is the description that you're supposed to honor to not Right.

To, to, uh, um, gather the data or, you know, so like, they, like, you're right. It's, there's so much hypocrisy here. We, The other thing too is a lot of these companies are pushing back because these same companies have been boasting about how much money they're spending and they're throwing into ai.

So Microsoft has a corporate blog where they talk about we spent 80 billion in hindsight, that probably wasn't the smartest thing to do. Yeah. Meanwhile, there's a story outta the New York Times that met us gonna spend 65 billion.

And I, of course, there's the big story that, uh, supposedly the Stargate thing is gonna be 500 billion hundred Billion. Oh, we talked about that. That's, that's all smoke and mirrors, by the way.

That's one thing I totally agree with Elon Musk about. There's very little money there. Yeah.

And, and, um, that's, that's kind of a counterpart that it was interesting to have the bookends of Stargate extravagant spending data center expansion. That's right. Yeah, totally.

6 million. We can do it all together, But, and let me, let me just, uh, inject a little bit of facts into the deep seek discussion. Sure.

Okay. And so first off, um, I wanna call attention to everybody's, uh, so Eckery, uh, Ben Thompson wrote an incredible FAQ about deeps seek, the tech is real. Um, the, the group seems genuinely, uh, honest about what they did.

And if you read their paper, which is not a scientific paper, but is a paper describing what they did, they're pretty detailed about exactly how they achieved this. 576 million include only the official training of the one model deep CP three, excluding the costs associated with prior research and experiments and architectures and data. In other words, the magic that they're claiming to have achieved is essentially being able to train this model on, uh, H 800, which are the export limited GPUs with limited memory bandwidth, and only a certain number of them, and the amount of time that it took them.

And they go into great detail about how they achieved this. They, uh, it's a mixture of experts model, which means that it's a combination of lots of small models that make a big one, uh, that's gonna take less time to train. It has less parameters.

It's also, as you said, a quantized model. Um, and it was also based on previous work that the company that they had done, uh, over the last year. And of course, it's also also based on every bit of work that's been done in the AI industry to date.

So even if you do take it at face value, it's not quite the amazing, you know, Sputnik moment that people are taking it for. Gina, I Think the other thing we have to start talking about is the data that it was trained on. Number one, I'm looking at a post on LinkedIn from Tim Ger, who is the, the woman who got fired from Google.

The ethicist got fired from Google. Um, but she mentions for it to be open source, we have to know what data it was trained on and evaluated on the code, the model architecture and the model weights, which of course I think everybody's focus on open source is the model eight, talking about the data in that paper, they talk about how that data is synthesized from other generative AI programs that they worked with. Let's also talk about where this is coming from.

This is coming from China, which is controlled by a communist country, which means the, the content and the information is controlled. So already we know that this model doesn't know anything about Tiananmen Square. We don't know what else that is.

So Actually, a friend of mine did ask them ask about T Square, and there was nothing, Nothing. A scenic square in the heart of the city. No, no, we don't, we Don't know anything about This.

We don't know what it is. But that, that's part of the thing where we're talking about a model that does generative ai. Generative AI depends on the data coming in.

And whatever it does from the data coming in, that's what you're allowed to build, build from. Now we might be able to make it, it's, it goes faster. More people be able to use gen ai.

Maybe you can train it with some rag or your own specific things, but what happens if, even if you tell it no, let's teach you about about Tianmen Square, what will happen. But that's the meta, the meta point there you just brought out, which is brilliant. Which is, there's two differences there.

Is their there API? No, they, they're outta box models which are built, or for Ruben Cohen did this on, uh, on the Quinn models. Their body used to the Quinn models.

He wrote some great stuff like a month or two ago. Oh. About all the inference fragility.

And one of the things they found is the more constraints, you know, like them blocking actually makes the model weaker in general, the more constraints you build in, and yeah. Those constraints are built into the data sources that are trained. The thing that is glass half full is if you can dissect all the technology it takes to build those, use your own synthetic engines to build your own data without the constraints, then you might have something really interesting that I think that's the, and look like taking their models at face value and throwing up in a SaaS, you're going to have all these inference issues, right.

Which are gonna create constraints. Constraints could create constraints. But the interesting thing about what happened in this paper is if this, all these technologies pan out, play out the way they described, and you build your own models using that technology, that is the thing I think that scared the heck outta Wall Street.

Hmm. May add a may I, I add a cynical note. Um, this is the week where Microsoft and Meta are gonna announce their results.

Oh. And I can Oh, bet your bottom dollar that they are gonna be inundated with questions about. Sure.

No, you've spent X amount of money. Like, hey, meta, you, you, yeah. You pursued the Metaverse and that kind of exploded on you.

Now you're doing this again, and how do you explain this, this company Deep seek and where's your monetization plan? The, the pressure has really been intensi intensifying the last couple of quarters among the analysts for any type of proof of a return on investment. And so again, we're, it's, it's, it's a very impatient culture, especially Wall Street, which is an unrealistic culture.

Totally. Yes. And, um, this just raises the andante and create, creates craziness.

So, yeah. And let's, let's wrap this up. One, one point that I, that I, that I want to add onto that is, ironically Meta and Microsoft are two of the companies that are best positioned to benefit from the deep seek moment.

Because if there's any validity at all into Deep See's technical achievements, and I think there is Yeah, I agree. Then Microsoft and Meta are about to get bigger, better models that perform much, much better on much lower hardware with much lower resources in terms of, of operating resource. Mm-hmm.

But also in terms of capital resources and, um, that's how AI has proceeded to date. And that is how AI will continue to proceed as Satya Nadella said, Jevons Paradox. Right.

Exactly. I was just gonna say, Evans, as uh, AI gets more efficient and accessible, we'll see more use of it. And that's what's gonna happen here.

Microsoft and Meta are gonna have AI everywhere, and thanks to this, uh, development that came from Deep Seq, but also from literally the entire AI industry, 'cause everything Deep Seek is doing is something that researchers have been working on for a while. They just did it. And that's gonna benefit.

It's Gonna increase demands. That's what, And that's terrifying because go back to the data, and this is gonna come from closed systems. How do we know that that data is appropriate for our, is fair to the whole world?

Mm-hmm. And, you know, it's, it's, it's a big question. We are going to be talking about deep seek, the deep seek moment for a while.

I, and I think that what this will do is it will only accelerate ai. It will not hold it back. So, uh, let's move on here for a moment though.

Uh, we're gonna take a quick break. We'll be right back with more text, strong gang, Discover, tech strong group, the epicenter of tech innovation. We are your go-to for reaching IT, leaders and practitioners worldwide.

Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more. Join our satisfied clients.

Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Techron Group. So welcome back to the Textron Gang.

We are now turning the page a little bit on deep seek mania, and we're gonna talk about what's been going on in Washington DC Now, this is not a political show. If you've watched, sometimes it does sound a Little, some People a little political with some people. Um, but it is undeniable that what's been coming out of Washington has been, let's say, earth shaking in terms of the, uh, executive orders that, for example, have paused all dis disbursements of money to basically, well, literally everything including, oops, accidentally Medicaid.

Um, also the dismantling of federal organizations, uh, the, uh, reassignment of, uh, leaders across the Justice Department across most other departments as well. Uh, the loss of institutional knowledge. Boy, this could not have happened at a worse week in terms of the global, uh, impact of all this deep seek nonsense.

Uh, John, you wrote an article about, uh, the changes at the CISA and and so on. Uh, talk to us a little bit about what's going on in cyber security. Yeah.

You know, Monday Morning, you know, I had known about DC in fact, you know me as being terrible at marketing. I probably could use your help. I, I did a podcast with a good friend of mine who is incredibly deep into all things training models, ai.

And we have, uh, two podcasts queued that won will come out next week, where we talked about deep seek the prior week. We talking about the architecture, how it's different oh 1, 0 2. So I, I've been familiar with it.

So, but on Monday warning, when the storm started, I started thinking about what happened the prior week, right? This is before Monday, right? This is even before what happened sort of yesterday, which was, uh, and I don't like to be a chicken little person, but if you think about this per I wrote, could a perfect storm be building in the deep, you know, trying to be clever, right?

The, uh, but like the, you know, CISA has been reasonably dismantled. Jen Ley has, has been pushed out. You could say she resigned or whatever.

Uh, the bigger thing is the CSRB, um, the CSE, it's called the Cyber Safety Review Board. They're the ones that have been tracking salt typhoon. If you know what Salt Typhoon it, it is a notorious group of Chinese actors that are infiltrating our, um, telecom infrastructure, um, at, at, at alarming rates.

That whole board 18 members, experts, some of the top cyber people in the world volunteering at time have been dismantled. Um, you know, there's cuts in, in csa, there's cuts to support of DHS, even in a department. Um, department of State has put a freeze on cyber money going to different diplomats.

You know, like all this happened last week, so you couldn't, I was telling you last night, I like, if I wish I could have wrote a Tom Clancy novel at the end of last year about this like perfect storm of all this stuff happening. It happens in every presidential transformation, right? This, this is not, some of this is uncommon.

Some of this is common. Some of it's uncommon. Yeah.

Um, and then, and then talk about this unbelievable monstrous new AI thing just flooding where millions of people are downloading it. I, I don't think you could, you could have wrote a believable, um, you know, fictional story. Yeah, yeah.

No. You mentioned, it's funny what, there's salt typhoon, and there was another Ty, I, I can't remember the name, but one of the things that gave me, uh, assured me was CISA and Jen Easterly. I mean, that was one of the organizations that I thought, okay, they know what they're doing.

They're working in hand with the government and, and the private enterprise. That, that gives me hope. And then this, Well, you know, The bottom's Gone.

A good friend of mine worked for her and he didn't have great things to say. And I've been tracking her for like the last couple years. I think she's done an incredible Job.

I, I do too. I, too Incredible job. And what they've done in CSA and all that, just gone like, wait a minute, let's freeze.

Let's see what we can learn from them. And, uh, you know, I, I think there's, there's this incredible, um, convergence of a gap that is really scary between the technology onlay of what's happening with AI and what's going on with the cyber. One other thing I'll say, and I, I will shut up.

One of the questions I got mostly yesterday, excuse me. Which was, would you be complaining about this if it came from Switzerland? I'm like, yeah, but that's not the point.

Mm-hmm. It's China and like they, there are some bad actors trying to take out our water supply, our telecom. Right.

And, you know, so like the infrastructure asking me that question as if I'm some bias against, you know, I've been tracking cyber infrastructure for the last two years and it's Scary. You know, I would be channeling my Alan Shimmel, but I won't, I won't do it because, but I gonna say this is something goes beyond tech. It's the institutional knowledge and the disparagement of it, or the dis total disrespect for it, which is going to affect all of us in many different ways, not just through technology.

So I just, I, I, this is just part of a larger problem and I won't go any deeper than that, but I just think it needs to be said. Well, It's not like they didn't tell us what they were gonna do for through use Exactly. That we knew this was coming.

Exactly. But I think the thing maybe pe the reason, say things about, we remember, we call it cyber and people our age thought cyber was something different 30 years ago. Right?

So I will never get ever calling this cyber. And then the, the second thing is, is, is everything is digital. Everything we do is digital.

They're, they've, they're, they've, the water supply, everything is run by computers. And because of that, and because, um, even before now where they've decided not to fund anything, for some reason everything was critically underfunded where they couldn't update things. So it was easy to find things to hack into and, and crack open and, and make disruptions.

I mean, I went to one time at Defcon, I saw the most awesome, um, presentation about how to take over a country. And he had like this video where he took these chainsaws and put 'em together on a drone and you could just go and cut the, the power lines and all of these things you can do. This is not, yeah.

I mean, Years ago, years ago we went to some hearing, uh, a, a meeting of, of government officials and they talked about as asymmetrical warfare and how That's what this is. Were testing how to cut off the water supply in Iraq. This is, This is what it is.

And it is, it's, it's state sponsored. We do it too. We do it too.

Exactly. It's what is done. And so it's not, it's not about we're against China, we're against Asia, but China is a foreign adversary and they are going to do this.

And, and just by the, the, so my friend Josh Corman, who's built this thing called UNR 2027, and he's just getting very simple that that like literally bad actors, and again, probably from China, could take out a water supply in a major city for two weeks. 'cause the infrastructure is sold. And if in two weeks a city doesn't have water, there'll be millions of deaths.

Yeah. Right. Like that.

It's that simple. It doesn't have to be some grand scheme on a big board about how you take out a whole country. Mm-hmm.

You literally just gotta go to Des Moines, Iowa and take out the cyber, the critical 40, 50-year-old infrastructure and, and, and, and dismantle it. I wanna personally apologize to the people of Des Moines, Iowa, who we have just thrown Need to do it. But read, read Josh's.

I also thinking about autonomous vehicles and like how we, we were gonna want set up this, this utopian idea of, of cars moving and flowing and and you can disrupt that. Well, that that's what, that's the thing. I mean, you wanna disrupt the, you wanna disrupt the country.

Most vehicles, there's a story that came out this week as well about, um, Subaru and how connected their cars are even. And you don't think of Subaru as a leading edge technology company with autonomous driving. No, they don't have any of that.

Their cars had a flaw that would allow all the Subarus to be stopped where they are. You want to disrupt a country stop all the Subarus May, and maybe that's, maybe it's Not, again, I apologize to Des Moines de Subaru Dealership De And it doesn't have to be the cars. Yeah.

What if you disrupt the trucks? Yeah. Then you don't have medicine, you don't have food, you don't Have that.

We could go down that road too. Right? Right.

So this is not, and, and it so fine, it's fine that you connect to an API of a new model that was trained in China that doesn't recognize Tianmen Square anymore. Totally, totally fine. Yeah.

So, so here's the, here's the problem. And, and to me, again, without getting too political, you don't have to be political to say, why didn't somebody think this through? That's right.

You don't have to be political to look at, for example, the freezing of all Medicare spend or Medicaid spending and the shutdown of all the Medicaid disbursements in the United States, which literally happened yesterday to say, wait, maybe somebody should have considered the impact. Well, of ordering all funding to be frozen. You know, and you don't have to think about like, I think that the federal government, I think that even the Trump administration would be pro, uh, CS a's goals.

I think that they would be supportive of what CSA is doing. If only they took five minutes to think about it. That's a good thing.

The bros are, are, are in his corner. I mean, this is, this is like the absurdity of it all to me. Yeah, totally.

You know, look at, he's surrounded by these people who should know better, but they're too busy. He doesn't, doesn't care. Well, personally, they're greed, interest care, self-interest Care.

That was my point yesterday. It seemed like the people that should have knew better. Like I, I am not that smart.

I listened to really smart people and I try to grok and I try to teach. I always say they're the unicorns and the horses. I'm a horse that tries to understand unicorns to explain to other horses, right?

And, and yesterday just seemed like the unicorns were failing us. They really were just going off on like these tangents about how great Deep Seq was. And, and I get it.

Everybody was trying to get their social branding out. I'm the first person that's what person. And we all were kind of doing a little of that, but the, I was just having this tug of war with people accusing me of being anti-this or anti open source.

And, and really what I was just saying is, can we all just calm down and figure this thing out before, uh, you know, millions of downloads happen and uh, and it just wasn't happening yesterday. And then you add the fact that we, that was all this stuff that happened last week and even in the, in the, the Medicaid thing got all the press, but there was like a $300 million DHS grant for cyber that was ca canceled as part of yesterday's, that's not even cutting the executive order stuff from last week. And, and it gets worse.

I mean, um, all of the, uh, inspectors general for all of the Yeah, that's crazy. Branches were fired. Um, again, these are people whose job is to rule root out fraud and abuse and overspending and so on.

And, and, and they're all gone too. Well, um, I I'm sorry to do this, but we do have to wrap up. Gina, I'm gonna give you the last word I was gonna talk about the Office of Personnel Management.

Yes. It's just such, uh, it's very much from reports that being run by, um, Elon's Pinch people and, um, the fact that they went in and set up an on-premises exchange server and expected people that they have trained very well not to, to click on links from suspicious looking emails. Uh, like what are they, they're not thinking.

They don't care. This is, is the, this the equivalent of Musk going into a data center and pulling a the plug? Nothing happened.

It's fine. That's right. It's exact same thing.

They, they do not care. They are not going to look at what they're hurting. They're not gonna do it unless enough people push back and protect the people that will be the most vulnerable and most impacted by this.

Because they don't care to know. And that's, they don't care. And that's a problem and need to, but Gina, you said this earlier, we're not acting like engineers.

No. And that's the thing that sort of really is to all these points, we're not acting like engineers. Well, we, we have to wrap up for today.

Uh, thank you for joining us on The Gang today. Gina, uh, welcome to the Crew. Thank you.

Fun. I think that you acquitted yourself. Admir Admirably.

She's, she's, she's gonna be a regular. We're gonna bring in, we're gonna bring you in I Vote Fridays. So we, uh, that's, we are gonna be live streaming, uh, tech Field Day on LinkedIn as well as on the Techstrong websites.

Uh, we're talking about ai, we're learning about ai. We're gonna have some private conversations. We're gonna hear from uh com, you know, organizations like ML Commons.

All of that is gonna be live, uh, Wednesday and Thursday. So by the time you watch this episode, it will have been live all day Wednesday. Go to LinkedIn, go to the Tech Field Day page on LinkedIn and you will be able to watch all of these presentations and our discussions and, and everything from Wednesday right now.

So, so go look at that and then we will be back with another episode of Techron Gang recorded Live right here in San Jose tomorrow morning as well. So keep an eye out for that. We're gonna shift up the guests.

'cause I don't like any of these people. Yeah, we will keep Gina. I'm coming.

There you go. Um, we'll, we'll be back tomorrow with another episode of Textron Gang. Thanks for listening.

Have a great day.