Hardware’s Comeback, Quantum’s Reality Check, and NYC’s AI Hiring Boom

Happy Monday, everyone. It's Techstrong gang time, and I am excited. We've got a great lineup of articles today.

Great gang lineup. Lot of smiling faces out here, and my friend Stephen Foskett sporting a, what'd we call it? Noir?

Or moire? Well, I'm going to put on a fedora and go black and white, and I'll call it noir. No, film noir.

Anyway, welcome to our Monday edition of Techstrong Gang. Let me introduce you to our gang, and we're going to jump right into things because we got a lot to cover. First of all, the quintessential queen of New York, Yvette Schmidter.

Yvette, good to see you. As always, Alan, great to see my fellow New Yorker. New Yorker.

Have a little coffee. And joining us, another favorite person of mine, Dedicki-- Dedici, excuse me. Dedici Sanyaka.

Hello, hello. Dedici, you came back from war. I'm so happy to have you here, man.

Thank you. Thank you so much for inviting me. I love being- Oh, it's a pleasure ...

just being part of the gang, brother. Also, part of our Monday regular gang lineup, or whenever we can drag him onto the show, my friend Stephen Foskett. Stephen, good to see you.

And then- Reporting for duty. Reporting for duty, as ordered. And then, of course, anchoring it from the high atop the Harrison Mountains of Harrison, New York, Mike Bizard.

Hope everyone had a great weekend. Of course, the news didn't stand still. There was always seven-day news cycles, 24/7 news cycles.

What a time to be reporting on what's going on out here. Mike, you want to kick it off? We got a bit of an ironic story to start with.

Well, this is true, and normally we don't go down this path, but it's worth having a deeper conversation because, well, there's a new VC fund out there, and it's put forward by the folks at Andreessen Horowitz, and they are putting together a fund that's dedicated solely to AI infrastructure, AKA hardware. This from Mark Andreessen, who coined the term, Software Is Eating the World. So, Alan has a column pointing all this out on Digital CXO, and you should all check that out, but it does beg the question about, well, where does the line between software and hardware going to be in the age of AI?

And Dedici is our resident software engineering expert. So Dedici, what do you think is going on here? And how is all of this evolving?

Well, I would push back on the setup just a little bit. 1 billion against the 90 billion under management is in the rotation. Hardware was already 20% of their deal flow, and they announced what, 15 billion, if I'm not correct, in new funds back in January.

But nothing left software. Right? Nothing left software.

So what people are reading as a rotation in my opinion, they're two separate things. There's AI is compressing software multiples, and then the physical constraints got real. So those reevaluations and the build-out, it's natural.

They're not the same event, but it's natural. And, the build-out is genuinely a new thing. We've got the rack power going to five kilowatts towards that megawatt path.

It's not a software problem, but Stephen, you probably got better numbers than I do on this. 1 billion, that used to be a lot of money once upon a time. " It's not a lot compared to what else is being invested in this stuff.

I'm suspicious that they're investing in-- Well, I suspect that they're investing more in startups and less in infrastructure per se. So basically, the startup's designing the infrastructure and that makes sense because all this software, it's got to have a place to run, right? Exactly.

1 billion isn't a lot of money. That's almost what Andreessen and Horowitz give to Republican causes as political contributions. They're the second-largest political contributor.

Here we go. Alan- Okay, let me just say it upfront. That's what this is.

You want to know what this is? That's what this is about. This is about currying favor with the administration so they can get in on Starlink and the rest-- Not Starlink, whatever the hardware build-out places are.

Oh my God. Right? All right.

Well, if you're going to open that Pandora's box, then at the same time, that very same government is allegedly investigating Andreessen Horowitz for manipulating board seats. So how do these two things correlate? Can I take that one?

Yes. Just because the one thing that's really odd about that whole thing is the investments came first, right? And so it wasn't like you were investing in a competitor and you needed the thing.

There was no competition at first, right? The investments came, and then you became competitors. I think it's going to be a reclassification issue, if anything else.

That's one of the things that based on the things that we're building in my company, we're really keeping an eye on that. So yeah, it just feels weird. It is weird, but like I said, giving enough political contributions has a way of making that go away in this current political climate where evidently everything's for sale Mm-hmm.

Now, but that being said, Mike, there is another point I wanted to make, and that is Mark Andreessen was not wrong when he said software was eating the world. It did, in fact, eat the world, and every company became somewhat of a software company. And it's still true today, but what we're seeing is AI eating software.

Right? AI is increasingly the vehicle that's writing the code, that's testing the code, that's deploying the code. Yes, with human oversight, but AI is eating software.

I think the software ate the world, so little fish eat big fish, big fish eat... You get it? But the thing about AI, much more than software, software was always portable.

Right? You could plug it in, run it on any cloud, read, write once, run anywhere, all that good stuff they fed us. But AI, clearly, in its current iteration where it's all about the LLM, is tethered to these expensive hardware racks like nothing before.

Right? And so if you're going to believe that AI eats software, you've got to also buy into that AI, at least until we get off the LLM heroin, AI needs its plugin. So I'm going to take a different tact on this.

The history of IT shows that increasingly over time, software functions, especially middleware, wind up getting embedded into the processor core eventually. And you can look at things like CUDA, and it's headed that way and getting deeper and deeper into the GPUs. So is this Andreessen-Horowitz just making a bet saying that the ecosystem of the future is going to just revolve around these hardware investments, and then their software companies will be conveniently tied to that investment?

Mike, I think you got this bass-ackwards. I don't think so. I think the history of IT shows me that that's what happens.

Increasingly, stuff comes off of hardware and gets put into software. No, that's definitely not what happens. Let's look at virtual machines, and it gets embedded deeper into the processor, and the same thing is going to play out.

Yvette, am I crazy or no? You are a little spicy. So yeah, that's a little crazy.

But no. I'm focusing on the cash, right? Because we're talking about software getting eaten, but here's the thing.

Casado is still running software infrastructure. From my take, it's the capital that's chasing the compute bottleneck with billion-sized dollars commitment. That's from my take.

I think at the end of the day, it's going to flatten out like it always does. All I'm going to say is, remember VMware, that's now Broadcom, and what happened with that? All I'm going to say is remember that.

Just remember VMware and how that kind of transmuted into something crazy. I think AI is going to happen the same way with software, hardware. It's going to be a mess.

So remember who you get into bed with because you remember when your VMware, your license went one way, and then the next year they went a whole another way, and you couldn't get off. Just start thinking about that. I think that's what's going to happen with AI, with the software and hardware.

That's my take. " Exactly. That's where I believe it's going to go because that's exactly what VMware Broadcom did.

Like exactly. They had it all one way, and then within six months, when it came time for renewal, you were in a really bad spot. I was going to say a very colorful word, but I remember that this is a children's show, so I'm not going to do that.

But no, I really think that's the way we're going to go. And I do really think that people are forgetting because, again, you want to catch a squirrel, use shiny objects. They're looking at everything that is coming out here today- ...

and they're just taking everything without really thinking through, what does this mean? Should this change? All I'm going to say is SAP, now, you can't tie in with an API call with them.

They are rationing down on who can have access to their core. So just think about if SAP is going that direction, everything else is going to follow through. Again, that's my little conspiracy theory because I read the tea leaves and the notes, and I eat the popcorn, too.

So I really think that there's a there there that most people aren't thinking about. Mike- It seems like everybody's trying to figure out how to build a moat around something, Alan. Yes.

Is that what's going on? Well, they are, except let me give you the indispensability trap angle on this. And you know what?

September 9th is coming. That's when it's back from final edits, and I'll let you know exactly when we're publishing. But the indispensability trap tells us that this hardware feeding frenzy is almost entirely due to scarcity.

Right? And when GPUs aren't, or whatever the dominant inference chip or whatever is going to come next, is not as scarce as it is now, the premium on this hardware, just like premium on fiber optic lines in 2000, is going to go straight down to commodity. 1 billion go that I put into hardware that's now no longer scarce and has become commoditized?

Yeah, that's the cycle. It's always been the cycle, right? As soon as we have good hardware, we make better software.

Software arrives and everybody's writing software. Soon as the software becomes super redundant, people are going building the same code over and over again in other places. That gave the rise to AI because it was the opportunity to take what we already knew and compress it.

And so it's this continuous cycle of compression. And the question becomes, as we move from semantic interoperability to semantic orchestration, what is that impact going to have on AI? If we can reduce the cost of running AI before AI and this hardware market really exceeds and takes us to the next level, what happens next?

Yeah. I read an interesting post by our friend Keith Townsend, Steven. He says, "Dollar for dollar, you know the best machine to run AI right now?

" Jammed up. Now you're speaking my language. Right?

It's about $20,000. You could get it, I think, with 512 gigabits of memory- Yep ... and the M5 Ultra chips, dual M5 Ultra chips.

And for 20K, you got yourself a hot rod compared to some of the prices we're seeing on some of these dedicated AI machines from NVIDIA and so forth. I think the price of all this stuff is going to go up because we have seen that the AI companies are subsidizing usage or consumption, and they are slowly but surely moving towards trying to figure out how to get more bang for the buck out of the tokens that they provide. And so we are starting to see the price of the token comes down, but the consumption rate goes up, and then we're seeing them saying the higher-end models are a little more expensive.

And you put that all together, and I'll add one other component. I was talking to Steve Lucas last week from Boomi, and he noted training ChatGPT costs roughly the equivalent of a used car. Training the latest AI models are roughly the cost of an aircraft carrier or so.

How sustainable is that? Well, but does that include paying for all the IP you ripped off to train it with? That's probably where a lot of the cost is.

I think they consider that free, but that may explain why they didn't want to pay for the IP. Yeah. Nothing is free, my friend.

But yeah, there definitely is that. But I really think this begs the question of do you need to run AI on some monster LLM in some gargantuan data center that takes up half the state and sucks juice out of everybody's house like vampires, or can I run it on my phone at some point? If 80%, 70% of what I'm going to do with AI could be run on my laptop, desktop, phone, you know what?

That $8 trillion in data center costs starts looking awful bloated. I just want to say this quick. Please, go ahead.

Well, again, and I don't want to beat the drum about semantic orchestration, but what we're looking at as we're studying AI is what are the contexts in which it doesn't use a lot of tokens? And we're finding that as long as everything is deterministic, you get better results. But most of software development, most of software usage, is actually on the deterministic scale.

You always know as a developer what you're sitting down to write and what you expect it to do, just like the DevOps movement, and this is something that you and I talk about, Alan. Mm-hmm. We began to write things down because it was the tribal information about the greater thing that we were trying to collect.

Yeah. And understanding that those tribal pieces, those things that were our own identity specific to our location, helped us control that deterministic flow. And so we're going to run into that same process circling all the way back with AI as people figure out better ways to structure their data to get better results out of the AI.

And we're seeing amazing numbers in that space. And we're just going to keep pushing ahead and forward, and that's... I'm just going to say, hey, if you believe that there's a conspiracy afoot, then check out Watergate and start following the money.

DC, though, it's a pleasure having you on here. I love what you bring into this. But guys, we got to jump.

We're over 16 and a half minutes. Mike, we've got to go to section two for today, which is another exciting section. Yeah.

And this is just rapidly evolving, and I don't think it gets the attention it deserves, but things are happening in the quantum computing space, and IBM bought an outfit called HRL Laboratories, which gave them a framework. And I'm not going to pretend I understand the computer science entirely, but the qubit framework is a little more tightly integrated with silicon, so that they're kind of moving down a path where they're going to maybe integrate quantum and silicon together, and these two things will create some interesting synergies at the chip level that will eventually manifest themselves in software. But I know Alan's written a couple of pieces about what the quantum market is shaping up at, but let's start with Steven here.

What's your assessment of what's going on here? Well, not to be a little too on the nose here, but the quantum industry is sort of in a superpositional state of will they or won't they, right? We don't really know what's going to happen.

No. I'm actually being serious here. So, it's interesting because so much focus is on AI, to the extent that if you talk to quantum companies these days, which I have, many of them are actually talking AI when what they really want to be talking about is quantum.

Simply because if you don't say AI, you just don't get the call And so everybody's talking AI, but frankly, quantum has just as much chance of revolutionizing the industry as the large language models do. And we're seeing literally trillion dollars being poured into AI, but far less money being poured into quantum, but still a lot, and still a lot of very cool engineering happening. So, again, I don't claim to be a quantum computing expert, but I play one on television, and I did stay in a Holiday Inn Express.

So let me see if I can do it justice. So first off, what did IBM do? Well, IBM bought a company called HRL Laboratories.

Now, HRL is actually something very cool. This is Hughes Research Laboratories, which was established in 1960 in Malibu in a very cool building, if you're an architecture nerd like me. And this is the same place that developed the first fricking laser.

Seriously. It is a historic think tank kind of operation. It became part of Hughes and GM and Raytheon and developed all sorts of incredible technology.

Well, for the last decade or so, they've been working on quantum, and what they're trying to do is actually something I think that we can all get our heads around. You've heard about quantum computers being basically beasts, right? You have to super cool them.

They're really finicky. They're really hard to manufacture. They're really hard to use.

We're at the very early stages of making these things into a product. Well, HRL actually figured out a way to productize quantum computing. They developed what they call a silicon spin qubit- Qubit ...

which essentially takes the same conventional silicon-based etching and manufacturing processes that are used to develop conventional computers and traps some electrons in there and allows it to be used as a quantum computer, which allows it not to have nearly the same amount of refrigeration and finicky custom wiring and technical hardware and so on. And theoretically, they could manufacture these things in volume, and if they work, it could really revolutionize computing. So if you're IBM, which again, you may not know this, but IBM, one of the many product lines that IBM is involved in is quantum computing.

That's one of the things that they've been spending a lot of time and effort and money on. They're one of the leaders in quantum computing. They must have looked at HRL's work.

There was a recent article in, I believe, Science, which is the journal, and they talked about what they've been able to achieve, and I think IBM took a look at that and said, "You know, that might work," and bought the company. They liked it so much they bought the company. And again, I'm probably not the right one to judge this, but looking at it, that might work.

And it's really cool to see somebody focusing on something that can be manufactured and produced in volume and used at scale instead of these basically science projects that we've been seeing in quantum computing till now. Mm-hmm. There's a piece that I love Steve, because he talks about the historical aspect of Howard Hughes Lab.

But here's the thing- Fricking lasers. I know, lasers, but check this out. But IBM bought this because it was a lab run by Boeing and GM.

So it's that historical plus that context, and so let's just follow the thing. If IBM is positioned, or allegedly the leader in superconducting qubit, if they're the leader, they just bought this silicon spin qubit shop. So IBM just added a second horse to a race it's already technically winning with optionality on this qubit design, which actually scales.

So- Maybe it tells you something about the horse they're on. Yeah. But look what came with the lab.

Cryogenics, control electronics, packaging material science, and freeze cold plumbing, which works at scale. So that's the point I think that analysts not really digging into, and they need to dig into and hammer into some more. So it earns the same size room, but the smallest part of the puzzle, of that tech.

And what did we just say about software eating the world and hardware investments? This is hardware, baby. This is hardware with a capital H-A-R-D.

This is some tough stuff. I'm going to jump in, guys. It's a two-horse race, so ride them both.

That's all it is, really. One of them's going to win, and you're going to make money, a lot of money, regardless of which one wins, and it's a much- Well, you would think so, DC, but here's my thing. And you know what?

There's two other articles in here that I wrote. One is a whole big special report on the state of quantum computing right now in terms of commercialism and so forth. You know what scares me, though?

Yet another industry where IBM does all the legwork, goes through the jungle with a machete, hacking a trail, and then some other Silicon Valley or some other Johnny-come-lately just kind of says, "Thank you very much, sir," and runs past them to capture the market. Wasn't Watson supposed to be the AI, right? Was it- Wasn't OS2 supposed to be the operating system?

OS2. I have- I knew it was coming, PS. But you knew it was coming.

Do not forget the relational database. The relational... This is unfortunately- The storage array ...

and God bless, I have nothing against IBM, but isn't this their MO? Right? They do all this legwork, they do great freaking work, and then when it comes time to cash the check, someone else beats them to the bank.

I think we call that a national treasure, don't we? Yeah. To Steven's point, there are very few labs in the world that have the HRL, the Hughes Lab kind of panache.

Bell Labs was the crown jewel of AT&T. It invented the transistor. This is that kind of lab.

They invented the laser, as Steven said. Is IBM going to commercialize this? Because here's the problem.

As I think that DC said, they're doing a parallel track. Their lab just announced that they'd made some great cryogenic breakthroughs that allows quantum qubits to be in separate things and work together. But here's the thing.

Do you know how big the quantum market was last year? $500 billion. Now, there was a time where $500 billion was real money.

Used to be a lot of money. Yeah. Yeah.

But- 500 million will buy you 10 quantum computers, right? Yeah, but think about that. $500 billion market without a real product, without a working computer, without a true working quantum computer.

Where we are making progress is, and I'm happy to report from a security point of view, with the post-quantum algorithms and quantum-proof algorithms and so forth, that we are already getting ready for the Q day when and if these machines work. But please go read... I think the quantum one is on Techstrong IT.

That special report gives a great overview of the whole industry. So correct me if I'm wrong, but the one thing I do like about this new approach is that it involves existing silicon and the DC... Last time I checked, the pure quantum computers, there's no compiler.

So how are you going to write software? It's going to be hard, and this other approach maybe tells me I have something that maybe makes it possible to write the actual software. Just saying.

Maybe Claude can just whip one of those up for you. Yeah, you can just vibe code that. Yeah.

" And there's somebody out there that's literally going to type that in, Alan. Never mind. So I'll leave that alone.

It sounds like a $500 million business to me. Yeah. Now, seriously, you're right, Mike, and last I checked, it was still very difficult to actually get these systems to do any kind of real work.

I think we're still on the basic science of it before we get to the real practical applications. But having something that's manufacturable, and having some big pockets and some big push behind it, maybe we'll get there. Maybe Q day is coming.

There is a real challenge in that space in writing software because we really don't understand all the complexities of just using the quantum system in that way, but we're still aligning those with traditional processors that do let us try to use that and take that to another place. The things that fascinated us in our own research was the error correction systems and how that's used to resolve some of those issues as you do all of that entanglement. If you're looking at the gate controls, if you're looking at the standard way in which processing is handled within those systems, there's a lot of opportunity, but it's just that translation piece.

And the thing that worries me most about how people frame the market is they talk about a quantum chip, a CPU, but it's really all of the other pieces that you guys have been mentioning. It's the refrigeration- Right ... the hardware, the transition, and understanding what those transitions mean, but it's an old school analog system at that point.

And I don't want to dive into the engineering portion of that and keep that on a level that isn't really science-y, but it's going to be there because it's usable, but it's just understanding those constraints and making it work. It's not five to 10 years out anymore, though, and that's- Exactly ... the important thing, right?

Exactly. I do think both Google, IBM, are all calling for '28, '29 as the year we get a real Q day, as they call it. We shall see.

And all kidding aside, AI may in fact become a catalyst for quantum becoming real. It may, at the end of the day, really help push this thing over the goal line. I'm betting on it's not one or the other, but all the above at the end of the day.

There you go. All right, Mike, time to hop to number three. All right.

Well, as it turns out, there may be a mass migration underway that ISE hasn't taken notice of. There it is. And- What is happening here?

We improvised tonight. And then- I love it, go with it. But anyway, there's a report out from one of these real estate firms, I think it's CBRE.

I don't really know if that's an acronym for what it is, but they were pointing out that, at least from their way they count it, the IT sector in New York is now bigger than it is in San Francisco. And if you live here in New York, you are aware that Brooklyn is filling up with geeks, and they're all over the Naval Yard and all kinds of places. But Yvette, I know you live here, you've seen this.

What's going on here in New York? So just like I love this Alicia Keys, Jay-Z, I love that Empire State of Mind. So yeah, there's I would say yes, with an asterisk.

So the numbers, sure. Raw head count, I think New York sits at 394,000, and then Bay Area's 375,000. And this is 13 years that CBRE has been tracking this information.

But there's a gotcha, because you know I love math. So here's the math got you. On their own composite scorecard, CBRE indicates that the density and R&D depth of the total head count, San Francisco is still number one.

And then when you look deeper in the numbers, New York lands at number four, behind Seattle and Toronto, which makes sense if you think about it. Right? So two things from my perspective I believe happened at once that kind of made that flip for New York, is that New York added over 30,000 tech workers while the Bay Area, they lost 24,000 because of people being riffed, whatever you want to call it.

Right? So I think that's the headline grab, but when you dig into the numbers, New York is not the new tech capital of the world. You know what?

New York is the engine for Wall Street being the capital of the world, and they're hiring tons of AI talent around fintech, health tech. So a developer at a bank now is tipping that technical head count. So I just wanted to throw out that clarification.

New York is still my favorite city. I still get teary-eyed when I fly in and I see the skyline, but it's not the tech capital of the world. I keep running into startups that are all based in New York, and I've also seen another mass migration of security companies out of Israel landing in New York.

And so there is this kind of- And Turkish ... part that's certainly happening. And Turkish.

Yeah, and Turkish. It's Turkish companies. Yeah, because we can go coffee.

I can tell you all the Turkish shops, and you can listen to all the stuff that's happening there. But yeah. And the question then becomes, is this report a leading indicator rather than a trailing indicator and other things to come?

And I noticed Steven here in the sidelines says his own kid now lives in Brooklyn. Yep. Well, Queens.

They were in Brooklyn, now they're in Queens. But yeah, data science jobs and working with a bunch of data scientists. Most of their classmates who graduated from Northeastern, which is computer science, they all- Great intern program ...

moved to Brooklyn and Queens. Yeah. And they all work in tech jobs.

But interestingly, the tech jobs that I'm hearing about they're working in are very different from the tech jobs that I hear about starting employees going to in the Bay Area. So in the Bay Area, it's, oh, they're so lucky. They got to go to Meta or Google or Microsoft or OpenAI, or Anthropic, one of these kind of real tech tech companies.

In New York, no. We're hearing about people going to finance, a lot of healthcare, a lot of other industries. Security, you talked about.

It is a very different picture, and it seems like New York is more focused on using AI and data and tech to solve other problems or to get into other industries. Whereas the Bay Area is still very much focused on sort of tech for tech's sake. Pure tech.

Yvette, does that seem like what you're seeing? Yeah, so fintech, and Mike, this is going to ring a bell because you and I talked about this earlier this year, is so fintech leads in New York. I can tell you that because you go to any of these networking funding rounds, it's like 90% fintech startups.

Okay? So fintech is the largest share of the city's open AI startup roles are there, and they sit in finance with healthcare coming in. Healthcare and legal coming in second and third.

So that means that we're going to go back to what happened earlier this year with the New York Raise app, right, around AI. This law doesn't touch anything that's coming out of these new startups. Why?

Because they're not at the frontier model size. So that means the models that are deciding who gets credit, who gets loans, who gets the referral, is now bringing me something to where no one's talking about, and it's the people, is that these startups are building things that impact people that look like me, who may not have all the information that they need, who may not even be included in the data, and they're running these things at scale. And I actually went to a networking event and I ran into 10 fintech startup founders, right?

" And they weren't. They weren't at all. They weren't.

Like, five out of five. I asked five brand new startups. None of them were thinking about bias.

None of them were thinking about the person who sits at that percentage point. There's a name behind that person, and for me, that's scary for me because we're the capital here. They're building these platforms that don't include everybody in it.

They will when they have to. But look, what does this really mean? Harkening back to an earlier era than Alicia and Jay-Z, if you can make it there, you can make it anywhere.

And that's what we're seeing, right? Look, my very first tech company, we started in Long Island, but very quickly, Mayor David Dinkins had opened what they called Silicon Alley in Downtown Manhattan, and we moved to 55 John Street, and that's where I started my first... It's where we grew our first company.

I wound up selling it. com and all those high-end design shops, they were competing ... with the Valley in San Francisco and all of that.

And of course, San Francisco, especially the Valley, accelerated past that. But it's an interesting phenomenon. When I went to law school 150 years ago, there were a lot of people who were accepted to law school but decided not to go because they want to go to Wall Street and be brokers, because you'd get a $500,000 Christmas bonus and go buy a Porsche or whatever.

And so people were running to Wall Street. I've seen it with my kids and their peers, where a lot of people go to New York to work in finance, but they're not brokers per se anymore. They're working for commercial real estate, packaging up syndicates, syndicated real estate deals like a BlackRock does and stuff like that.

And here's an interesting phenomenon. Two of the kids who I'm friends with their parents, recently left their jobs because they were typical Wall Street young kids. They make good money, but they work 110 hours a week.

And what job did they take? They went to work for AI firms based in New York. They're fintech firms.

They're selling AI financial packages to hedge funds, family offices, the banks down on Wall Street, because that's where they perceive the money is. Right? And this is now three generations, four generations.

They go to where the money is, and they're seeing that's where the money is, and when you're talking money like that, New York has it. So- If you get on a subway in New York, you not see nothing but startup tech ads. That's all you see.

And it's - I was up there for Platform Con, and yeah, we took subways and stuff. I'm one who I love the drive from SFO into downtown San Francisco. Just- Yeah, all these companies you've never heard of, and you're like- The billboards.

You know? Yeah. The billboards, man.

They're great. But you know what? New York has their own version of that on the subways now, and I noticed it when I was there.

So, hey, as a New Yorker, I'm thrilled to hear this. I wonder how much of it is Brooklyn, Queens, Manhattan. What's going on in Jersey?

Huh? How about where in Jersey? Where in Jersey you from?

New Jersey is last. Let me just put it this way. I would say Baltimore is doing more stuff.

New York is doing tons of stuff. They have a commission, coalition. New Jersey not doing anything, which is super, super sad because the outgoing governor, his administration, their administration, they had incubators, and they were AI first.

Yeah, but it's just across the river. It'll be a trip. If you go to work in IT in a big bank in New York, the only part of that that's in New York is your interview.

The rest of it, you're working in Jersey City. Yeah, Jersey City- Or Connecticut ... was always big for the back office.

That's true. You know? That's true.

Mike, what about up by you in the wilds of Harrison? Has AI civilization reached up there yet? No, man, because IBM moved everything downtown, and so most of the people who work for IBM up here are now commuting down to Union Square or any one of their offices.

M. Pei built it up in Armonk or whatever- Yeah ... that was.

They can't even figure out what to do with that thing, and they have security guards there to keep the kids out from vandalizing the place. It's kind of sad. But another thing that did my heart good in this section, reading the article, was why are people doing the New York thing?

Because they need to be in the office five, six, I hope not seven days a week. But- Yeah ... the return to the office, that whole- It's hard in New York, yeah ...

I get stuff done in New York. That has nothing to do with it. It has everything to do with how the hell am I going to meet somebody if I don't go to work and hang out in the pubs with them after work?

There's something to be said for that. It beats the hell out of drinking at home alone. Well, you got to like...

Never mind. I was going to say something- ... really, really smart, but anyway, I'll leave that.

I'll just be quiet. I'ma be quiet. But it had me thinking maybe I could use this in my never-ending battle to talk Bonnie into moving back to New York.

Hey, man. We miss you. Come on up.

Come on home, man. I wish. I wish.

Anyway, hey, we're about out of time on the gang today. What a great three segments today. We had quantum computing, which it's sneaking up on you.

Don't let it sneak up on you. The city that never sleeps, I guess we won't let it go there either. And everything else going on here.

What about data centers in Manhattan, though? You don't hear much about that. Oh, the governor put a kibosh on that.

So- Well, I know she has. Right. But I'm saying, back in my 55 John Street days, there were some nice data centers right around downtown.

Mm-hmm. Yeah. But that was a different generation of data centers.

There were some old data centers that they tucked into places like Tarrytown back in the day, and all up here in Westchester. I almost bought that data center, an IBM facility- Oh ... when I was at Interline in the late '90s.

They built- But they were not so loud, and they were tucked into places where they didn't disrupt everything. And they would stand a nuclear holocaust. That's how well they were built to it.

They had everything on-prem from water to power. Wait, wait, wait. Until Superstorm Sandy hit, and I was literally, my team was one of the teams that we had to go up, walk up flights of stairs to pull down servers, because the water mark was getting so high.

I'm like, so yeah, Southern Manhattan There were data centers tucked in there, like whole floors. Superstorm Sandy, they nipped that for everybody. If they had a data center down there, it ain't down there no more.

Yeah. Well, we shall see. " Mike, I assume we'll see you tomorrow on Tuesday, gang?

We are. We're, of course, celebrating the Yankees-Red Sox series, so it was all good here in New York last weekend. Why do you have to do that, Mike?

16. Why? 16.

16 times. Wait, Mike. Shots fired.

16. Shots fired, Mike. Who was the Red Sox pitcher?

All right, I'm done. All right. Typical Red Sox fan.

He can't take his medicine. Oh my God. No.

Hey, you know what? We'll see you in the wild card. If you don't win the world championship, it don't mean a thing, because in New York- True that ...

either you're the A number one or you're not. That's right. We'll be back tomorrow, though.

But hey, as I always say, Monday to Friday, we're here noon every day live. tv, Techstrong TV YouTube channel, or our OTT channel on just about any screen you like to watch videos on. We've got Techstrong TV.

Stephen, Field Days. Yes. We got something soon?

Absolutely. We are getting ready. We're gearing up for Cloud Field Day, which is our next one, coming up here end of September.

Then we've got Networking Field Day 41, AI Field Day, Security Field Day. Networking Field Day, let me tell you, is going to be off the hook. It's going to be four days.

We're actually going to be getting bonus delegates in there because we've got so many companies coming in. So if you're interested in AI infrastructure related to networking, that one's going to be the one to go to. But of course, we're also going to KubeCon.

We'll see you working with our- Mm-hmm ... our friends here at KubeCon, at Supercomputing, at re:Invent, and maybe some other ones that haven't been announced yet. com for the schedules.

Absolutely. All right. All right, everyone.

With that, I'm going to end it. Again, thank you all, gang, for coming on here. We will see you.

This is Alan Shimel. We're out.