$6T AI Spending, Military AI Expansion, and the Growing AI Knowledge Gap | TSG Ep. 1013

Gartner predicts $6 trillion in spending the military ramps up. AI and Anthropic opens the black box today on Techron. Hello everyone.

Happy Thursday. It's great to have you all here on, on the Gang. We appreciate you.

Um, it's an, as I've been saying all week, what a crazy week for news. We, we keep going back and forth. It's so hard to just pick three things.

I feel like, I feel like I, I need Adderall or something for A DHD. There's so much going on out here. Oh.

Um, but that being said, we, we've picked three good ones today and we've got a really good gang to talk about it. Let me introduce you to our gang. First of all, in all her Valentine's Day glory.

It's our friend, our friend Ann, our whole award. And I love that background. Thank you so much.

It actually makes me feel kind of cheerful too without coffee. It's, it's waking me up. Hey, if, if it feels good, do it.

Next up we have our friend Fred Wil. He has his normal background Baha mug. Valentine's let's too old for Valentine's Day for some of us.

And then we have our friend Gina Rosenthal. And out in the valley we've got John Schwartz, and of course Mike Ard. Gang.

Welcome, welcome. Happy Thursday to you. Um, so Garner, well, we'll talk about Gartner, but Gartner is saying that we're gonna have a $6 trillion it spend this year or next year.

We're up this year, 2026. Yep. 6 trillion bucks.

That'd be almost enough to man the data centers. $8 trillion build out. But, um, in any event, that's an awful lot of teas in 6 trillion, awful lot of dollars.

It's a great headline grabber for a company whose stock is dropping like a rock. But Mike, what do we got? Well, this Has always been something of an inexact science, but it's an annual tradition now, I think where Gartner comes up with their 2026 annual forecast and they've done it every year.

And honestly, I don't know if anybody's ever gone back and actually measured like how accurate these forecasts are and historically, but it's better than nothing. And I guess, Ron, I'm gonna get your read on this whole thing, but, um, you know, it does seem to me like there's a lot of, uh, inflated AI spending in this projection. And I can't help but wonder if that's like, you know, five companies spending billions and everybody else kind of just watching, all watching along.

Yeah. Yeah, you're right. It's, you're right.

It comes down to about a handful of companies doing the, the vast, vast majority of the spending. 15 trillion in 2026, which is up about 11% from 2025. Majority of it, the engine of the pri of the growth is gonna be data centers building out infrastructure.

As, as Alan mentioned, um, the data center investment's gonna go past 650 billion evidently this year. Um, other highlights, and there are a lot of highlights, and I have to look down just my cheat sheet to make sure I have the numbers right. Software sector's gonna continue to grow.

4 trillion. Uh, gen AI is going to skyrocket in terms of spending. Uh, however, on the flip side, let's mention some other things.

The consumer spending on, on spending on consumer enterprise devices is kind of hitting a plateau because of the rising memory prices and supply shortages. You know, it's one interesting sidelight is, um, there's been this big tech stock selloff, and one of the factors that went into the selloff was, I think it was like a trillion dollars was lost. 'cause investors are worrying that AI is going to disrupt traditional business models.

So I kind of found that interesting. Uh, on the flip side, um, there was another report that we, we mentioned the showing notes from Databricks that shows a, uh, a big spike in enterprise adoption of multi-agent AI workflows. So like 350%.

Um, it, it, this stuff is, it's, it's, uh, the number is big, but I think it's a kind of a duh type of conclusion. Um, Daniel Newman from Fu Futureum group been talking about this for a long time, around the same types of numbers. Um, it's, it's a build out still.

Uh, and it's, it's a growing, it's a growing industry. Um, the infrastructure is the story for now. Um, it will we'll kick the can down the road about profitability and about profits, but, um, profits in general.

But, you know, it's an interesting study. And again, you know, with, with Gardner what we need to do, I think what somebody should do, I don't want to do it, but what somebody should do is do, is look back the last 10 years in their predictions and then cross-reference with what really the really numbers were. Um, I used to do that with sports.

I used, I used to read the predictions of the, of the sporting events of the, the standings at the end of the year to see how bad the predictions were. I think we probably should think about doing this as Gardner and other groups. Sean, I thought you were gonna say, put them outta their misery.

Well, we, we mentioned yesterday, uh, the stock, it was it like, I wanna say 80%? No, no. Well, it's down like 75% in the year.

5%. Yeah, that's what I Meant. Or it, it dropped, it dropped just the other day alone.

Uh, some outrageous number, like 35, 40%. Uh, maybe it ended up in the twenties, down 20 something percent for the day. I'm not entirely sure.

I, I'm not entirely sure I trust the sources that Gartner's using. 'cause they're surveying the IT people. And I'll, I'll put two things out there in perspective.

Well, maybe three. So DataWorks 300% increase. So 300% on top of one is, you know, three.

And if you look on the today, there's a couple of stories up on text drawing AI talking about some other surveys where Salesforce just announced that they have a Salesforce, where they looked at it and said, yep, people are deploying AI agents. And on average they have in these organizations deployed 12, that's an average. So some have 20, some have two.

And they're also forecasting that by the end of the year, the average will move to maybe 20. Okay, that's nice. And then I go look at what those AI agents are I wanna be talking about, well, their AI agents then were exposed by Salesforce or they're exposed by, um, you know, we'll take for example, Adobe.

I happen to like their IA agent, but it's stuff that, you know, is in those SaaS apps. It's not like I went out and build it. Now Zapier is also saying though, they'd come this year, um, a quarter of the IT folks that they surveyed expect to have very, you know, sophisticated AI agent orchestration frameworks and automated workflows.

But here's the thing, and in my experience, and Gina, I'll ask you this, I have never met an IT person who did not overestimate the impact it was gonna have on the business. They just, you know, it's all rose colored glasses, and then you go talk to the business people and they're always like, come again and say, what? Didn't you see it that way?

Yeah, I totally agree with you. I was cracking up at your, the first thing you said about the 300%, because as a marketer, it's awesome when you walk into a business that has no marketing at all, because every single number you do for the first year is like 14000% increase. So, um, yeah, but back to your idea of, um, of the business not seeing the impacts.

Yeah, I think that's, that happens all the time. Or you think you do this wonderful thing, especially when you talk about IT and operations, they didn't even notice it at all. They just didn't complain about anything for a year.

And really that's where the metrics should be. Maybe, You know, Mike, I've got a shimmy law for this. Here we go.

Here we go. Right? Hey, hey, AI taketh right?

For everything AI giveth, AI taketh away. So for all of this $6 trillion that they're attributing to the AI infrastructure, boom, we're losing money and market capitalizations and CapEx opex faster than, you know, than the tornado that carried Dorothy to Oz. Right?

This is, you know, you look at a company like Salesforce down 50% on the, uh, 35% on the year service metrics down 50% on the year, uh, Microsoft, down 4% on the year, Oracle, which was a high riser for a while back, flat Gartner dropping like a rock Forrester. I don't, I don't think Forrester could go any lower, right? Mm-hmm.

There are so many businesses being upended disrupted and their, and their market caps, you know, reflecting this, even though some of them are still making good money, they're making their Wall Street estimates and everything else, and they're getting punished and killed on the thing. And at the same time, Databricks says, well, we're gonna have a 327% surgeon enterprise AI systems. You know what, what AI takes with the right hand or what gives with the right hand, it takes with the left hand, and I think the left hand's taken more than it's given at this point.

Yeah. I also think that there's BS at work here that goes like this. So let's say I'm a CIO and I wanna buy some more switches, or I wanna buy more storage or whatever it is.

Am I gonna put that in as a request on its own? Or am I gonna attach that to the AI budget and say, yes. And of course, we need all this stuff and all this extra gear, and I may not be using it all for ai, but as far as the CFO knows, well, yeah, it's all for AI because, well, that's the strategic imperative from the board.

So this is like a bill moving through Congress and I just attach all kinds of crap to it. How much of it is AI related? I don't know.

Does that ever happen, Gina? Yes, it totally does. You know, and just to put like, uh, maybe a little bit of positive spin on this, um, there's so much hype that the business leaders have no idea what AI is and what it's gonna do and how it's gonna move the needle for their business, not necessarily for their usage of ai.

And there's really not a lot of there, there yet. But when I was at AI infrastructure field day last week, um, I, and it came away actually with a really positive feeling that, hey, the infrastructure people are getting our stuff together. We're making it so people will be able to, when it becomes a little more mature and things are available, they'll actually be able to develop agents and AI systems that will move the needle for their business.

But we're in the very, very beginning. So I guess they gotta keep that hype going, but as long as they do, there'll be places to stick your request in for critical projects that need to get going. Mm-hmm.

And a great point, and I wanna kick this one over to Alan, but, um, correct me if I'm wrong, but I believe that every aspect of the AI supply chain is constrained right down to the metals that we're using to create the processors, the processors themselves, the data centers that the processors are gonna go in, and the skills we need to build all this stuff. So, you know, is 2026 really the year of ai or is this the year where we kind of finally get our resources together and 2027 might be the year of ai? Yeah, I gotta tell you the truth, Mike, I'm feeling a little bit like we're on Hollywood Squares and I'm either Charlie Weaver or Paul Lin.

Uh, wow. 1975 is calling right name. Rolling.

All right. They want their, they want their game show back. Drew Barrymore is now the center square.

Oh, wow. Okay. There you go.

There you go. Excellent knowledge. I Love match game, old seventies match game.

I I aspire to be Charles Nelson Reilly for maybe we could do a extra gang match game episode one day. But in any event though, to, to answer your question, you know, in a world where we have a a, a Church of Crusta Arianism, okay, what are we doing? What did it, we might as well throw darts at the border Yeah.

Or blindfold me, twist me around and tell me to pin the tail on the donkey. This is, we're living in it's crazy times. It, it's, you know, so trying to make rhyme or reason out of an illogical timeframe, you know, in my best Mr.

Spark is illogical, right? I mean, it's, it's, you know, what do I, I just, I shake my head, you Know what fun, fundamentally, I think what with the problem I always have with these types of reports, and it went back to Data Quest, which I think was one of the worst defenders. Their clients are the companies they're running the reports about.

So they're getting paid by these people, and they are, they are dependents on the information they get from the companies. And I remember this, this years and years ago I was working at a newspaper, I think it was a Chronicle or something, and there was a dispute among the companies over the final numbers of a data Quest report. So I, I spent like two days looking into this, and I found out that they were massaging the numbers to make everyone happy.

And it was, it was, it was something they do all the time. And eventually Data Quest, actually, God, I shouldn't say this, but they actually said to me, so what, what are they telling you? Which, which numbers should we we change to?

I mean, I, that to me was the eyeopener. I was like, okay. Uh, well, I think, I think there's a couple of real concerns here.

Uh, and, and just quickly, I'll, I'll pause it. This, the, the growth probably hyperscalers most of this here, but if you look at the average age of the board of directors, right, of the Fortune 1000, you're not talking about folks that are making decisions on millions of dollars of revenue, is billions and trillions, and how many have access and understanding and context on what AI does for the business or won't do for the business. I guarantee you all of these companies are trying to figure out maybe desperately in some cases, how to make sense of their business model moving forward.

And so is there profitability happening in the marketplace? Sure. Is it being completely undercut by things?

We don't understand it. I'm sure that's true too. And the concern, or the interesting part I'd love to see is, I mean, we see these, these are all trailing indicators.

Uh, the prophetic notion of what, you know, Gartner talks about, that's great, but the trailing indicators of the market, and so interesting part for us is like, where are the leading indicators become AI oriented things that, you know, businesses can operate on instead of, you know, the reactionary force. So I think this year and the next year and the year after that, it's AI is the future. So it's not a bubble.

The market of where we're spending on AI may be a bubble, but that's, it's a new way of life. We gotta get comfortable with what that is and how it impacts the marketplace. I agree with you, Fred.

I agree. Hey, you know what though? Whenever companies are looking for money for a particular market or anything, one of the lowest hanging fruits and the fattest of pork is the military budget.

And especially in today's day and age. So let's, let's jump to our second, uh, segment here. The militarization of ai.

Hey, smells like victory in the morning. Yeah. Well, we have a post up on digital CXO that kind of points to this, but it seems to me like the issue is, of course, the military is gonna be using ai, but, um, the way we structure things in the United States government now, it's like, well, the line between the Department of Defense or Department of War and the Department of Homeland Security is very porous.

And a lot of the technologies that the military is using for AI are finding themselves into a Department of Homeland Security that's starting to use these things in ways that may not be in keeping with our, shall we say, constitutional values. Gina, what's your thought on what's going on here? It doesn't seem like we have a whole lot of oversight That's putting it lightly.

Um, yeah. So I, the Department of War and the defense Secretary, Petes Hegseth is only wanting to, he wants to make sure all the AI models are there for him to fight wars and wants him to fight war, and doesn't want them to be so woke that they won't do what he wants them to do. And specifically talking about anthropic, anthropic seems to be the only model that, um, is worried at all about civilians being killed.

That's quite problematic, was the quote from the CEO and, um, is worried about it being philanthropic, being used to specifically target people and, and whether they're in the country illegally or not, or whether they've broken some kinds of laws or not. And that's kind of why you have to look at it. So what, what exactly is the big thing that ICE is using?

ICE is using, um, something from Palantir called Elite, the Elite Platform, and that stands for Enhanced Leads Identification and targeting for enforcement. And it pulls information from all sorts of government databases, from Medicaid, for the health records, from travel records, from driver's license information, passport information, all of those kind of things. It maps the neighborhoods where they want to go and target people for, um, um, deporting them.

And it also generates dossiers about the individual that they want and a confidence score of whether that person will be in the area or not. Um, they're able to do that because they also use Clearview AI that uses Face, uh, and a product called Mobile Fortify, which I think is the app they use to scan people's faces. Basically, it's, that's facial recognition.

Um, technology they can use up to 200 million images at a time, which come from all of the government information, phone tracking tools, license plate reading networks, and even our favorite real time location data that comes from those commercial brokers we talked about a few weeks ago. Um, you, they can, that lets ice identify individuals, even entire communities, and track them and decide when they want to, um, move in and, and do things. But all of that information being pulled into one system and then being used this way doesn't just impact, uh, immigrants.

It impacts people that are, are citizens, the several who have been arrested, um, or killed or killed. Exactly. It also, not only, not only does it, those people come up in the searches for that platform, but it also, all of us are probably in that system.

So our identities are, faces are our roots to work or wherever else we go have been pulled into a surveillance infrastructure, which is completely un-American. And I think if my grandfathers were alive who both fought in World War War II and Korea, they would be absolutely out of their minds about it. So it is a very intrusive targeting pipeline.

Yesterday when I was preparing to talk about this, I, um, there's just so many, there's so many correlations of what happened, um, with Nazi Germany and IBM, and if you've never read the book about mm-hmm. IBM for Holocaust, you absolutely need to read it. IBM, um, continued to market and sell their, their mainframe products in Germany after the rest of the world.

Boycotted selling anything to Germany. They provided equipment, they provided services, they provided support, everything that they needed. And what they, the Nazis did with the mainframe was they did the censuses, but then they also were able to track people, but also neighborhoods and decide when they were gonna walk in and take those people out and send 'em to the concentration camps, much like they're doing today.

ICE is doing today with the immigrants. They also, the tattoos early on that people had on their arms were the primary key in those mainframes from, to locate them. So they wanted to track them while they were working and what they did and how productive they were.

And then finally the end when they actually were murdered, it got so bad they stopped using the IBM numbers because they were just bringing too many people in, and they couldn't wait for the mainframe to catch up with numbers so they could get them tattooed. So they used another method because they were just bringing in so many people. There is absolutely a direct correlation to this, even though copilot wouldn't allow me to, to bring that information together, I kind of had to pull it together on my own, which was interesting to me because you can't talk about people being like the Nazis, even though that's what SAP actually happened.

Where for watching, again, is technology is being used for absolute evil. Gina Bravo. Thank you for that.

I, I appreciate those insights. And I also think it's important to mention that all four major AI companies that consumers are using are also working with the Pentagon Open ai, Google Anthropic and Xai all have contracts with the Pentagon. They, they're all every, And, and let's get something clear.

You can't change the name of the Department of Defense without an act of Congress by Fiat. But what I find really interesting is we call it the Department of Defense, but the drunk is the secretary, uh, the Department of War, but the drunk is the Secretary of Defense. Yeah.

Weird, right? Shouldn't he be Secretary of War? Does that make him like Mars or Aries or something?

What? The whole thing is ridiculous. So there was a more troubling story, too, and I think it was in the times where, um, they said somebody wrote a letter complaining about what was going on with ICE and sent it off to the Department of Homeland Security.

And then, you know, shortly thereafter, they got a notice from Google saying that, uh, they were the subject of an investigation from the government because, you know, somebody was looking into their mail and their systems and all the other things. Now mm-hmm. Is there is causation and correlation firmly established there?

Maybe not, but it certainly is trembling that suddenly, you know, some of this stuff is being used by maybe the Department of Justice now as well. So, you know, we're crossing off Now. He's being naive.

Kay, come on, Mike. Of course there's correlation there. Why are you even, what are you trying to ham and horror and, and cover yourself trying, obviously trying To be a fair-minded journalist.

My journalism soul says I need to put the word alleged in here. Some, yes. Bravo.

It's my opinion, it's a good way to start. So I am a, I am an immigrant to the United States. Uh, I got my citizenship in my early twenties, and we thought that George Bush Jr.

We thought w was gonna do what's happening now. And so my family got citizenship. And in some of the work that my father did, I learned very young that the FBI had a file on us because we were immigrants.

And that never really bothered me because, you know, you kind of wanted to trust your government, but we're going from data collection to making moral judgments, to building profiles and making these sort of gray area judgments that a human, I think, really needs to sit in on. And that to me is the scariest part, right? This, this, this is, it's not mature enough that it should be used.

And then obviously the same issues that have been true with facial recognition this entire time are still true now, which is that they have a harder time interpreting black and brown faces. Um, and there's an inherent bias there, and it's been heavily litigated. I mean, Amazon abandoned facial recognition for quite a while due to some of their, their lawsuits, especially with Clearview.

So I, I don't get how, in a matter of less than five years, I mean, I get the advances have been sweeping, right? But I just don't get how this is where we are. It just seems like poor judgment all around.

Absolutely. If, if I could, uh, let's disambiguate a couple things. Uh, department of War, department of Defense, and what their operational charter is, is different than what Homeland Securities is.

Absolutely. And, and be clear about that. So the intent and the action and everything that he has to say about it has defending the nation at heart.

Let's put that aside for a second. Uh, I, I, I wanna talk about the surveillance apparatus and the collection of information. This has been going on since the fifties.

To be clear, the method of analysis may change over time, but, you know, when we had the Red Scare and McCarthyism this was happening, then the difference being now we have more points of telemetry. But let me double click on that without the moral, you know, the moral fabric of this conversation, I think to try to keep it somewhat agnostic around that, uh, whether or not there are the appropriate controls for the access and the actionability, that information, those are good questions to ask. I think, uh, Palantir has been used by the, by the military for quite some time.

And I wouldn't say that it is immature. I would say it's, it's actually relatively mature. The question is, is again, how is it used?

When is it brought to bear? And when we look at the rule of laws, it used appropriately for American citizens, the requirements we have for trying to solve some of the problems we're currently koston, make that a moral question for everyday Americans. But I just, when I hear things about the surveillance state, you know, we've been doing this in this country and another country for a long, long time, and it is not uncommon for folks to have a record somewhere that captures all of this information.

But remember, it's a lot of data and not everybody is going to be the subject of an investigation simply because data was collected. Fred, with all due respect, with all due respect, you're not wrong. What you're saying.

What, what, what the, the key ingredient in the formula that's changed is weaponizing this for retribution. Right? And that's, that's where the line has crossed.

That's where paler has become Halliburton, and this government's private army and Peter Thiel and his group of white nationalists get their, get off and make all of this money, right? They're, they're, they're a creature of the government. And any time, and if you've watched any movie or read any books, you know, those kinds of organizations are prone to abuse to abuse.

And that's the problem we live in today. We have a government that is prone to abuse, to retribution where the, we, we used to protect whistleblowers. Now we prosecute them, right?

And, and that's the thing that people gotta be afraid to speak out. You know, speaking of speaking out, um, what happened in Minneapolis and what continues to happen there elicited two different reactions. I think about the tech CEOs who were fairly muted about this.

In fact, some of them were showing up at movie premieres or other, uh, sundry events. On flip side, I wanna mention that there were 60 large Minnesota companies and the CEOs of those companies that include Target, best Buy, general Mills, 3M, they, they spoke up. I mean, they, they actually made a statement.

Um, I, I just wanna throw that out there because I am, I mean, I'm not, I've been selling this over, over again, but it's just, it's just very agree. Wait, I wanna ask Ann a question about this. 'cause there's, there's iron, there's irony in, in all of this, right?

Mm-hmm. So for years, the folks on the right complained that the government was targeting them, and they were overly intrusive and they were using the apparatus to target them. Now, the left is pretty much saying the same thing.

So, am I gonna go to DC one day and find two protests from the left and the right about the same thing? Absolutely. Yes.

Because we're operating in a society where it's low, low trust. We're in a low trust society, and we have decided to carve out roles of us versus them. It, it's, it's, it's very weird.

I, I remember a time where we had moderates. It was a valid answer to say I'm a moderate. Mm-hmm.

But you say that now in polite conversation and people will press you to go, yeah, but are you really left or right? We, we still have moderate, they're just not as loud as everyone else, right? It's not a spectrum.

It's a horseshoe. When you get to the polar ends, They're about the same. Yeah, I agree.

Go ahead, Gina. Yeah. I wanna say one more thing.

The reason I brought up, um, IBM and Nazi Germany is because of the technology standpoint. Yep. So technology, because we are the keepers of the data, is never neutral.

The reason that the Nazis used the IBM platforms was because, um, it was the latest and greatest and it could do what they wanted to do in really short amount of time. That's why they used it. That's why they went after the money that was there in Germany.

And they also played the other side over in the States talking about both sides, right? It is very much a parallel to what's happening now. It is all about the data.

Palantir's figured out a way to put all of this data together, the government data plus the social data, plus the broker data that's really, um, timely to do the same thing to target individuals and communities for, to, to move, remove them from this, this country. Well, Here's, here's what else makes me sick and ashamed. Gina.

You know, John, you mentioned the Minnesota CEOs and good for them. It, it used to be tech CEOs who were those people? Yes.

Speak out. Now you've got tech workers trying to pressure their boards and their CEOs to speak out, but very few of them actually speak out. Yes, I was gonna say that, bros.

Did you read How, how, but it's, It's only a small, very, very, Well, I got two people at once. John's and him. Okay.

It's very important to mention this. The number of people who speak out within the tech companies and, and these are the large companies, is a very minuscule number. I mean, there was a, a petition signed by less than 500 people.

And you think about those 500 people worked at companies that employ a million people. And the reason why, there's a couple of reasons why they work for companies that are in the mid midst of slashing their workhorses. So they, the last thing they want to do is create or raise the, the target to themselves.

The second thing is, it is just they are terrified of, of job security. Um, it's, it's kind of this, uh, close your eyes and pretend nothing's happening. And, um, you know, the workers at these places have to, and by The way, Just going out at the top To Gina's point, that's what's happened in Germany.

A lot of people dis oppose their eyes in the thirties. And what do you have to say? I am sort of saddened by the amount of tech CEOs who have genuflect genuflected to the current administration.

And if you look at Tim Cook's statements just a few years ago when there were, was strife or other issues, uh, if you look at what he said about George Floyd, what happened there. And if you look at what he is just said about Minnesota, it was so passionless middle of the road. It was like, where did, where did your inner compass go?

Models go. Yeah. Yeah.

Because everyone could always kind of look to Apple as the good guy, even if they weren't always. They were sort of, and have been historically seen as the good guys. And when I read Tim Cook's statement about Minnesota, I was like, This is, that was after he trust people.

That that was after he offered up the golden calf. I mean, apple, the golden apple. I would just add one last thing to this whole thing.

I have serious doubts that there are moderates, and I'm going back in time. And I remember that there were friends that I've asked, and they said, oh, yeah, I'm a moderate. But when I poke at them and I asked them what they feel about X, Y, Z, and ZZ well, it turns out that they're far from moderate.

Well, Exactly My point, Mike. Exactly. I, i, guys, we could talk about this all day, but we'll, we'll all wind up under investigation if we aren't already.

Let, let's move along. Let's move along and talk about understanding ai Andro, who, you know, their clawed constitution and everything else, uh, seems to be trying. But anyway, they, they've got some new research.

Mike, what's this about? Kudos to philanthropic. At the very least, they're making us think.

And they have a new report out that says that, well, if you're using AI to, uh, to acquire new skills, turns out you may not understand it as well as you might if you had learned it yourself. And I'm gonna pass this over to Fred, but it kind of reminds me of, you know, back in the day, I guess, you know, I used to have a decent understanding of how my television worked. 'cause it had cathode braid tubes, and I went to the store with my dad and we would replace them.

And then once it became all solid state, I lost interest. But I don't know, is that the same thing here? Robots today have been gold?

No, we have dudes. I actually terminated, uh, uh, somebody for overuse of AI who was a developer this last year, uh, who had a, you know, a master's in computer science and a amazing pedigree. But I terminated them because their overuse on ai and they were too young in their career and not developing the skills and reasoning and, and following it when it was wrong.

Uh, O Okay, I'm gonna, let me, let me talk about the, the AI tool chain here and the conversation that anthropics forwarding. I think the, the submission would be they did a, a, a random blind controlled study on 52 junior software engineers and make that point so that you understand where they are in their careers, right? And basically gave them, uh, a, a Python library that they hadn't used before that did basically some asynchronous programming.

And, and they asked them then challenged to say, Hey, please, you know, build something with this and, you know, complete these, uh, uh, some coding tasks based on that. And some of the group was given limited, uh, access to some documentation. And some of the group was given just access to, uh, AI and the use of the, uh, of, of the library.

And, you know, things that we know, uh, if you don't have to work very hard to get an outcome, did you learn something in the process? Uh, the answer is yes. The question is who is going through the process of evaluating this?

So is, is AI a learning aid, not a substitute? Well, most of the folks that have been software engineers for a long period of time understand the value of testing and the iterative process of improvement of their code base and the understanding of optimization and refactoring and things like this, and folks that are trying to get something with speed and completeness, right? Comparatively showed that they weren't picking through some of the logic problems to understand the context of what the, uh, library provides for them.

So, I mean, while that's an interesting assertion, uh, I, I think it's an obvious assertion in the same sense. And so the question that we have to ask is, you know, in the triangle of things, you know, quality time and, and, uh, and, and, and completion of something, uh, are we, where are we spending the resources in order to do that? If you were to ask senior engineers instead of junior engineers, their stance on this, I would argue they probably both are arriving at something closer to equality around that, because they've already done this a million times, and those skills are codified.

So there is truth in the concern about whether or not AI is gonna help you learn new skills effectively for something you don't know if you don't know how to do that learning by yourself. But also that there is some assertions that, you know, speed may kill, you know, quality and, and knowledge context, uh, in the immediacy long term. I'm not sure if that's a great thing.

Yeah, I I think long term, Fred, five years from now, you look back and say, this developer got, you know, laid off for using AI too much. And they'll say how quaint, right? I'm gonna be talking about this on Shimmy, says later this afternoon about, we're in this scarlet letter phase, I call it, of AI right now, where if you admit to using AI to write code, make music, draw pictures, write articles, you, you, you know, you're, you get that scarlet a on your forehead or your chest or whatever.

And you know, it's akin to when we used to think that stuff, stuff that said made in Japan was cheap, like an old Godzilla movie with the thing on cardboard buildings, right? And then all of a sudden Japan became one of the top producers of quality products in the world and changed the way we manufacture stuff. And it went to China, and now made in China doesn't necessarily mean it's bad quality either.

We're, we're here all on Chinese equipment, probably. Um, it's the same thing with ai, right? It's a scarlet letter time that we're living in, and that's what we see here.

Well, about that. I, I think, you know, just coming from an educational background, what makes me a little worried is that if junior engineers don't get the chance to fail, because that's really where you learn it is. Mm-hmm.

You break something, you fix it, you learn. So if they don't get the chance to do that, will they have the skills that more mid-career and senior developers, it's not just developers, it's everyone. It's people who write, it's people who do any kind of thing.

If you don't have the context for something or you don't have the history of something, how can you write an article? How can you do anything? But, but here's the thing, But, but just, But maybe the, the job will be different than too, Gina, Maybe so, but I, how I can't see a time where people don't understand the basics of what they're doing because they haven't had, there's no, if you don't provide that intrinsic, um, force within a person to learn, you have to learn as you go on the job.

Nobody, you can go and learn coding in college all you want to, but when you get on the job is when you really learn how to do things. When you break something in production, that's pretty epic. You learn a lot from that.

So I would say the best developers are self, self-taught. You know, in my experience, I mean, I, I had a little bit shown to me and I figured the rest out. Um, but if you look at the layoffs and who is getting laid off in these big tech layoffs, they are the senior experienced people.

And so our economy is now saying the us all big tech is saying to people, experience doesn't matter. Yep. We don't wanna pay Experience.

So what you're gonna get is less quality. That's just, is That, is that just a function of ageism or is that really experience doesn't matter, or is just those people were the most expensive developers, so we cut them loose. It Was the, you guys are both your vault, right?

And we think that we can replace them with cheap tools. So, so a friend of mine who works at, uh, without betraying their confidence, someone, a friend of mine who works at at Amazon just got a memo. They're a senior, um, executive high up, they've been trained specifically in a certain area.

I don't want to say where it is, but they've basically been told we are sunrise your group and the people like you. And I think it is Mike's right, it is ageism, but I also think, uh, Ann is right too. Is this, is this kind of like the senioritis they have.

Um, and this is, it's gonna accelerate. It's only gonna accelerate with ai. So my first job, my first job at EMC was training the mainframe people, how to attach Unix systems to symmetrics.

I knew nothing. I was so green, knew nothing, but they had already laid off all the people that knew anything. Mm-hmm.

There was one dude and I, and I was really, really nice to him. And he was very, very nice to me to help me understand how to talk to these mainframe guys. The mainframe guys were getting retrained because they were moving into something else, else.

What I'm saying is, if you, and the, the other position I've been in my life is when during outsourcing, they pulled the same thing. They got rid of all of the, and I was kind of almost mid-career at that point. All sorts of stuff you need to know, and all the people are gone.

So yeah, I'm all for ai. And don't get me wrong, like I love ai, use it every day. I think there's all sorts of ways it can be used.

I think there is a danger of us having this missing middle. We'll end up like the mainframe people because we don't have people to come up and take our places so we can retire, Like calling out of the middle class or, uh, you know, it just, you're right, Gina. I mean, it, I you people, I mean, this is gonna happen, especially, I'll give you a simple, I'll give you a simple example.

So when we were kids, right, and you were learning math, they said, don't bring a calculator in the class. And then you got into college and whatever, and everybody had a calculator, and my kids all had calculators in their early access to it. And I swear their math skills suck.

I think That may be true, but I, I'll, I'll go back to Alan's point. The the market is shifting. The same skills aren't gonna be required, right?

If you look at, and so Amazon's layoff was not principle, just principle people, it was spectrum of entry level through VPs, right? A whole spectrum. In fact, business units, right?

Summarization for business units included. And some of the key things, the track are more like the hiring trends. How many junior engineers got hired last year went down by 60% based on job listings.

Like it's, we're not hiring people to learn things. We're hiring people to execute things, or we're not letting people go to execute things. And I think there is a shift in that and what your knowledge needs to be, the dawn of time of the micro application is here.

SaaS products are probably going to slowly die off. And you're going to get a, I don't need to write code. I need to know how to write something a customer needs, and I can write all of that stack and I can provide that to a customer and operate it that way.

So you can think about where the independent wealth tree comes in here for people that have competency and have learned how to do those things. But, uh, you know, I just, there's, there's not as, I don't, I don't have as much doism on whether or not we're going to, you know, completely optimize the workforce of talented folks. That that's impossible, right?

It's gonna be impossible for quite some time. But we're certainly seeing like, hey, if you got a computer science degree, if you notice the number of, you know, folks that are graduating with a compsci degree, four or five years ago, that trend started to go down as the number of graduates. Now you don't see folks wanting to graduate with, uh, a comp sci degree to write code.

You just don't, I, I didn't pursue a comp sci degree because I was working at startups at night and coding and doing things that were way more advanced than what they were teaching. They were two years behind. But something makes me think about, I think about something that Ray Kurtz while said when I saw him speak years ago, which was our notion of privacy changes over time.

So that makes me think that maybe that's applicable here. Maybe our notion of technical literacy should also change over time. Maybe That's, and and that's what I was getting at Ann.

Yeah. We're developers tomorrow don't write code. They manage AI that writes code.

And, and it goes back. AI doesn't take your job. Someone who uses AI better than you takes it.

And unfortunately, like it or not, this is the world we live in. This is the world, this is the time we are living in now. Stuff is shifting.

It's disruptive. We we're seeing you play out in the financial markets. We're seeing you play out in our personal liberty and freedoms.

We're seeing it play out in the job market. If you don't, you wanna stick your head in the sand and say, bull, bull crap. You know, I'm, I'm sorry, but it's not that.

That's what's going on here. Anyway, we're about outta time, guys. What an impassioned panel today.

And not only impassioned, intelligent, forthright, and just some really smart as heck stuff. Who needs that ai? We got smart people right here.

Um, of course we may need jobs, but that's another story. Anyway, um, and Gina, Fred, John, Mike, thanks for joining. Thank you for watching.

Hey, a quick reminder, I am on Shimmy says at two 30 Eastern Time. Today I'm gonna be talking about this AI Scarlet Letter stuff. If you wanna join in, uh, we've got text Drunk TV immediately following this.

And another reminder, starting Monday, Textron Gang moves to noon Eastern time. The network bosses have put us in a different slot right behind the Match game. Um, so stay tuned for us there.

But until then, is Alan, she, and we're out.