Intel’s Reported SambaNova Play, AI Agent Adoption, and OpenAI’s Cyber Warning | TSG Ep. 989

Hey, everyone. Is Intel trying to spend their way out of their doldrums? You're watching Textron Gang.

Hey everyone. Happy Thursday. You know, it's beginning to look a lot like Christmas.

Um, we've just got a few more shows for Textron Gang. Let me give you a heads up, right? We are doing today's show Thursday.

We will have a fresh show on Friday, like always. And then next Monday, uh, we're going to do something. Is it next Monday or Tuesday show?

I'm asking my producer Tuesday, for next Tuesday, right? That'll be our last fresh show of the year. And we're gonna have a, a super duper extra Cali Federalistic panel that I think we're gonna break up into three different sessions.

We're gonna do one tech strong session on, on ai, because why not one tech strong session on it? Again, why not? And one text drunk session on cyber.

'cause we have to. And, um, so stay tuned for that show. It'll be on next Tuesday, and that's how we'll end things out.

And wish you all a Merry Christmas and happy New Year. We'll be running some other stuff during the week, but it'll probably be prerecorded. Um, anyway, just wanted to share that with you.

I don't know why I felt compelled to do it, but there we go. Um, Mike, well, let's introduce, we've got a great gang here today. Let me introduce you to them.

We've got the lady behind in front of the wall, Terry Robinson. We've got Crackerjack Futurum personality Analyst, professor Dave Nicholson. David, it's good to see you, of course, the one and only John Schwartz also.

Well, he's one and only two. Mitch Ashley and the Bard, Mike Ard. Uh, ladies, gentlemen, welcome.

It's a, it's an interesting day. 6 billion of our hard earned taxpayer money on, uh, I don't know if it's from the taxpayers, but, but, uh, but in, in, in the acquisition mode, Mike, what, what's this about? Well, we've seen this before.

I think it's one of those examples of when you can't beat 'em, you buy 'em, right? Mm-hmm. And so they're looking at this situation and Intel is saying, Hey, why spend all this time trying to build our own AI accelerators, some of which they have tried to build, and the reaction from the marketplace has been, shall we say, mixed.

There's a lot more competition in this space, and I don't think they have a lot of time to figure out what they're gonna do. So it seems like this might be the most logical thing to do. But John, you wrote this story.

You live in the heart of the Valley. What are the folks saying? Yeah, so you're right, Mike.

6 billion. And basically they're plan, they're trying to play catch up in the AI hardware market. Um, and Samba Nova to Intel Lease offers a potential shortcut, uh, without devoting years and billions of dollars to internal development.

So, simply speaking, Intel has been so dominant with traditional computing with CPUs. Meanwhile, Nvidia is captured a dominant share of the AI chip sector, which has now more become critical for training and deploying LLMs and other machine learning applications. So it's a, it's an issue of catch up.

And it's also, uh, interesting issue because the CEO of Intel, lib Bhutan is the executive chairman of Samba Nova's board and Tans venture firm. Walden Internationals an early investor in the startup, which is kind of interesting. Um, it's, it's a, again, it's a, it's a game of catch.

And I think just kinda stepping back, it's also a, a sign that we're gonna be seeing a lot more deals. I mean, this is this, this week alone, I just want to kinda look at the bigger picture. We have Nvidia ServiceNow making acquisitions or about to make acquisitions.

Uh, Coursera was another one announced on when on, uh, Wednesday. Uh, there's a, I think a collective $13 billion of acquisitions here that we're talking about. And, uh, that market's gonna take off.

And, and I think AI is spurring this. And so Intel in sense, was forced to, Intel's an extreme example. You know, they're, they're in a position of struggling and they've gotta do some drastic things very fast.

Um, but on the flip side, I think we're also seeing this AI boom and this necessity to, uh, be on top of things, spurring all this, especially from the big players. So, um, we'll see how it plays out. But in Intel, it's, it's, it's a, I wouldn't say desperate, but it's a, it's an urgent move, Dave.

I'll say desperate. Let's ask Dave. Let's, we haven't had Dave on the show in a while, so let's get Dave in here first.

Dave, what's your take on all this? Is it, is this a smart move or is this gonna be another one of those two little too late kind of things, Two rocks decide to go for a swim, kind of, kind of a Situation? No, so I mean, I, I would say this is gonna be sort of a cultural litmus test, frankly, for Intel.

Uh, I have one word to consider, and that word, it's not a Spanish word, it is not the capital of Cuba, but it is Habana, uh, $2 billion acquisition by Intel. And the, and under previous management, they drove it into the ground. Uh, it was a classic, classic example of a firm that responded to an acquisition.

Like the human body responds to a virus, uh, you know, attack it with antibodies. Um, so I, we're gonna see, let's be open-minded. Let's be hopeful.

Frankly, I think there's a lot of good things will happen, uh, if, uh, if Intel can get this to work. I'm skeptical, but we'll see. It will be a, a big cultural litmus test.

The bigger issue is, I really think that 2026 is gonna be the year of the beginnings, at least, of recognizing that there are alternatives to Nvidia in the AI accelerator space. We're seeing that all over. Uh, we're seeing people be far less shy about presenting what they have as alternatives.

And I say by the end of the year, people will all accept that those 87% margins that Nvidia is getting on its hardware are absolutely unsustainable. So I, I'd give few thoughts on this one too. First of all, there's the incestuous relationship here between Tan Intel and, and the San Nova.

Now, San Nova raised money at a $5 billion valuation couple years ago ago. 6. That is there a reason?

Does someone know something? 6 billion. And, and, and on an inside job, t is saying, let's pick this up because otherwise, who, who knows what the future holds for it and and is an intel.

Yeah, that's desperate as they are, as cash strapped is, you know, losing money as they are gonna go. Is it right? 6 billion on Anso ran company when they're at, to Dave's point, when their track record of of integrating these acquisitions, it leaves a lot of eggs on the floor instead of in the omelet pan.

You know, you, you've gotta ask these things. And what kind of AI chips is this? Because the action isn't in competing on the GPU front.

With Nvidia. The action is in, is on the, uh, is on the inference chips that Broadcom and Amazon and Google and Microsoft, and everyone wants to jump in on that right now. Our own signal, six five labs from Futurum Ryan Trout, he did some tests on those Intel AI chips.

I think it was last year or maybe earlier this year. 'cause we're at the end of the year. And you know what, those Intel AI chips weren't that bad.

If you look at the, at the lab test Ryan ran, are they, are they just buying something? 'cause it's not Intel inside. 6 billion?

And, and I forget the, the brand name for the Intel AI chips that Ryan tested, but you, you could find it on the signal. Which one? I believe so, yes, you could go on the Signal six five site and get that report, by the way.

6 billion into continuing to develop that because this thing wasn't setting the house on fire, that's for sure. 6. That's my take.

You know, those who don't innovate, instigate, instigate an acquisition, then that's what're, that's what we're seeing here. If you can't innovate it quickly enough and that, and that you can't innovate it everything. Right?

But there's also the opportunistic, and I, I don't know if the price here is picking it up, but for a song, you would think that that would be more the case given the drop in valuation of, uh, samba Nova. But I'm, I'm, I don't hear what the synergies are, what the real kind of pickup is by getting these guys. So it, it does cause you to question, is it sort of, to use the rock analogy, putting another rock in your, in your bag to carry with you to say, I've got lots of rocks.

Um, and we'll see which ones work out in the end. Maybe that's the strategy. You know, With all due respect to all our friends who test things in the labs.

I don't know anybody in the AI space who puts an ounce of confidence in any of those tests, because there really don't reflect any of the real world use cases around AI these days. It's always been a problem in it. But in AI it's especially problematic.

And the marketplace has clearly voted. They all tested and they all looked at the Intel gouty chips and said, you know what? We're sticking with Nvidia.

So, you know, I'm just saying, um, you know, Intel has a severe problem here. And it's not because you know, somebody, you know, doesn't understand that the gouty chip is awesome. It's more the simple fact is that for real world workloads, Rena's not made Rena, it rename it.

com days, I, you know, I I helped build a company called Interline. We did 30 acquisitions in 36 months. One of my mentors there, Len Faler, who also mentored my friend Brad, felt Len once told me, you know, if you just put crap on top of crap, you know what?

You got a big pile of crap. And so, and I'm not saying Intel is crap. I'm not saying samba nova's crap, but what I'm saying is, if if samba nova isn't, isn't market, if it's not in the top three, get the hell out.

Right? And if Samba Nova's not that kind of player, all you're doing is piling stuff on top of stuff, you're not building something better. And, and there's a lesson there.

We, in a given intel, the state of intel and how it's floundered for, for God, God knows how long, maybe this was their only option. I mean, in terms of acquisition targets, I mean, they've, when you're, when you're a popper, you don't really have a lot of options. And I'm just wondering if this is almost like a Hail Mary in a certain sense.

I don't wanna be too dark, but I'm wondering if this fails, then what? So what are you telling me Intel's going to target looking on the clearance racks for AI chips? 6 billion cheap way to pick up engineering talent that I need.

Maybe that's what it is. Well, that Maybe it's, maybe they're gonna Thinkign all these people to the gaudy chips. Maybe.

I don't know. It just smacks a desperation. It makes me sad for Intel, you know?

Um, Yeah, I would say this though. Um, you know, so maybe they had a miscue both at Intel and at this new startup, but I'd rather have the people who've already screwed up once come to figure out what the next thing is because the, the other team's just gonna make the same mistakes over again. So there's something that we said.

So you're citing that old mathematical law Mike, of two negative numbers added together, make a positive. Hey Dave, I would say Wharton did Multiply them right to get I'll, I'll, I'll give you one example where, you know, that, where it did work in real life, right? So after Pearl Harbor, when they named Nimitz Admiral, he did not replace all the other admirals.

He kept them all. And then they went on to win the war because, well, they were all p****d off. So there You go.

Mike, Mike, you and Alan, haven't you gotten the memo? There are no lessons to be learned from history. Okay?

I feel like we spent the better part of a decade reinforcing that message. Exactly The last 10 years. That, that this is all, this is brand new.

You're all sounding like AI bears. There is no bubble. 6 billion versus 5 billion.

It's the same reason why size 15 basketball shoes don't fit my feet. Yeah. It's because I have size 10 feet.

Yeah. And this company was never worth $5 billion. The thing that struck me when I, when, from across my desk was that anova, I knew them as a company going back years at, at, at the Supercomputing conferences.

They were based upon this idea that you needed to embrace their entire architecture in order to leverage what they were doing to get value out of it. And, and that was, that's always a tall order. Um, to be able to say, look, if you embrace our ecosystem, forget about Cuda, forget about all this other stuff, you need to buy the entire stack.

And then it makes sense. So I've gotta go and do some research, but I'm a little bit puzzled here exactly what it is, what's the IP that they're really trying to get at? 'cause it's not the whole rack of samba nova stuff.

Um, and so, and that, so that makes it even more complicated. Kinda like a date with the big family. That's what you're saying, right?

Were these the guys that that sort of switched their focus though too? Weren't they on AI training early on? Or am I wrong?

Maybe They did Terry. Yeah. Yeah.

That could be. Well, Like The inference makes sense. They went, they've Been veering from one thing to another To another.

Well, they went from AI training, which really Nvidia at this point is locked up to ai inference. Inference. Yeah.

Which Would seem to make sense, but good luck, right? You're going against tra and, but that and the Google one and the Broadcom ones, and, But that's when their value dropped. I mean, their valuation dropped.

What I can, Yeah. What I can't get past is Tan's relationship. And though that whole dynamic that Alan brought up, that just always makes me, that's Unsettling to me.

Well, this, this smacks of insider wink wink deals. Another thing that we've learned a lot about in the last 10 years, it's been reinforced insiders. Okay, now, Yeah.

You know, I'll end it with good green life. Let me just propose something out at you, not only to our panel, but to everyone out there has Intel become the Cleveland Browns of the chip space. Ooh, ooh.

As a Steeler fan, I had to throw that in. Let's take a break. We're on tech strung gang.

We're gonna come back and talk about agentic AI acceleration. You've Earned it. The spotlight, the responsibility, the weight of teams, companies, and entire industries fall on your shoulders.

Lives depend on your decisions. Your home life included that work. You are protected physically and digitally.

Nothing gets through your team without a fight. But in a globally connected world, everyone sees you, including those who mean to cause you and your organization harm. And now home your sanctuary attackers see an opportunity.

Your digital front door is wide open. And what compromises your home can breach your boardroom. Because the devil's greatest trick isn't targeting your workplace firewall.

It's convincing you that your personal life isn't at risk. Black cloak, digital executive protection, defending the new attack surface your personal life. Hey folks, we're back and we've been talking a lot about AI agents a a lot lately for that matter.

And there's two new reports out, one from Zapier and one from an analyst firm talking about how adoption of AI agents in larger enterprises in particular seems to be happening faster than we might've thought. And there's a lot more of these things in production environments. And Dave, I know you're close to this whole business side of the equation as well as the tech, and you kind of sit between the two.

But is this surprising you in any level? I mean, this is still an emerging technology. There's still a lot of issues, and yet we're pretty far down the path, especially for slow moving enterprises.

Yeah, it, well, what it does is it begs the question, what are, what are they talking about when they say, oh yeah, agentic ai. Yeah, we're doing all that. All of the things, all of the agentic things, all of the agents, we got 'em all.

What does that exactly mean? Because, uh, a few years ago, if you ask someone, um, what are you doing with ai? Or if you were to take a look at, uh, say the students in, uh, one of my programs who are CIOs planning what they're doing with AI in the coming year, um, in the AI space, it's still a lot of machine learning.

So I suspect, frankly, and I know firsthand, at least anecdotally, that often people are referring to the use of agentic ai. It looks a lot like robotic process automation or whatever the heck UI app and, and the rest call it. So there's a fine line between automating tasks and the idea that, um, that you're using agents to initiate these workflows.

And I thought it was the, the key interesting point is contrasting agentic use versus, um, uh, generative ai. That's a, that's a bit of a flag there. 'cause it's like, okay, well hold on a second, really, the really the only real forward part of this is the, is the generative AI use.

And so if there's no generative AI there and it's ag agentic, where is the real autonomy happening? And, and I know the difference, but I'm just telling, I I think that the surveys are not as, as exciting in terms of the, the, you know, this idea that we're really leveraging ag agentic stuff in a way. And I, and I, look, I I do stuff with N eight N on a home system, uh, putting together ag agentic workflows.

And a lot of them look a lot like setting up, uh, um, scripts, frankly, Using AI to write your scripts, right? So, So, yeah. So, so Mitch, to David's point, um, you know, are we kind of roughly having the equivalent here of asking people who are not aficionados about wine and what the difference is between a $300 bottle and a $50 bottle is 'cause they just generally aren't gonna care and they're just gonna say, it's all red wine and thank You great Juice.

Well, I, I think you have to take this as directionally information, because what happens in the large organization is the same thing that happens in market, right? In the market. We hear people talking about, we're doing this, it's real.

Now we're, you know, this is, it's not just coming. It's already happening. And suddenly the market starts to believe it.

Same thing happens inside of an organization. So the p we're not doing a survey of 200 CIOs and asking them specifically about what's happening. We're, we're surveying whoever the people are, are, uh, responding to the survey.

So they're, they're answering as in, yeah, we're doing that. I heard we have an ai, uh, agentic process project, whatever that is. And I heard we're doing generative ai and I know I've got copilot on my computer and I'm supposed to be using it.

So I guess I should say yes. So in other words, it's, it's, it's directionally where you're going. I think it is the way to read this, unless you really curate who the audience is that you're surveying, You know, there, there're like, there's, there's been several studies about the age agent use, age agent AI use, and they run the gamut.

So they, they go across, across the unit. It depends on who's being cold in what industries and may maybe even the, the, uh, amount of knowledge the people who are being pulled have about agen ai. So in a sense, they might be lumping it in with some of the other things or they're projecting.

So what the thing always comes back to me is that this is a very imperfect science. Uh, surveys always are. But this idea of how far has that agen AI gone, I'm not sure it's gone as far as some of the, some of these surveys suggest, but I'm, I'm not, I'm not sure.

It's, it's, it's, it's non-existent. So it's somewhere in between, it, it is growing, but it's not growing as fast as maybe, and I, I wrote one of these stories. I don't think it's growing as fast as some of these folks say, because I think they're, the way they define it has lot to do with the answers and what's actually going on within those companies.

I think the other kind of delve in the details is what are we talking about doing ag genetically? I, you know, to your example, Dave and I could use AI to write a script to say, send me an email in the morning and say hello. It does it automatically.

It's, it's on, it does, its on, on its own is agen. Now that's, that's a kind of an absurd example, but you could call that ex uh, agentic. So how sophisticated are the agents?

How sophisticated are their workflows? They're probably not super sophisticated yet, right? We are climbing up this hill, getting better at it.

Frankly, the technology has to mature. We're just getting longer running agents now that's just coming onto market. That's a pretty new thing.

Um, being able to control and share memory and, and, uh, manage multiple agents is still a big challenge. Still areas for growth and innovation. So, you know, we may be talking about problem sets that, you know, a Zapier could easily handle or maybe push Zapier a little bit.

Um, of course Zapier has their own, you know, AI kind of stuff too. So it it, it's relative to where you are. I think also, Well, and I wanna make sure when I understand this.

So you're saying that if I go onto Adobe and I click on that thing that they call a, an AI agent for their PDF files, and I use that to summarize some PDF report, that I am not on the forefront of the agent workflow wave. I'm shocked and appalled. You were on the, you were on the forefront of the Mike Vard AI wave on, On, on the floor or something, Mike, I don't know if it was the front of the back.

What you may On the forefront of is creating security problems though. Yeah, I Not, but guys, lemme let me, let me, let me put this in. You know, Mike, I bought into all this stuff until the last segment when you told me that no one pays attention to what the labs and the surveys are about.

So I said the benchmarks, the Surveys are Oh, benchmark, okay, was just the benchmarks. Well, now he's splitting hairs. Okay.

Oh yeah. Now, yeah. Now he's, he's, he's, uh, distinguishing, we call that in law school.

He's distinguishing. But, but seriously, these surveys are all over the place, right? This reminds me, this reminds me of the Wharton MIT thing, right?

And mi t 95%. Yeah. Don't see a damn thing Wharton says, up to 70% think it's pretty good.

I don't, it depends, You know, then Harvard. Harvard actually credible one's credible. And what is it?

One is incredible. That quickly. Well, there you go.

Spoken from the Wharton guy, John, what, what were you say? Well, I was gonna say mit, MIT has that, that it is terrible. I mean, terrible results.

But Harvard came out with their own study. They basically poke the MIT in the eye. So, I mean, there's always a, there's some always somewhat of a hidden agenda in all these things.

It depends on who's producing it and what they're trying to sell and who they're trying to sell to. Yeah, I mean, you're right. It's just, it, it's all over the map.

So imagine, imagine for a moment, uh, whatever news outlet you find, not credible, okay? Just imagine in your mind for a moment, I'm gonna have them recast the headline of this story. The headline of this story is, businesses Everywhere tried to make money or save money with generative ai, and they couldn't figure out how to make money or save money.

So now they've moved on to ag agentic, Right? 'cause this time in Headline, headline, generative AI has officially failed. Everyone has moved on.

Everyone is is is desperately seeking ag agentic at the moment. That's another way to look at this, because, So, so let me, let me, let me turn that on a sec. Which uses a, everyone said gaudi's, the is is is a bomb.

6 billion on Sam Bon Nova. Yeah. I like, like, I like my headline better Bullwinkle AI this time.

For sure. Yeah. We said it Rocky all Alan, hold on a second.

Hold on a second. I think, I think it's important that we be accurate here. 4 billion.

That's a good, that, that, that's, that's how the VC looks at it, right? That's what running the messaging. 6 billion.

It Sounds like my wife shopping Grace, don't, she saved me money. So much money. I saved money.

Is that the headline From VC Today? Alright. Hey, let's take a break on this one.

If you are using agents out there, agentic AI agents, let us know. 'cause we'd like to do a little informal study and, and not the Adobe ones that Mike's talking about. But we're gonna come back here.

We've got an open AI advisory. Ooh, sounds cyber. You're watching Textron Gang.

Marks a turning point. Artificial intelligence is no longer just a tool. It's shaping industries, accelerating innovation and redefining how humans build, create, and solve problems from engineering and medicine to finance infrastructure and everyday life.

AI has become one of the most influential forces on the planet. That's why for 2026, we recognize AI as tech Strong's person of the year. Not for what it replaces, but for what it enables a future built together humans and machines predict.

2026. Join us. Hey Folks, we're back with our third segment and we're talking about, well, at least to my eyes, an unusual notice from OpenAI, a little advisory that says that the next edition of their models are gonna be a lot smarter and therefore might be used in nefarious ways.

And everybody should, I guess, buckle up buttercup. 'cause things are gonna happen. But Terry, you wrote this story.

What's your take on what's going on here and why did they feel compelled to do this? Um, okay, so paint this as my surprised face. Um, I really honestly, um, think there's maybe a little CYA here.

Um, let's send out a warning now that what we're doing, uh, has the potential to, to open up a Pandora's box when it comes to security. Um, and I I I, I think that might be really mostly what's behind it. I, I mean, applaud them for maybe sounding the alarm.

Um, but I, I, I can't go too far with that because I, I think they have their own interests at at heart here. But yeah, it's, it's an, it's an interesting thing. You know, they're, um, they kind of compare it to other dual use domains and say that, you know, the defensive and offensive, uh, cyber workflows, I mean, are gonna be equally used Right.

And exploited. And, um, that you need to sort of be, uh, buyer beware or be on the lookout. Um, I'm not sure if that's like taking the heat off of, uh, them, but, you know, um, I mean, it, it can't take the heat off them fully.

But I believe that's kind of what's behind, um, this, that's just, I mean, that's my, my feeling now. They did get applause from a lot of the security people that I talked to, um, for this story for, you know, being honest and blunt about what's to come. Me personally, I feel like I just walked into the back room of Rick's Cafe and was shocked to find out there was gambling going on.

Exactly. I know. Well, that's why I say this is my surprise face.

Who knew? You know, I mean, serious. I, I Was, I was thinking this is the equivalent of the labels on the cigarette car and this cigarettes cause cancer.

We still selling To you, So this coffee is hot, you know? Yeah. Right.

So, Alan, do you think that this is a somewhat, I don't know, disingenuous effort to prevent the label reckless being attached to a future lawsuit? 'cause at least we can say, Hey, we warned you. Um, I I, I don't wanna be that cynical that it's going to give them a lawsuit.

Immunity. Um, let me just say this though, that they, they waited this long to say it. Either they're incredibly dumb or, or someone, you know, whispered in their ear, whispered in their ear, something like this.

You know, Mike, I, I remember back when the no SQL databases came out and I talking to some of the no SQL companies, I asked them, there's no SQL, Stanford, no security. 'cause they knew, they said, you know, this was the CEO of Mongo at the time told me when our customers demand security, we'll we'll build it in. Well, did some customers just ask open ai, Hey, is this thing going to be used by cyber criminals?

Is this gonna be a risk? And they said, yeah, now that you mention it, it might be Well, and then, then they, they're facing, uh, or everybody's facing that. The issue that we've talked about a lot is how to put the guardrails up while this thing is moving, you know, and out of the gate.

I mean, how are people gonna actually be able to do that? I mean, what, how do you respond to a warning like that? What are you supposed to do?

Yeah. Well, That's part of the issue. I'm gonna come in from left field on this one.

I think this is, this strikes me as the equivalent of me announcing to everyone. Hey, I just wanna warn you, I'm opening a new restaurant in Manhattan. Yeah.

And the food, the food is gonna be so good that You may not get in. They're gonna be willing to kill each other, right. To get into the restaurant.

Right. So be warned That there's, there's a two hour wait on the line and Studio 54 is empty At least two hours and you better watch your back. That's how good the food is.

Yeah. 2 comes out and yeah, there are some benchmarks where it's, where it's, uh, significantly better than five one. But I'm wondering if, again, this idea of a agentic hype, how much value can we get out of it?

If basically they're saying, oh, no, no, just wait for the next one. It's gonna be so amazing. It's gonna be dangerous.

See, I thought maybe, I thought you were gonna go with the food poisoning at your restaurant. No. Well, that could be too.

Sort of a, it's sort of ridiculous. It's like, it's gonna be so amazing. It's gonna, you know, Hey, I can't, we can't, we can't promise you that Sam, your wife's not gonna leave you for it.

It, you know, your wife might leave you for this, this version of ai. Sam Altman is the Pi Piper of ai. He's the closest thing we've seen to Steve Jobs since Steve passed away.

He's an Dave. I wouldn't put it past him at all. I think that is par for the course there.

But let me, let me put on my take off my pundit hat, put on the cybersecurity hat. When I first saw this, I, I was of two minds. Are they warning that cyber criminals are going to use this right as, as a tool to make their cyber criminal activities more effective?

Or were they warning that the model itself read between the lines? I'm rushing out, I'm in a code. Red Claude is, is kicking my butt with, with better coding AI and Google's whole integrated stack in Gemini.

Man, I'm feeling the heat, I'm putting out my next model and we haven't really done the whole security testing thing on it. Oh, I think that's a good part of it too. I mean, I, I do, right?

Is that, is that what's Going on? It's not cyber criminals. The, you know, and it's the people that just use it day in and day out, you know, but, but it's creating, but It's spinner.

It's like, it's like Tom Sawyer painting the fence and now, you know, people are walking by and he is getting them to help 'em paint The fence. Because Mike, I know you went to school in the Bronx and, and the Bronx was not known for higher education. I think it was Huck Finn.

I think it might have been Huck Finn. I think. I think, I think, I think, I don't think Huck Finn was painting the fence.

I'm pretty sure It was to, it was. It was Tom Sawyer. It was Tom.

I went to school. What do I know? So be careful.

Before you denigrate my Finn was floating down the river. Long Island Throwing No, it was Queens. It was Queens.

Don't blame me on Long Island. It was clearly St. John's University.

Rick Patino is now our coach. We're back. Good.

I wanna say that, that about this memo or about this post, whatever it is, I guess it's an exercise. If you don't have a message, any message will do because it's got multiple things in it. Yes.

Cyber criminals can use this to do bad things. Didn't we already know that? Yes.

We're doing some things to try to improve our security. Okay. We assume you are.

You weren't very detailed about it, but, okay. That's good. Alright, we're starting this Frontier Risk Council.

Okay. That sounds interesting. Who's part of it?

Nobody. It's just our thing. We start it, you know, save, save that until it's real, where you really have some people, some industry backing behind it.

So I think it's, you know, they hired a new marketing firm or something and they weren't quite as, as, uh, experienced as the last one or something. I know. Well, yeah, that's true.

And when you look at the things they say they're doing, it's the typical stuff, right? Training the model detection systems end-to-end red teaming. I mean Yeah.

Should be standard stuff. Standard stuff. But yeah, that, that Frontier Alliance, I mean, that would, that would be, that would've been more interesting if, you know, and more real if they'd had some names out There.

Well, but in, in today's world, that belongs, like with the Linux Foundation, you know, I, uh, I interviewed someone earlier this week, you know, the new foundation, the Linux Foundation spun up. Was it Agen AI Foundation, something like that. That's exactly, you put, you put something like this in there and you got broad industry support and all of a sudden you, you get taken seriously.

And we won't be making fun of you on here. So, so how serious is this though? Because, you know, we're talking about, you know, this like Yeah, it's, no, you know, it's just yet another risk.

Or is this like, you know, an Oppenheimer level risk and people are gonna be like waking up one morning and basically they're saying, yeah, we created something that might change and destroy the world as we know it. We Created no, the mass destruction, The model just have a big jump in in productivity for capture The flag is your team is keeping pace with a 200% increase. No, but, but this goes back to what I said, they are feeling the heat, right?

Uh, Claude, uh, anthropic, right? They're doubling down. They're they're doing things Yeah.

Help that Is, How does this help that, Alan? I, I agree with you. They're feeling he, but I don't understand.

I think it helps the way Almost. I, I think, you know, all PR is good. Mm-hmm.

So while, while we've been talking that Frontier Advisory Board has just been valued at a hundred billion dollars, But Intel's buying it for $2 and 98 cents, First time in history, that a, uh, that, that the, the advisory board has had a valuation of, uh, of over a hundred billion with no members on the, on it. Can we, can we also assume that Google and, uh, philanthropic and all these other people are building similar levels of AI models and this will not be the only one capable of this Stuff? So I, that was, oh, yeah.

That was also, when I was writing this, I was wondering, are they gonna come out with similar warnings or are they just gonna let these guys lead the way? Only Only if the government mandates it. Yeah, I would say I would, if I were Google, I would be all over this.

I would be saying, yeah, we're not gonna have that problem because we actually build the hardware and the software and, and I would go into some whatever is, whatever the story is in terms of, uh, chip level security on TPUs. And I mean, who knows if it's, if that, if that's a genuine defense, but, but I would seize upon this. I wouldn't, I wouldn't seed that high ground to say, oh, your model's smarter than ours.

Therefore it's, it's more dangerous. No, no, no. That's not smart.

That's, that's called being stupid by ours. It's the secure one. Yeah.

And it's, it's integrated. Yeah. Well that's, I mean, yeah, I think Google should seize on it.

I I don't think they should make any promises that it wouldn't happen with their model, because this, as soon as they do that, something will happen. You know, I think they should all come with a subscription to cyber risk insurance. You know, kinda like your credit protection.

Right? Crazy. Just crazy tough.

But look, this is the brave new world we find ourselves in. I can't wait for next year to see how all this plays out. Yeah, I'm, I'm hoping we make it through the year.

Yeah, I was just thinking that, Wait, this is really snarky Thursday on, I guess, I guess it's, I guess it's, I'm sorry. It's a tough, I'm just, I'm just, it's a tough crowd out here. I'm just trying, I'm, I'm treading relatively lightly because I don't know what Terry did to make you guys put her up against the wall.

That's right. Well, I gotta tell you, last week she wasn't, there wasn't a brick wall there anymore. There was no Wall last week.

Right. Last week was all, now she's back to got the brick wall. She got out.

Well, and after the first of the year, there may actually be artwork back there. So just telling you, keep an eye out. I think the artwork could be a, a, an inclination of what AI's gonna be like for this year.

It could be some hidden meaning, like if you spin the artwork backwards, it says Paul is dead. Well, Funny you should say that because one of the pieces is actually a sound wave from John p Prime's, angel from Montgomery. So you don't know that when you look at it, but that's what the sound wave looks like.

There You go. There you go. Crazy.

Hey, we're gonna wrap up our Thursday show. What a great show, gang. Thanks for joining us, Terry.

Always great to have you, Dave. We gotta get John here more. Mitch, John, I, uh, Taylor's reaching out to both of you for our year end, uh, retrospective next week.

I hope you guys can make it. Um, Mike, I'll be in touch with you. Yeah.

All right. Hey, Mike. I not a word about the Knicks.

We won some crappy made up tournament for Gibo purposes. Hey, I, Todd, it's the first time we've won something since 1973. Yeah, Embrace it.

Yeah, embrace it White. Are they still using facial recognition at the garden though, for nefarious purposes? It comes with the Peace prize too, by the way.

As a, as a New York sports fan, that's just like a guppy that should be tossed back in the water. It's Like, oh, stop. We're gonna actually raise a banner on the roof for the garden.

The last time I did, you know, when, we haven't done that lately, other than when some of the great k Nicks passed away or something. I, I would, I would be, I would be embarrassed to see somebody hoisting a banner for this stupid tournament Regard. You mean you don't like a made up award, Mike?

It's not a made up award. Like good Citizen or something's a, it's A, it's, it's a yes. Nick's favorite gambling trophy winning team of 2025, voted by your friends in Las Vegas.

I mean, whatever. Go New York. Go Ns.

I'm proud of him. I don't care what Mike says. We'll see you tomorrow here on Textron Gang.

Thanks for joining us.