When Data Centers Fail: Outages, AIOps, Burnout & Observability | Techstrong Gang

Hey, everyone. Good day. Welcome to Textron Gang on Thursday.

It's a Thursday kind of gang. I'm excited to get jump into things today. I've been excited for the last couple weeks.

We're, we're living in historic times. I'm telling you, you're going to look back on this one day and say, those were historic times and there's so much news. How do you figure out just, what do you want to talk about?

I'm, I'm doing a shimmy says later today, but I'm doing two. Shimmy says, 'cause I couldn't make up my mind about two different topics. Well, I, I can't make up my mind a lot of days, but that's a whole nother story.

Let me introduce you to my, uh, gang for today, who we're joined with. We've got a Fred Wilmot back home out west. We've got Ann Al Ho Award recently coming off of TV appearance in hometown Houston.

Yes. Our man in the valley, Jon Swartz and our man in the city. Mike Vizard.

That sounds like NBA uniforms, right? And She waited in the nick of time. I had every faith in you.

I said, leave the door open. She's coming. She's coming.

So good for you. Muted. Gina, You're a little muted, but we're here.

All right, Mike, while Gina fixes her microphone, let, let's jump into what we're talking about today in these, as I said, historic times. Yeah. Well, To your point, you know, every morning now I go out on my deck and I look over the horizon just to make sure that the four horsemen of the apocalypse aren't on the way.

But short of that, um, That's not, that's that Catholic high school upbringing for you. There. There You go.

Mm-hmm. Thinking that the four horsemen are coming. But a funny thing seems to have happened on our way to AI Nirvana here, and there's a pair of reports out about it this week, sort of.

But one talks about how system admins are still as stressed out as every, in some cases, maybe even more so, because there's more workloads than ever. And the cybersecurity folks are reporting the same thing over on security below, where they too are stressed out. 'cause there's just more things for them to keep track of as well.

And a funny thing seems to happen here, and I'm gonna go to Fred with this one because I get that we're supposed to be using AI to make our lives easier and, and automate more things. But I think it's exposing some interesting underlying issues that we've ignored for many decades, which is the underlying systems are a freaking mess. And so now I just keep adding all this AI stuff on top of that, and I'm not really getting the, the true benefits of all of this.

So, Fred, at some point, are we gonna just have to rip and replace all the underlying mess that we have to make it truly AI native? That's a great question. Uh, and, and I think that probably starts with the humans part of that.

But, uh, I'll let, I'll let us cover that a little bit in the third segment. But fundamentally, uh, you know, I think there's a convergence zone happening here. And, and these reports illustrate a couple of pre top line concerns.

One, um, the, the pace of, of new technology and the rate of change is massive. In fact, it's so massive that, uh, you know, 10 years ago we would talk about, uh, how do I get to send my, uh, it admin home early, right? And now, right in the case where we look at the number of types of occurrences of, of technology influx, the integrations of all the broken things we already have with the new things, we don't maybe fully understand the implications of them yet.

And the consequence of the immediacy of those changes that now the pace of the industry is demanded. Uh, we've seen a pretty massive influx of both additional work, but also additional stress. Uh, some of that can be, uh, sort of analogous to what it looked like when we first started adopting a brand new technology, uh, industry-wide.

Um, now what we look at as, as AI adoption, while we're more comfortable with it, and a lot of the, the security leaders might be more comfortable with the adoption, uh, and they might have budget for it. The, the concern is oversight and governance over technical, uh, uh, prowess for these security leaders. And also the cross-functional responsibilities, not just sort of the standard security or the standard IT concerns.

It's, it concerns might be breaches first. And cybersecurity concerns might be, you know, hey, there's HR implications for these types of, of occurrences, which is a broader set of interest than, you know, we've probably reflected on previously. And some of that we can, we can owe directly to ai, including some of the budgeting requirements.

Historically, uh, it hasn't really had a problem with a budget, uh, or a budget crisis. Cybersecurity has slowly, uh, over the last handful of years become more resource bound, uh, sitting beneath in a lot of cases. The CIO most folks feel as though they've got enough financial wherewithal or resources to deal with AI tool spend and things like this.

But there is a, a, a growing gap on the ability to weaponize the technology through the use and the value of employees that already understand the business pretty well. We have seen a great rate increase in the adoption, uh, of, of technology just in the last year alone, uh, moving from 50 to 75% per the, you know, the, the simplicity study. But, uh, the question remains, what does the end goal look like and how much of a human governance problem are we not addressing here?

Mm-hmm. I think part of the issue too is going back to the tech, it's like, so all these things that we deployed had some sort of graphical tool in front of it designed for a human. And the APIs were, shall we say, uneven.

And the AI agents are looking for higher quality APIs, but those are not always as robust. And heck, some of these systems we have don't even have an API in it. So I guess I'm kind of wondering if this is all gonna beget some sort of reinvention of our underlying IT environments that maybe is AI agent first.

I doubt very, very strongly that most companies have any kind of real governance structure for new technology. I, I just can't imagine that that's a thing that anybody was dream, dream scaping for themselves. So the way I interpreted this article was that AI didn't really actually create a leadership crisis so much, is that it, it revealed it to us.

Um, and security teams are just the first ones that are feeling it, and they're telling us, 'cause you know, they don't like it. I think the other thing too, if you look at it from an ops perspective, and I know this was the, that they, um, interviewed the, the survey was of leaders. But we've been trying from a data center ops perspective to automate everything PO as possible for longer than I've been in the business or old enough to be in the business even.

So that's a pretty long time. But, um, so I I, they, one of the points was that operators like, yeah, heck yeah, we want AI can help us do a lot of the things we couldn't do. Um, but I see a people problem in this a, you know, not just a human, just a human resources problem.

You also look with, oh great, we've got ai, we don't need any young people coming in here 'cause we can run it with agents. And if there's no on-ramp to do the boring things that nobody cares about unless something blows up, how do we find people to learn and take some of the weight off of the senior assist admins? On the other hand, the senior system admins have to take on more responsibility until figure out, like, what does that mean to have agents running autonomously in a data center and making sure that they do the work and making sure that there's some sort of oversight, as both of y'all have said already to if the work breaks, then what do we do?

Then when you add the other human part that then, especially generative ai, generative ai, the interface is that chat box that anybody in the organization can use to access IT systems. And if they're able and, and have, um, the, the access, they're able to do that with ai, with, um, APIs. So that adds a whole nother section of people breaking things or people complaining when they try to use their AI chat bot to talk to the LLM to figure out how to do things that they don't have access to the things they want to do because they probably shouldn't have access to it.

So there's a whole back to the, back to the future kind of thing where this is basics blocking and tackling and what are the governance around what, what business people are allowed to do and how much then they impact the IT systems, and how do we build a collaborative environment that's never really been there before. Yeah. I think also one of the other things I hear from folks is that, and they're not complaining about it per se as much as it is an observation, but they're taking on more tasks and things that other folks used to do, they're now doing inside of their own workflow.

And that's not a bad thing, but it also is increasing the cognitive load for them. And so they're getting a little more challenge. I think Mike froze.

Oh, yeah. Um, so I wanna pick up where Mike's until Mike comes back. Um, so I talked yesterday the Dropbox VP of engineering, and he, he was saying a lot of what Fred was saying, that the technology integration is outpacing the human capacity and governance and we're creating this kind of stressful, exhausting, uh, situation.

And then one of the things that I think in the, the simplicity report that jumped out at me was they would talk about this six day work week. So basically people are working at least an extra 11 hours per week, and half of them are saying that this is, their roles have become exhausting and they're not rewarding anymore. And, uh, that's a patch we're gonna go through.

I think it'll, it might get worse. I think it'll probably eventually ease up a little bit as we get farther down the trail of the path of what these folks are going through. So let me, let me bring another, another view to this here.

A lot of people losing their jobs to ai, those people complaining, you know, they be careful what you wish for, right? Because be happy, you know, and I don't want this to sound callous, but if you're that busy, you got a job. Mm-hmm.

Right? Hey, I, that's a fair point. Haven't necessarily replaced you.

Now, You know, I'm reminded Fred, you'll, you'll appreciate this 20 years ago. I'm in, I'm, I'm selling, I'm selling. We're, I'm, it's still secure.

And we got this product called vm, vulnerability Asset Access and Management, vulnerability Assessment and Management. Mitchell, Ashley and I are sound that we meet with some cisa. It, it wasn't a ciso, it was a CIO.

And we said, look, we've spoken to your, uh, help desk team, to your patching team. Adopting Van with the workflow for remediation will mean the difference between your guys not having to come in on Saturdays to do patching when you know the network's down. Um, or, you know, the network's not that's used versus not.

Do you know they won't have to come in on Saturdays anymore? You know what the CIO told me? He said, we, we pay our people by the week.

Why do I care? Mm-hmm. That's real truth.

That's what the guy said. And that's when I realized what I was truly up against in trying to sell automation and remediation and so forth. Um, what I'm seeing is actually quite the opposite.

I'm seeing a lot of people very excited by the possibilities of what AI is bringing to their roles. Yes. There are people who are anxiety riddled and afraid to embrace it.

Right? Like it's fragile, it'll crack, but other people are embracing it. I got a problem.

I've had people sleeping in the office on weekends here because they're in love. They, they all wonder You were conflating two things that do not relate. You're doing an apple on an orange.

Just because I'm stressed doesn't mean I'm not enthusiastic. Well, the Apple stories later. Yeah.

Um, it doesn't mean that I'm not enthusiastic about it. It just means though I am getting more stressed because there's just more things and more moving Parts. Yes, there are more to do, but, but here's, so let me bottom line it for you though.

'cause you are conflating two things. What you're conflating here is that people like Jack Dorsey, we're gonna talk about it later, have taken the hatchet to their workforce and blamed AI for making it that stress. The stress we're seeing is because we we're running too thin in a lot of organizations that laid off a lot of people, and that that is stressful.

You got people having to do two jobs, three jobs because that's, you know, they've laid off people in a lot of places, those cuts have been to the bone, not the fat. True. And to your point about that, also, the roles are changing and that's expected, but I think a lot of people don't even know where their job begins and ends anymore because there's just so much that they're doing.

Um, some of it's just on the fly and that's not a bad thing, but over time it takes a toll. And it seems like, honestly, that sometimes speaking to that Mike, that you've got system administrators and network administrators who are responsible for security and responsible for the operations of the data flowing where it's supposed to go, which is critical in ai. You've got them holding the fort because things aren't changed over to AI yet, but then they're also building the airplane while it's flying to make sure everything goes, maybe that's why they're stressed out.

They're doing not just two jobs at the same time, but two different eras at the same time, trying to keep it all smooth so business stays up and those AI expectations are met. Yeah, I think that's a fair point, but I also don't think it's new that we stressed out cis admins. Remember when everybody started bringing phones to work?

Yeah. Mm-hmm. I remember when it was a big deal to put an Apple computer on a PC network because mm-hmm.

Apple desktop bus was, or there was a claim that Apple desktop bus to USB, they didn't conform. And so Oh, you put a Mac on the network, you broke the network. Yeah.

I, I think there's a, They, they Don't like change. People like things to be just, so I think that is part of it, but this is not the first time that we have angered the CIS mens and it won't be the last, well, I don't think it's they radio group. They're radio group.

No, no, I, no, no, no, no, no. I don't think that's a fair thing to say because, um, yeah, dumb stupid things. P**s CISs mens off for sure.

And that's not bringing a new technology and I'm making it work. A lot of times the cismen are on the back end researching, trying to figure out how to make the users happy, explaining up the chain, why it costs so much money, and the, the, the good things that'll happen, the bad things gonna happen, and how they're gonna mitigate all the, the air, the security issues. So I don't think that's fair to say.

I do think that cis mens know more about the organization probably than anybody else. And they know who's needs something and they wanna get to that. But in the meantime, they're dealing with all the extra pressures that are coming from AI enabled in the whole organization with no real oversight, no real ground rules of how far you can go with it, what you can can't do.

That stresses the out. It's not new stuff because they're probably been thinking about it. And I know this was a case with me thinking about it and preaching about it and building plans for it, building these transition plans way before the rest of the business ever got hold of it.

So they're not the block point for that. That's a lot of times a business and a money situation. I'm gonna come, I'm gonna come back full circle on this though, because with the, we never promised you a rose garden with the phones thing, right?

So we told you when we were getting the phones that we were gonna screw with you and that was that. But ai, we kind of, we've kind of promised these people that somehow rather AI was gonna make their individual jobs and lives better. And we're not quite there yet.

Maybe we'll get there, but right now it seems like we're getting A lot more, I, I just think that it's a major messaging issue. I mean, from the top down, there's this urgency, there's this hype, and it's this not filtering down to the rank and file and rank and file is confused. They see these layoffs, they wonder what their responsibilities are gonna be.

Like. They're working harder than ever. They're wondering if they're gonna be eliminated eventually.

I think that's just like a real concern and it's a legitimate fear that's going on right now. We're gonna go through this patch, but the next year or two People who are paid to do tasks and they're compensated hourly for their time, CI as a direct threat to their income. Because if they do things more efficiently, they have been reverse incentivized.

Right? And that is a comment I've had from my own team. That's a comment I've heard from other people.

Well, if I use it, then I'm gonna get paid less. I'm making myself obsolete. I've heard that too.

I'll add to that. 'cause I was at a conference last week in New York about that, and they were touting this new AI tool that was aimed at the service provider folks, whether they're professional services, taxes or IT or whatever. And they were all saying, this is gonna be great because you can reduce the amount of time that it takes for you to do any given task.

And a couple of people in the audience started figuring out that, well, the customer's gonna figure that out too. Mm-hmm. And so more revenue that I can charge and the number of billable hours that I can get away with might be dropping very harshly soon.

And then their business models are all upside down. So it's gonna be a funny world. So I, I think there's a couple really big important points to, to talk about here that make this different than everything that we've seen before.

Uh, I don't think it's fair to assert that the same old problems we've had are the same new problems we have. You know, meet the old boss, meet the new boss, same as the old boss kind of a thing. The difference is all of that workload, all of that expectations is all asymmetric back in the day.

Right? If we wanna use the error quotes back in the day, it's gone past asymmetric to symmetric. And now real time there isn't a, let's have a window for blankety blank.

There isn't a, here's a map for how AI's gonna be used. There isn't a map for how it's going to provide, uh, an understanding or governance around risk. There's none of that.

So everyone is all paddles in the water paddling as hard as they can. And the challenge that presents isn't just, uh, you know, sort of the standard human capital problem that's always gonna be there. Sure.

But what we're talking about here isn't just whether or not the bottom sort of tier really gets stressed and, and maybe not unionized folks that are hourly, uh, are gonna be exposed Here. We're talking about the top end of the supply market of the most talented folks becoming crushed with the level of execution required, whether it's force reduction or reduction in force, or it's a startup that doesn't hire 10 more engineers. 'cause they've got as much as they want out of the folks that they have and they continue to push.

The difference is there's no, there's no directional strategy is when the horizon is something we achieve. So if Mike is looking over the horizon every morning for the horseman of the apocalypse, right? There isn't a 23 mile distance he can see to or any of these folks.

And so they're just getting more and more of the business, uh, accountability, uh, layered on top of them because it is a technical drift from what they do today. And that really becomes the challenge. So if both security and IT professionals feel the biggest concern they have might be breaches, might be data exfiltration, might be, you know, the information loss to the business.

And that tells you how much of that centered actually in the technology that is ai. So is this the wizard and the wand to quote Fred Wilmot? Absolutely.

Oliver, Fred. Hey guys, we could talk about this longer, but I'm, I'm afraid I have to move us along. Move along now, move along, go play in front of your own house.

Uh, we're gonna next talk about data centers. You know, we're building 'em, we're not building 'em. We're spending money.

We're attacking them with drones. What, what, you know, the outlook is, what did they say? Cloudy with a chance of meatballs or whatever.

Um, Mike, what's this one on? So there's a report talking at about, well, data center construction. And it's interesting in the sense that it says that it's clearly there's a surge in demand.

Everybody can see that. But there seems to be a little softness suddenly at the rate at which these things are being built. And it, 'cause it looks like the first wave was built using money from, you know, the, the big tech companies that had stashed away all this cash.

And now the next wave is supposed to be funded by various, um, brokers and all these other folks and banks. And some of those people are starting to maybe pull back in the rate at which new data centers is being constructed is softening, possibly. Or it could just be a blip in the moment in time.

But Alan, maybe, is this the first signal of a bubble? What do you think The first signal? The first signal?

Um, so look, there's a couple of things going on here, right Mike, you're right. The, the, the first wave of this data center build out that we saw over, let's say the last two years came from the hordes, the hordes of money that was stashed away by the tech bros, right by the, the Googles, the mes, the AWSs, not the apples. Not the apples.

I have an interesting, uh, shimmy says this afternoon that we're gonna talk about that. And I have an article, apple actually spent 19% less this past year on infrastructure and data center. The other ones are spending about a hundred collectively about a hundred billion dollars a quarter.

It's a little crazy. But, you know, here's the bottom line. We all agree that we we're at 125% of what our data center needs are or that, you know, uh, to what data center we have.

We're, we're overextended, we need more data centers. The move to OPM to other people's money to build data centers was obvious. You couldn't keep even with the hoards of money, the, the, the the tech bros had, you couldn't keep spending that kind of money.

You had to go to the markets for money. Um, so that becomes an issue because now you got the bean counters involved there and they wanna know what their, their returns are. Secondly, you have this whole nimby movement, right?

The data centers are not going up as fast because people are realizing, you know what, I don't want it in my backyard. I don't want it in my neighborhood. Hey, if you're in the Middle East, you probably don't want it near you either.

'cause they're attacking them with drones. They're gone now. Yep.

So what do we, you know, but the funny thing is, is Mike, I remember sitting here with you and John, you were here and Fred er I don't know if you were on the gang yet, but I I'm gonna bet Ann was. We talked about so much of that $8 trillion that was pledged to ai factory data centers was pie in the sky. Yeah.

And that we wouldn't see it. Right? And, and now here we are, you know, a year and a half later or so and the pie is in the sky.

I, I agree with you and I think, you know, I remember in high school teacher saying that the greatest wars have all been fought over water. And I think the future wars are going to be fought over power. Right?

Electricity, the arms race that we're seeing is over power excess and grid infrastructure. Not just chips, not just, you know, rare earth metals. It seems to be where we're heading.

Um, Yeah. But interesting, Leanne, I I saw an article where the, the, the big data center builders are saying, Hey people, we'll, we'll, we'll absorb the cost of the electricity. Yeah, there was, that was, uh, yesterday.

Yesterday, uh, Trump, uh, and his tech ai tech bros met at the White House and came up with this voluntary protection pledge, basically. So where the consumers are not gonna get screwed by the energy demands. And the people who were there, I'm just gonna make this really quick.

The people who were there, who were very upfront were Microsoft Open, ai, Oracle, and AWS SpaceX was there too. The one company that was not there was Anthropic. And, um, Well, for good reason Broad is over there.

But, but that's all we talked about this a week or two ago here. We did that. That whole idea, they keep trying to phrase it in a different way that we'll take care of it, we'll provide the energy, don't worry.

But they're all providing energy on site building things that are polluting the neighborhoods that they go in. I'll give you a new example. I was waiting to get on and be able to talk about this.

Um, there's a dinosaur national Park, one of the state parks I think in, in Texas. I haven't been there yet. I love dinosaurs and we gotta go before this happens.

But they're building one of these data centers really close. And the the first defense he put up against all the people that are wanting to protect this place that will go away and that, that that environment will be destroyed because of it, of, of the data center. Don't worry, we're gonna build our own energy.

Well, you won't have to worry about the energy on the grid, but that's really not the point. It's more than just that. It's, it's destroying, um, natural resources that can never ever be replaced in search of the data center.

So there has to be a better way. But I'm one of the NIMBYs. I hate to tell you, I don't, I don't.

Well, John's point for You, to John's point, I would just point out the word is voluntary and, you know, voluntary means Sure. Someday maybe, if I remember I'll get around to it and I sure hope that you Are you impinging on the, on the character of the president and the tech pros? Well, I'll tell you what they'll do.

They'll just, they just, what Musk did in Tennessee, he'll put up a whole bunch of, um, natural gas fired generators that belch really toxic smoke into the poor neighborhoods that he went and took over to build for his big new super computer. So yeah, it's voluntary. They'll get you off.

We'll get, they'll get off our back. We'll get off their backs. Texas Just passed California to become the leader in scaling solar.

9,700 megawatts. Yeah, But they didn't put out a press release, did they? Oh, I'm sure they did, Because that's not something they're proud of over there.

We don't participate in any other grids. We like our own grid even. It's about the Texas grid is powered by solar for sure.

And you can't go any place, like all around Austin is farmland. Like, I'm going out to my son's town, which is a, a few miles, you know, 30 minutes east, it's farmland. But any place you drive in the back roads is all solar farm.

Huge. Huge. Well, not only solar farm, my my younger son used to work out near Abilene and all you saw there were the wind turbines for, uh, miles and miles.

Right? So solar and wind for all of Texas is bluster is what power is Texas. Let's see, Let's call this what it really is though, right?

Because ultimately a lot of those people are just gonna move to another town and somewhere else. So basically, you know, if I pollute the environment and there's no water and it's too damn noisy to live there, people are just not gonna live there. You know what Mike?

That used to work and they would say go west young man, or something like that. But there's nowhere to go. Where you going?

Well, I think too, Everybody's going to Idaho. So to Idaho Exactly. Hold up there.

One of the things I think is real interesting too is, is this something that can be solved by technology, right? So you can't just put a data center anywhere, it's got to be near centers that have the, the fiber connections and has, you know, a massive amount of people and it has the power infrastructure there to tap into. So the the problem with that is Why, why do you need a massive amount of people?

Because they're not gonna build the power infrastructure where there's not a massive amount of people, which is some of the whole Issue. Oh, so you need it for the power infrastructure, not the data center. Right?

But you need Right, exactly. But that if you look and if you look on maps where, like in Texas, the big chunk and the big build out is all up and down the I 35 corridor, which is where the fiber cable is, which for the internet, which we need that and where power infrastructure is. So that's what's driving up the prices of the existing data center, is everybody needs those buildings that are there, but they're running out of power.

But maybe this is something that technology can solve. We're getting smaller and smaller drives, we're getting more efficient ships. So maybe, I don't know if we can go fast.

So I, I I have, so I have an article that just, uh, I think it just got published right before we came on. Here's the thing, what about if I told you the future of AI may not be in the data center, that the future of a ai AI may be on that phone in your pocket. It may be on that Mac mini on your desk.

And that in spite of us badmouthing apple, crazy like a fox Yeah. May become the inference AI compute engine for AI on the edge. And that's where a lot of AI's gonna take place.

That's their long View. I think You're right. I think it's a view right now you can't buy Mac Minis right now.

Mac Studios are on back order. They're coming out with with models. Uh, you could run on your phone.

It turns out those M1 and better chips, the m threes, the m fives, because they, they have the neural, uh, processors and all these things and they're, they're perfect for running inference. Who the hell needs that billion dollar data center? And I could buy a $600 Mac mini, I think for certain workloads that works.

But the reality is gonna be much more distributed and there's gonna be more caching of local AI models at the network edge, and then there'll be between the edge and the big data centers in the cloud and they'll be everywhere. And it's not just running on your handheld because some of these tasks and workloads, if you try to run 'em on your handheld, then you're the require the processing power required would melt your hand. I remember HuggingFace came out last year and they were like, oh yeah, we're gonna do mini models, mini models, run 'em on your phone, run 'em locally.

I have not Heard anything against. We'll, we'll get, I, I do think that's the future. I I don't think LLMs are the, the dominant future.

Uh, well, so the foundational models in the, from the Frontier Labs, I think that's all great, but today people are already able to run models on Mac minis and laptops And so on. We're doing it here in the office. I bought three Mac Minis this week.

So there's, there's two things here, right? This, this, the blood of what's happening for data center acquisition, that's a real estate company problem, right? Yes.

Tech bros are investing in all those, but I mean, at the center of this BlackRock. So when we think about what that growth model looks like, my concern, and I think probably a lot of folks' concern is what does the control space look like for that? We also have data centers in space, right?

Where, you know, water cooling and, and these types of concerns that we have here on earth not gonna be the same problem because of the temperatures and operational capacity, bandwidth requirements, all of those things laid that we've already, uh, basically paid cost for. None of those things are inhibitors, uh, for data centers in space. And finally, I think agree with, with shimmy and also, uh, with, with this group, basically models aren't just gonna be cached on your laptop, on your phone or whatever.

We're gonna use this shared compute infrastructure that exists, whether it's in a data center with, you know, today we, we do this a lot with browsers already, right? JavaScript, all the other things. And so we're gonna use every available shared compute that's, that's presented phones, all devices, your tv, all of it.

Mm-hmm. And that's going to be the way of the future. So that distributed model, we're gonna go back and forth through this like we did when we decided that hey, we're gonna have 32 70 terminal emulators.

We're gonna connect up to this shared, you know, SNA mainframe, and then we're gonna have personal compute on laptops and desktops, and then we're gonna go back and I mean, it's gonna continue to go that way. Wherever the shared compute is, we're gonna use It's, and that's been the history. Can I Borrow a couple of CPU cycles to run this AI thing for a couple hours?

You got got, I've got 11 extra hours, Mike. I'll use that Time sharing. All right guys, we gotta go into the future.

Back to the future here though. And, uh, go to our third segment. AI pumping dump our friend Jack Dorsey.

Oh, he's the bad boy again. He's always been a bad boy Jack, hasn't he? Mm-hmm.

I guess, I guess you know that there's just an amazing number of people out there who just can't seem to read the room or whatever. But when everybody else got crucified already for, uh, blaming layoffs on ai, Jack Dorsey comes out and does the very same thing for block. But you know, I guess more people cottoned onto it sooner and just pounded on 'em.

But John, you covered this and a lot of people are just asking, what the hell were they doing with this number of employees in the first place? Exactly. Exactly.

So this is a live show. Can I say b******t? Yeah, you could say whatever you want, John.

It's an adult. We're gonna put the flashing lights and adult language thing on the top right corner. Oh, there you go, John.

Come on. I I know, I know this, this is utter, utter pile of b******t on his behalf. So Jack in his inimitable way kind of comes into like, he is not reading the room because he's kind of absent.

I'll just put it that way. And then I'll, I'll, I'll reference what I mean by being absent, um, in a few seconds. So he announces this mass termination, 4,000 of the 10,000 employees during a very strong quarter and a letter to the shareholders.

He says, this is not corporate letting, but it's part of this AI evolution and we're more efficient than ever before. We don't need all these people. Well, think back to COVID.

They, they were up to about 12,500 people. They were totally overstaffed. And by the way, they, they have a lot of, uh, excess, and I have to be careful how I put this, but they've held parties that have been documented that were in the tens of millions of dollars that involved recreational drugs and mid midget wrestling.

I'm not making this up. So there's Lot of shades of the Wolf of Wall Street, huh? Exactly.

And Jack is, in a sense like that he sly a fox, but he is also a wolf. So this whole AI washing argument that he made is basically all about, and I think it's, it's, it's more about, uh, the market cap to headcount ratio. That's really what it's about.

And in a sense, it's, it's masking what he should have done through management, his own management or his, his his, uh, missing and action management, which is the way he runs companies. That's why he has very strong CFOs and, and other folks around him because he pretty much leads them to run the company while he does his excessive behavior. Um, he, he does this, my fear, and I'm just going to, I'm gonna look back a couple years to what Elon Musk did post COVID where he started eviscerating X and say, we don't need this many people.

We can get rid of more than half, 75% of our employees. I think we're just seeing a repeat of that. And with AI being used as an excuse for mismanagement of a company in a woefully bloated workforce, and I think it's gonna happen.

He's opening the floodgates just as, as must. And I think other people are gonna follow. And they're not gonna care whether people like us are offended by this because all they really care about is shareholders and their bottom line.

I'm done. Sorry. I pretty much agree.

We're done running. I'm sorry I was. No, but I agree with a thousand percent of what you said.

I think anybody who thinks Dorsey cares what they think is woefully incorrect, he has shown himself to be somebody that does not care what other people think. Um, and just like a weird dude in general, like with all the fasting stuff, and I haven't eaten in days and you know, there was a movement in San Francisco where people are like, is Jack Dorsey okay? Like, this is not who I would describe as a stable individual.

However, there are people that, people that follow him, He's a guy, he's a guy who's like, he's an addictive personality. And I'm just gonna mention this because I used to cover, actually went on a trip with him to New York once. It was very bizarre.

This guy had a fascination, an obsession with punk music, with pizza, with architecture, with, I mean, I can relate to that at least. Like, I know, I'm like, like, that sounds like me. What the hell?

That's okay. But, but, but guys, lemme lemme, Where's the weird part here? I was just gonna, lemme bottom line this for you.

Are you blaming a tech bro for being a tech bro? Mm-hmm. You know what?

Of course the tech bros, this is what, this is why we in the mess we're in. Yeah, I think that's Yes. This Weren't, this weren't as as brazen about it as they are now.

They just don't give a flying F about it. Well, because they got the big protector back in Washington. Exactly.

Exactly. They bought and paid for that and they, they, they got air cover. What I think if we tie it back to what we were talking about earlier, like there's no, there's there, we don't really see any standards coming out of how we should handle ai.

How we should figure out where's the horizon, what we're trying to get to, and what everything should look like. Um, and when you see business leaders like this just dump almost half of their employees and say, figure it out. You got ai that just is adding to the problem.

And it almost extends the, the idea and the hype of what AI is and what it can do. Instead of saying, this is a fantastic business tool, we're gonna figure out how to make it a great tool for our entire organization. And we're gonna make, we're gonna do it in a way, you know, that'll feel bad for a couple of quarters and then we'll get it all under control.

I just, that's what frustrates me is it's such, we we're the examples of the companies that are doing this in a great way and trying to find that horizon and what they want their known good to be. So Let, let me play devil's advocate though. You brought up Musk at x He did, he cut it to pieces.

But have you seen a lot of Twitter outages or X outages or whatever you would call them these days? Do they have the same load that they used to have? How many people left?

Do they have the same amount of use? I, I don't think it's, I think it's the Level of engagement is I, I think it's down. I think Gina's right?

Yeah. It's, it's a little, it, it's down, but it hasn't fallen on its face and he's trying to make it profitable. Here's, here's the thing, and this is a malaise that affects our federal government.

It's affecting the tech bro companies. And it all comes from the same, the same mindset. Yes, you may cut it to the bone and even cut bone and things will get by.

We're gonna lose a little here, we'll lose a little there. But for the day to day they get by until something happens. Until something happens.

You cut, you cut health in human services, and then one day you've got a pandemic and now you got the cupboards bear, you cut fever and a hurricane comes. How about cutting foreign services during a war and then a war comes Cut, cutting CC and then going to war, you know, take it from the government to this, you know, cutting your engineering team and then an outage happens, uh, cutting your, your, your team at, at block and all of a sudden people can't, uh, do their credit card transactions. And what the heck is a, a small business supposed to do?

Things are okay until they're not. You could run things on a skeleton until the, you know, what hits the fan and then all of a sudden you really, you, you start seeing that those cuts you made were too deep. And Tell you two things about these kinds of cuts.

One is, hey, the people who got cut, we don't know if they're just the hangers on, on the back end of the party or if they're actually the good people. And then the worst people are still there. But I will tell you, if the good people who still are there, they're leaving soon because they're gonna be like, yeah, this is not stable.

I need to go someplace that is stable. And no matter how bad things get in the economy, the good people always find another place to be Agreed. I agree with you, Mike.

Mm-hmm. I agree with you. Mm-hmm.

Anyway, speaking of another place to be, we have another place to be. We're about outta time here today. Um, what a great discussion.

I mean a little, a little, some on the downside, but, you know, a week full of great discussions. Uh, Fred, Gina, and John. Mike, thank you so much.

Thank you for watching. Hey, we've got more text on TV immediately following this, and I do have my double shimmy says today. So stay tuned for that.

Live. I think at two 30 we'll be on there. Um, other than that, we'll, we'll be back here tomorrow.

Actually, I won't be here tomorrow. I'm going to Boston to visit my oldest son we haven't seen in a few months. So Mike, you'll, you'll be staring this ship yourself.

Well, I'm sure you'll have John or someone to help, but it's all yours. How many dents am I allowed? As long as you know, as long as you have the insurance, it's okay.

Alright. It's okay. It's your car.

All right. Until next time, this Alan Shimel. Thanks for watching.

We're out.