Google’s AI Shopping Protocol, AI Coding Advances, and Quantum Funding | TSG Ep. 999

Hey, everyone. Have you gone AI shopping yet? Hey, it's the latest thing you're watching, Textron Gang, welcome, welcome to our live Textron Gang.

For those of you who haven't had the chance to, uh, log in or, or, you know, jump in, or since the first of the year, we've moved to our live format. So it, we like it. It's, it's, it's, I think it's snappier.

It's more on time, you know, it is live TV or live video. So bear with us if anything happens. But it, it's a good time.

We've got a great panel for today's live Textron Gang. Let me introduce you to Chris Blak. Kate Scarsella.

Sid Nag. Sid, I always mispronounce your last name. Is it, is Nag the right?

No, That's fine. No. Okay.

And then of course we've got Dan o and Mike Ard, lady and gentlemen, welcome. So Mike AI shopping, right? I, you know, I, I remember when e-commerce was the buzzword.

Is this a whole new wave? A whole new generation? Well, the National Retail Federation conferences this week in New York.

So everybody was talking about this particular initiative because Google launched it. Uh, they've been talking a little bit about AI shopping before, but now they've introduced this protocol with help from Walmart and Tech Targets behind this. I mean, not, sorry, not tech Target.

Target is behind this, um, and a bunch of other folks including, um, uh, Shopify, which is probably their most important player in that set. But it kind of says that, you know, we're all going to let AI agents do shopping on our behalf. I don't know how much agency we wanna give an AI agent.

I'm a little cautious when I give my money to anybody that they're gonna spend on my behalf. And of course, I'm sure there are a million security implications. But, Chris, what's your take on this?

What are you think? And, and are we gonna let AI agents just go to the store first? Us?

I Like it depends if it's your AI agent or somebody else's, right? You know, like a lot of things we're changing. We're going from a advertising to intent to a transaction.

You know, I gotta put something out there. I'm gonna see it. I'm gonna want to buy it.

I'm gonna buy it. You know, now we're getting into the, I mean, this is Disney World, 1971. You know, the, the world of the future where your fridge knows what you want and orders things for you.

So we seeing, we're seeing the big players trying to navigate, be that broker in the middle. And, you know, from my my opinion, I think we'll do a lot of that. I think we're actually moving in a different direction.

I think it's more when your fridge really is ordering for you. I mean, if, you know, it's being run by the corporation is profiting. Yeah.

The trust level just isn't there. And I don't think that's the way ag agentic AI works at the personal and the household level anyways. So I, I think we end up in a, in a model, and we're starting to move that way where you, your fridge is actually yours.

You can trust it because it's not being run by an ad agency. But we're in that middle part right now. So this is the layers we see as the big players try to own that broker space where all the transactions will go through that.

I Don't know. I'm trying to figure out where's the AI agent that's gonna say no, because, you know, the conversation in my house kind of goes something like this. I go, Hey, we should get this.

And then my wife looks at me and says, no, we're not getting that. 'cause we can't afford it. So I, I gave Gotti ever has been having a great, him and Bruce Schneider have been ho hosting this, uh, series of, uh, prompt or GTFO, uh, challenges of really, really good people giving like short talks.

And the first one I gave was on, uh, AI that can say no. Right? And I said, my piece there, and in the chat someone says, you could just imagine if your AI could say no to you.

My response is, just imagine it can't, right? If it can't, if your, my AI in my house running my house in my fridge is unable to say no, like a lot of these commercial products are, that doesn't lead anywhere happy. So yeah, you've gotta be able to trust your agents to say no on your behalf.

Otherwise, we're just, we're just marketing fodder. But you know, Mike, 'cause you wanted to buy something extravagant or, or too much money to Chris's point about the AI agent working in your fridge, these are staples. The milk's empty.

Do I need more milk? Hell yeah. Do, do I, you know what I mean?

How often am I gonna say no, I don't want milk, I need cheese. I need, you know, staples. And so I don't want my AI agent shopping for a leather coat I'm gonna buy or something.

You know, that's very stylistically, uh, dependent. But for staples for the block and tackling the food, shopping, the same stuff that's on my list every week when I go to the market, man, I, I take that in a heartbeat. A heartbeat.

You, But you, Martin knows a lot of that about us, right? I mean, Google's, you know, if they, they've got the identity side of the salt. Everybody's got a a Gmail account.

They've got, you know, years worth of your transaction history. Every receipts going into your inbox. Um, they've got the payment infrastructure with Google Pay.

Uh, to me this really reinforces their moats around search, around, you know, commerce, uh, advertising. Uh, I mean, I think this is really allowing them to kind of take it that last mile where they've done the search, they've done the advertising pieces historically. Now they're really letting you kind of go all the way through.

Um, you know, I, I like that they're again, leading, you know, on kind of, you know, open source, uh, you know, infrastructure as well here, you know, donating, uh, you know, kind of the code, uh, out to, out to the community. They got a great set of partners on the retail side, Walmart, target, Shopify. I mean, I, I think, think this is a pretty big deal.

I think this is, uh, you know, gonna take a little while to get people comfortable with. But, you know, as long as you've still got your finger on the button on whether you're actually agreeing to buy this or not, um, I, I think they can really personalize this and, you know, make some strong recommendations. And, you know, to Alan's point, kinda automate a lot of the mundane that a lot of us don't have to deal with.

Um, you know, kind of the, the ref refreshing of the staples in the pantry every week, right? Those sort of tasks I like, I like the, I like the fact that, I like the fact that Google is actually proposing a standard to these things, right? With the ECP, because honestly, what, what also caught my attention as a side note is that Amazon was not part of this protocol development or this consortium, right?

Which is the biggest e-commerce, uh, vendor that you have in the market. But I like the fact that UCP is defining an open standard where you can do, you know, the whole entire commerce journey seamlessly, right? So in terms of sure, the, the, your payment schemes, your credits, your shipping information, just reducing the whole friction in the cart experience, right?

I think that's positive. What I don't like is the fact that it may turn out to be a cabal of a few people who are going to control the e-commerce, uh, space, right? I mean, it's like the oil companies getting together and fixing prices of the pump, right?

How you don't see price difference in the pump could go from 1, 1 1, uh, gas company to another. So I'm, I'm worried that they could potentially retain, you know, drive control of pricing fulfillment and the whole relationship experience and sharing information, the whole issues around privacy and, and all of that, that comes into mind for me, right? So is that good things around the ECP concept?

There are many negatives that I think we should be worried about. Well, I'm Clear how this will actually work, because let's think this through for a minute. Let's say that you have an AI agent as a seller, and it's designed to sell something at the best price possible for you.

And I have an AI agent as a buyer, and I'm gonna buy some, I'm gonna program that AI agent to go buy that thing at the lowest price possible. Where will these AI agents negotiate with each other? And won't they eventually just cancel each other out and call us for help?

Maybe, right? But, but you know what Dan and, and Alan said, right? You know, you know, Alan, you, you know, we, we think we can trust our fridge to buy milk for us.

But, you know, to Dan's point, if Google, Google hasn't let, let's say it out loud, these egg companies know more about us than we are comfortable with now, but they're still not capable of making the decision on what kind of milk I'm gonna buy. 'cause they don't know us well enough. So we, so we had to extend this to the point where we will let arbitrary e external sources opaquely hold literally everything about our our lives.

So we can know what kind of cottage cheese we like to buy. That's not the way it's gonna, but they Already know. They don't, they don't know enough.

They don't know enough. Look To Sid's point, you know why Amazon isn't in this? 'cause they already got it.

They already got it. I'm, I'm a Whole Foods shopper. Mm-hmm.

Do you know every day, oh my God darn Amazon Echo thing. Hey, uh, do you need Nancy's probiotic cottage cheese? 'cause they know that's what I've been buying the goods on sale one, which is they're competing the good is on sale.

Do you wanna get that? Instead of Nancy's, is it time for more organic kale? I mean, Amazon's doing this with the new Alexa plus already to a certain extent, but they already have that.

But here's the fundamental thing, guys. Dan, you touched on it, said you hit it head on. Chris, you bit around the edges.

We are gonna have to come, we're gonna have to come to an understanding with big brother, with big tech, with ai, whatever, that if we wanna reap the real, the total benefits of it, the maximum benefits of it, we're gonna have to get comfortable with giving up a little privacy. And maybe you don't want to, I think that's an individual choice. Chris, you may want not want them to know that you like organic skim milk, not whole milk, but From Wisconsin.

But you we're going need, you know, and this, I'm, I'm being facetious. There's, there's bigger implications, but we're going to need to each of us make a decision how much we're willing to trade privacy versus convenience. Oh, yeah.

That's always the, the, sorry, go ahead. Yeah, this privacy, more importantly, there's the, uh, pricing issue, right? Who is your best agent?

It's you not an agent, right? So I'll give you a simple example, right? The, uh, the other day, my, my carbon monoxide detector detectors had to beep, right?

So, uh, Three in the morning probably too, right? The whole I unit, right? So my first reaction was I go to Amazon and look at the prices unit.

It was 69 point 99, right? Then I, then I, then I, then, then I started, look who's actually making this product? So I went to their website and it was 55 bucks, right?

So Amazon is making almost $14 markup on their product. Now, if you let these guys, or form a cabal, whether it's Amazon, Shopify, Walmart, they're going to more or less be in the 69 price range, right? And if you, the agent is gonna just search their, their respective ecosystems.

But the reality is that I can be my own agent. Or maybe those vendors have their own agents on their website that give you a lower price. So that's an example where I think these kinds of things end up being a, being a sort of a gathering of interested parties that, that could, But, but Sid, let me ask you a question.

Does it make a difference whether it's you as your own agent, or you have your own personal agent that says, Hey, before you buy the Amazon or the Walmart, go on the web and find the lowest price for me, that's an easy task for an agent. There are no agents already. There's a, the Capital One, for example, has popups.

Yes. Uh, capital One shopping. Yeah.

Shopping right when I go shop. So there are those things already. So there's no reason why each vendor can, like a manufacturer of that, of that, of that, uh, uh, mono carbon, mono detector can't build its own agent within its own Sure.

You know, uh, website. And then many of those smaller vendors manufacture the actual carbon mono detectors work with these Amazons of the world to sell them as retail. Uh, companies could fund their own agent, uh, sort of consortium and, and find the best price there.

And completely eliminating these guys, because these guys are generally more predatory when it comes to pricing or the people like it or not, right? I mean, there are, you do get the convenience of shopping one, uh, one-stop shop, but the prices are not always the best. Right?

That's my worry. So anyway, I, I think for me, when I read these stories, it made me think of the 1984 Apple commercial, um, that famous one, right? Where you have the runner throwing the hammer.

And the reason I say that is because in 19, right, 1984, apple showed us what it looked like to break this monoculture. And today's standards like UCP are doing the same thing, not by controlling choice, but by finally making individual choice possible at scale. So for me, you know, personalization here isn't creepy.

It's respectful of time and intent. And as a person who can't match to save my life clothes, like if I had like girl animals that were digitized, and they could say, now You show your age with the girl animals. I'm sorry.

Well, but then this is, you know, to to to all these points, Alan, right? You know, I personally, you know, I am a, a a, a risk to person. My privacy and security levels are way, you know, wider than a lot of people's, I, I bought architect structures that everybody can fit in.

And okay, well, you're saying to be clear, I want these automated systems to show me the ads that I care about, because they, they know me well enough. However, my point right now is that the only way to get there in the past has been to get, let some huge corporation hoover up all my information. Because locally, I can't do that.

Now we can, and in fact, we can't have them, you know, have Google and Amazon know, say a thousand times more about each of us, what we actually know about ourselves. They couldn't even do it. It wouldn't run.

Now we need to take it back and have our information locally. And we can do this. I can have an AI running here.

I do the ax on my behalf knows me well enough, and it doesn't have to, I don't have to give all my information to Amazon. So that's where we're going. In the meantime, yeah.

I think all these big companies will do really well. They'll, they'll own that broker level in the middle, but the value they're adding by knowing us better than we know ourselves is going away. Yeah.

I wanted to shout out that comment that just went by, because it also says that there's not enough transparency to what's going on. 'cause a lot of this stuff is already happening behind the scenes, and there is a massive level of disruption. And to Sid's point, you know, a lot of these companies exist on that margin.

And if people are just gonna go buy direct from the manufacturer, well, what will become of those Amazons and ecosystems? 'cause the, uh, the whole ecosystem will just get flattened, right? I'll Buy direct.

No, I, I agree. But you know that in that particular example, when you buy stuff from Amazon and Amazon's making 14 bucks on it, no, Amazon's probably not making 14 bucks on it. Probably there's a third party seller who's selling through Amazon and paying Amazon for fulfillment or warehousing, and they're splitting that or what have you.

Yeah. Um, but you know, you're paying for the convenience. It's the market, right, Alan?

And then it feels like the ENT way could actually drive a lot more efficiency, right? Yeah, I think so. And That's agents And who wins Looking for best price?

And Dan, who wins on that? We do. Yeah, Absolutely.

Yeah. I, I think, I think I, I think the good, the, one of the good things about this, it, it's, it's very much like what H-H-T-T-P did for the web, right? It's a shared framework.

The whole idea of UCP is kind of analogous for shopping like what HTTP did to standardize the web experience. So UCP essentially enables retailers and payment providers to plug into an AI based shopping ecosystem with less engineering overhead, more flexibility, so people can, others can plug in too. It doesn't pre preclude others to plug in, but it's still, they're still worry about whether these big guys are gonna do something predatory, right?

But I like theorization and the plug and the, and the UCP protocol, the shared framework, that's a positive, right? Hey, we're about outta time for this. I'm gonna take prerogative to say the last word.

If you look at the history of e-commerce, there was a time where I thought, like my wife, for instance, we'd never buy clothing online because she's one of these touchy feely types who has to, she wants to feel the material, she has to see how it looks on her. Does it fit right? You know, I just buy it by size.

But lo and behold, my wife does a lot of shopping online. And Kate, to your point, I think all of us always wanted a personal shopper to help with the grow animal color coding and everything, right? So this look, get out, lead follower, get out of the way.

But AI commerce, AI shopping's coming, and it's here. Let's move on though to our next subject, if you don't mind. It's AI coding in 2026.

I thought AI coding was so 2025. Mike, what's the story? Well, the Harvard and, uh, meta are working on something they're calling Confucius.

And the idea here is that the AI agent will be smarter and will, instead of just working on the coding, we'll also, we'll start doing all the scaffolding and a lot of the other scut work that developers hate. And soon, I'd like to get your opinion about, well, what do you think is the current state of AI coding? I think we keep talking about AI as if it's a constant, but it seems like we're just rapidly advancing.

In fact, I'm having a hard time figuring out, well, do we standardize on tools today? Or are we just rip and replace them now when the next great advance comes in about two months from now? Yeah, I think the key word there is scaffolding, right?

I mean the, uh, I, I think the, or the whole meta, what's it, Confucius code agent, is that what they're calling it? Yeah. It really kind of demonstrates how scaffolding with things like, or for an agent, right?

Things like memory management, context protocols, execution structures, and data structures. And all of that can really materially affect the ability to operate reliably in a real world code base, right? Uh, because enterprises are not really ready in many cases.

So it underscores the enterprise readiness for Gentech applications, right? Which really demands, you know, rigor, uh, and not just bigger model. I talk about this a little bit in one of my, uh, LinkedIn posts around vibe coding, and then the whole idea of, or, you know, rational or reasonable learning, right?

Reliable learning, right? So I think when you get into writing code or the production grade, the split between what the agent perceives, what needs to be done by that particular code base, uh, versus what humans observe, right? In terms of user, the, the UX EX and the developer experience highlights really the emerging needs for clear observability.

So things like, how do I make sure that a code base of written by an agent is reliable? Uh, can I crack the metrics of his performance, right? When do I know it's going to go into an infinite loop, right?

What is it going to cost, uh, in terms, will it bring my iPhone down, right? Uh, or will crash? Will that application crash?

So things like that I think are really key. So that's what, uh, this whole Confucius code agent announcement of Meta is talking about. So it's really all about bringing the rigor, reliability, the actual, uh, user experience versus making it what we've been doing so far, which is a geeky experience of using all this technology, Google or Microsoft.

And so, so the idea is code development is not the end all be all. It's the actual experience, the measurement metrics of the particular performance of the code base, the application, user experience, all of that. So I think that's really the key here.

Yeah, I, go ahead. Look on, I don't wanna get her in here for a minute. Um, Kate, what's your sense of the maturity of these tools right now?

I mean, how much are you kinda looking at this and saying, this is, this is now, or is this something you're gonna keep an eye on but not use every day? I mean, where are you in your thoughts on AI coding and what's going on with all this stuff? Well, personally, um, right off the, you know, to answer your question, I'm, I'm for it.

Um, I am embracing it fully. What I like, and actually, what I'll say that I love is that with cybersecurity, one of the things that, um, we always asked in the very be very beginning in the early two thousands, is who's coming in to your enterprise? What are they doing once so they're, and how quickly can you prove it?

And what I saw with this overall block and the articles that we're looking at, um, from application to network, you know, at first when I was reading the Network, um, I and Shield, I, I sort of laughed to myself because always bring in the network people to slow down the app people, right? Um, that's something that, that we always saw. And, um, but wow, one password, right?

I mean, good for them. I, I, you know, they were out there at, um, field Tech Day, uh, in October, and good for them. I mean, they, they actually delivered, uh, very quickly on something that they were just talking about.

So, um, so again, what I see from an architecture standpoint, you have your identity, you have your network, you have your app. And I mean, so, so to to your question, um, I'm embracing it. I'm loving it.

And let's go, man. Let's, let's buckle. Let me bring everybody up to speed.

'cause you mentioned shield, which, uh, we haven't introduced yet, but Palo Alto Networks has a framework called Shield Out, which is, you know, helping people figure out how to use AI coding safely. And then while we're at it, we're also seeing some advances on the security side with, uh, one Password is working with Cursor to help protect secrets when you're using these AI coding tools. So there's a lot of rapid advances coming together at the same time.

I guess the one thing I'm trying to figure out, Alan, though, is, um, you know, this is a highly competitive space. So how do you used to be, you could build an entire company like Microsoft on the loyalty of developers, but I guess my question is, are developers just gonna be, you know, moving from one tool to the next and there is no loyalty anymore? You know, I don't know if there ever really, truly, truly was loyalty.

There was lock-in. And I think the, I think everyone wants to avoid lockin. I can put a, a couple of observations here.

I can't be the only one who's thinking to myself, when this Confucius tool tells me what it's doing or what it says, do I, do I say Confucius says, I, I just had a mention, but, but seriously, meta is this, is this what the billions of dollars that Zuck has spent, is this the payoff for it here? Because, you know, this is the first time I think we're discussing meta in terms of an AI Yeah. Tool set or something.

And Well, it's meta experiments on models too, right? Yeah. I, I mean, you know, um, I Think we did talk about a meta acquisition, I think while you were away last week, but Met is clearly becoming a bigger player.

I don't know how they're Wrong. Well, they should for the amount of money spent. You would think so.

And they are one of the, one of the giants, Chris, you got your hand raised, and I, I, go ahead bud. Yeah. So, so in this last year, in fact, this May of last year, you know, I haven't done a lot of coding in a long time, and I have generated, produced tens of, you know, software stacks with tens of thousands of files, millions of lines, you know, shout out to Ashraf or our CTO and, and millions and millions of documents, right?

And we learned a couple things. I've pushed all these coding tools to their limits and past, and, and A, they're fantastic. Uh, BI wouldn't trust any of them.

You know, c they're missing also, you know, we're, we're like, a lot of 'em, my commentary, we're pushing in certain directions where, where you can't get there from here, right? So, but you know what I'm finding really interesting about all that is 'cause we'll mark all that out. And right now the documentation is more interesting than the code.

You know, I really don't care anymore what the code is. You know, or our focus more is on documenting the system so that you can express that in whatever code is generated by what, whatever coding platform you're using, you know, whether it's, you know, literally the programming language don't care. Uh, the models don't care as long as it matches the, the intent.

I, we can now generate code and express code, you know, uh, one little piece we wrote, uh, the other day based on one of our folks experience, who, who did a lot of this as a, as a younger person, cracking software licenses is dead. No one cares. I will just generate your application easier than I'm gonna crack your licensing scheme.

But what does that even mean, right? So yeah, the coding, coding tools are going exactly where they're looking like writing code. People still need to do this, but that's not how it works anymore.

Yeah. What is, Well, what are the, Chris is right on, I think, um, Chris, I, I totally agree with you and you know, and right now we also talk about the speed of it, right? Which becomes so important when we're addressing, you know, starting up businesses and stuff.

We want speed end of the day. Give me speed, Right? I, I gotta to tell you something, today's Tuesday, tomorrow is Wednesday.

Most of us know that. Um, but tomorrow is also predict 2026, and we've got vir virtually almost the entire future of analyst team presenting at Predict tomorrow. Couple of sessions there that are right on here.

One is Mitchell Ashley session, kind of on the future of coding and stuff like that. Uh, Nick patients, actually, Daniel Newman talked about this as well. Look, what a great time for developers.

What a great time to be a developer. You've got so many options, so much greenfield in front of you, so much runway in front of you, and you don't have to be locked in. And you have, you know, all of this.

And that's professional developers. What about the rest of us, right? Chris, you don't consider yourself a professional developer, but yet we, we can contribute Dando, you're not a professional developer, but you, we all are, are, uh, you know, tingling around here.

Um, the, the important thing is, is that we can't think of just being a developer, someone who writes the code. That's the message. I don't wanna give it all away.

Come to predict tomorrow. com and register for it right now. Um, but the, the key is tomorrow's developers aren't necessarily people writing code.

They're managing the writing of code, whether it's vibe coding or, or some other agent processes that that's the job of tomorrow that we're gonna call a developer. And it's wide open for anyone out there who wants to be a developer. You know, this is probably the greatest learning tool in the history.

And I was talking again on Predict 2026, Daniel Newman's gonna talk about how he became a nuclear fusion expert in a month, right? Think about love. I love it.

Just think about that. So go, go to predict 2026, but if you wanna be a developer, learn how to develop, learn how to vibe code, learn how to use ai, learn how to be the, the director of the code, not the person who necessarily writes the code. I think, I think though, I think what Palo Alto is doing is spot on because vibe coding, I know it's highly hyped today, but there's, there's this, a serious risk of insecure code and malicious prompt injection, right?

So that's all the more reason you need things like shield to se do things like separation or duties. Human in the loop reviews and input out output validation and the security constructs that vibe injects should not by any means break the traditional security development controls. And so therefore, all this is cool, wipe coding, all that cool.

But you, you just have to have that shield, for lack of a better term to, to be that oversight functionality, right? So my coating doesn't go sunder. So that's, I think that's a message here.

You know, I, I thought you brought up shield because you know, know, I know that's the next segment, but all blurs together, right? You know, and, and so I've spent the last six years focusing on supply chain, right? How I have a software bill of material that's the inventory of some piece of code that I'm using somewhere.

And during that whole conversation, we were driving down from the, how do I have one and get it from me to one customer. And my focus has been about how do we do this on scale forever? And we talked this through and we literally, now I keep going this way, this big machine I have here that write its own code, they'll write a piece of code, use it and throw it away.

You know, not generating an bound for that. We need to be able to put this in context all the time, and we can now, right? You know, you could just put the rules in express the express, the, the, the adherence for those.

And that results in, for example, S bombs. But the fact that we're, oh, you know, without all this, we've been going through this process where folks like me have been looking at how are we developing code right now? Oh, nobody's keeping track of who was handling what code in the, in the CICD pipeline.

That's probably important, but we got away with it. We're no longer getting away with those things. So again, kudos to Palo Alto and Shield and so forth.

But I would argue these are primitives that need to be built into everything all the time. Not a special effort. When you're coding things, when you're agents are coding things, every bloody thing they do, they leave a receipt.

We don't say, Hey, maybe in this case we'll start adding receipts to our process. Build it into everything all the time. Fair enough.

Guys, we're over time on this one. I gotta move us along to our next, uh, next, uh, segment, the quantum imperative. Mike, what do we got?

Well, Congress has decided that regardless of what the president might think, that they're gonna start funding research and development projects. The National Quantum Initiative is part of that effort, and it's been on a lapse for the last two years. And there's an effort now that seems to have some support from both parties to put some funding into this.

I think it was north of $800 million. And wow, that sounds or used to sound like a lot of money. But I don't know, Dan, let's start with you.

'cause sometimes I feel like that was, this is the equivalent of grandpa giving you a quarter and telling you to go buy some ice cream. So, um, To that, yeah, I mean, bureaucracy is standing in the way of innovation once again. Um, I think it's interesting, you know, this is kind of a, a b part, a bipartisan, uh, you know, kind of reactivation.

Clearly they thought this was a good idea a few years ago, for whatever reason, they let it lapsed and, you know, now they're gonna victory lap as they, you know, kind of put back in place what was already kind of working. Um, and everybody had already seemed to agree to. So, you know, it gets frustrating at times.

We, you, you see Trump, uh, you know, calling for half a trillion, uh, dollar increase in defense budget next year. Um, you know, this does feel like pocket change compared to those numbers. And, you know, it feels like we keep learning the same lesson over again, right?

We did it with semiconductors, doing it a little bit with AI right now, you know, quantum next. Um, you know, it's a lot easier to get ahead and stay ahead than to loosely even try to catch up. And, you know, with Quantum being tension is so groundbreaking, you know, in terms of enabling, you know, novel, you know, in invention and innovation that, you know, you, you couldn't do without that level of compute.

Um, that this is the type of thing we wanna keep well funded, you know, build the secure supply chain, build the ecosystem, build the skills, um, that we're gonna need as this, you know, really becomes real over the next, next few years here. So, um, feels like a no brainer. You know, go ahead, approve it.

It's a big, uh, small drop in a big bucket. But you know what? I I, you take your victories where you can get 'em, I'm, I'm happy to see the Congress express a will to actually do something.

Yeah, it's not a 800, 800 million or 800 billion, not, it's not a lot of money. 800 million. It's not a lot of money.

But I will tell you this, these things have a way of like filtering down, trickling down, that's a bad word, but trickling down here in Palm Beach County where we're based in Boca, you know, they, there, there's an initiative under, uh, underway here to make this quantum beach. They're actually ruining a public quantum company right now with like $20 million in tax incentives and abatements and all stuff to move their headquarters here from a, from a higher tax state. Um, let's be clear, quantum, a a AI is in many ways the, uh, you know how for nuclear, for a hydrogen bomb, you need a nuclear fission bomb as the, as the detonator, if you will.

In many ways, AI is that detonator for quantum, right? Yeah. Combining those two is a one plus one equals five equation.

And so it's important. We know it's important. And Dan, you're right, it's easier to be ahead than to play catch up.

But you gotta start somewhere, guys. And, and quite frankly, something's better than nothing. Um, you know, maybe this is the first drop in the bucket that starts a cascade that we see, you know, smaller jurisdictions join in.

We see private companies, you, you know, encouraged by this to go spend money and do things. It's important. This, this is an important, important.

So I'm, I'm gonna be, Can I agree, Alan? Wow, I so much agree with that, right? I mean, let's get back to the two guys in the garage, you know, building, you know, a pc, you know, that's what we're talking about.

It's a signal, you know, and, and, and, you know, it's a signal. So yes, I Have a question. It's, and you know, and that's something we lose that I, specifically with ai, ai, especially at the model level and, and the level we talk a lot about here, the barrier to entry is so high.

It's such a big tech bro game. Yeah. That it's not the two guys, it's not Hewlett and Packard anymore.

It's not jobs and laws anymore. It's, you know, you need a hundred billion dollars table stakes anti up to, to play in that. But you don't, and and this is, you know, don't get me wrong.

We, you, we, we have big clients and big resignation states and all that. However, what we're seeing is that you can decouple this, and I've said this, you know, in, in, you know, uh, in the Middle East, we have now an entire city getting populated with ai, locally running on existing old hardware without anybody's connection to anything. And that is providing business opportunity.

They have nothing to do with AI or any of this. It's just better, faster, achieve information systems so that you can get your, your business done. And I think that's what we're missing.

I brought this up in the show 'cause 'cause we are who we are. We talk about the a hundred billion dollar investment in the big GP data stack, but it, it, it turns out, and yeah, as, as I say, I love this weekly conversation because I keep remembering the things we said three months ago and six months ago and so forth, and three months ago, two months ago, I was arguing these things. Now we're just doing 'em and they're going viral.

They're just spreading all over the place. Because if you wanna run a pharmacy and have better loss, you know, control of your, and, and I'll add this to this, but we talk so much about the individual rules. The rules that we're seeing developed by, you know, a sovereign AI system we're putting out in the world have to do with humans.

Because complying with your technical regulation about what you, that's automated, so it's all automated. The only rules and protocols that are interesting to businesses are when you go back to, to the back of the store, do these three things. That's an actual business rule.

Not if you're HIPAA compliant, fill out form X, Y, Z after a b, C process. That's, Yeah. I I, I think restore, go ahead.

I'm, I'm just trying to wrap my head around Quantum Beach. Does this mean like a grain of sand will be in two places at the same time and we can sell that beach property twice? Not, not once you realize grain is saying there, right?

I, I think, I think, I think first and foremost, uh, you know, restoring funding for federal agencies like OSTP, nist, NSF and NASA is great. The idea of cutting funding for those agencies for quantum research was a horrible idea. So I, I applaud that, right?

Uh, but to the, to the point of the AI connection, uh, uh, you're right, uh, Alan, it's really the trigger, right? That's going to drive quantum. But we should also keep in mind then Quantum for AI is not really about, you know, replacing classical AI or GPUs.

It's really used, uh, it's really gonna be used as a sort of a special accelerate. So if I may call it that specialized accelerator, right? So, you know, realistically, where can quantum really help, right?

It's best suited for things like I'd say combinatorial optimization, right? Not your classic model training or LLM inferencing. But, and it also helps in the areas of things like sampling probabilistic models and, and those kinds of things.

So there are very specific niche areas where quantum quantum can help, uh, drive the AI adoption and, uh, and research is needed for those areas to make AI even bigger things. So I'm glad that those agencies are being, being given the funding back. Yeah.

Yeah. Science is good. I would think that the AI winds up kind of being a front end to the backend quantum computer to make it more accessible.

So I don't have to know How to Oh, yeah. I mean, listen, as we get into, you know, real quantum applications, you know, they're all gonna be hybrid. It's gonna be that combination of classical, you know, uh, neural networks and AI compute, um, and quantum all coming together, you know, to to, to really enable these, these early wave of applications.

Very little you're gonna do in quantum that's gonna be done discreetly. And now the pure, like pure quantum, you know, I, I agree. And I think what there's gonna be a whole learning process there of, of how do you program that?

How do you, like how does this go in that tree? How does that go in that tree? How does it come back together?

You know, there's, there's gonna be just some basic research that, that hopefully is going on right now, right? Um, but It didn't really talk much about the participation of the commercial vendors. Like what's the role of folks at IBM, Google, Microsoft D-Wave, et cetera, going to be in this?

That's an interesting topic we should maybe cover, uh, in a, in some form of, well, I mean, To me it, that's the trickle down, right? You know, as, as it goes into all these agencies and all these academic institutions, you know, the, the technology vendors are, you know, really the ones providing hardware in that, you know, kind of early set of experimentation hardware that the, you know, those agencies and academics are gonna be used. And so, um, you know, it feels like there's, there's broad endorsement here.

It's like a great area for public-private partnership. Yeah. Chris, I think you're on mute.

Not sure really sure where Quantum takes this, but I see this sort of threefold. There's Moore's Law, we're all used to that. Then we get AI and we have semantic systems that understand human language and that that adds a whole nother uh, uh, acceleration of the, the trajectory.

And then we have quantum help, which is gonna speed things up more. And as a security person, like I said in a previous comment, we have gotten away with not doing things for a long time because we can't and whatever. I think all of those security holes need to go away all at once now and add quantum to that and that just more so.

But I think we're pushing ourselves one way or another into the kind of systems. I don't know, that may break us out with some of the cliches we've developed in the last decades, last 50 years. 'cause some of those structures are just arbitrary.

They weren't really good, but they were what we had in the meantime. But with this sort of power and this sort of speed, I think it will be interesting. At least we have plenty of time to talk about that.

I'm looking for a quantum powered ai autonomous robot. Get it. That's, That's, that's the, uh, triple parlay in DraftKings this week.

Right. Um, We'll have to change the way we think though, because quantum, by definition, in my mind at least, is a lot of things that are happening in parallel. And we're already seeing that with ai.

So, you know, a lot of what we built historically was 'cause, you know, our brains thought of things in a sequential fashion. So I just wonder if the way we think about things is just gonna fundamentally change. 'cause we're trying to mirror nature, which does all kinds of things in parallel.

So I'm, I'm looking forward to a whole new society, I think. Well, and that's the only thing I can say about quantum is quantum cryptography, really. Right?

You know, I have this, you know, I can break this big key my question and, and people ask, how do I stop that? My question is, why are you trusting one big key? To your point, Mike, nature doesn't, you know, I'm not alive because every single thing works all the time.

We have massive redundancy in all of biology, and we built all these corporate systems on brittle systems. Chris, you're you're right, Chris. Yeah.

Well, but I, I'll tell you something. You know, I saw a thing a while back, like when AI's talked to ai, the language they're using, right, is not decipherable necessarily the human language. I really believe that.

No, our brains may be hardwired. We're doing things in parallel. Like that is gonna be a huge leap, right?

It's gonna take another stage of evolution. But the AI way of quote unquote thinking the AI way of problem solving, I think it's probably more adaptable to that parallel processing, if you will, that quantum will enable. And so that's why it makes AI better.

Yeah. But we use the word parallax a lot, you know, there's, you know, looking at things from two directions every, a lot of the AI problems people talk about, again, it's, it's like the one cryptographic key. I'm like, why are you trusting one?

Would you trust one person? No. So yeah, you can build automated systems that take all the advantages of, of AI and just use the normal human system processing that you would put in a critical infrastructure, for example.

And to this point, maybe are happening inside biology because the feedback loops and so forth. But again, we have not, over the last 50 years, we rushed ahead built a lot of things at work, but they're brittle and fragile, a lot of individual points. You know, this era of, because of ai, and I'm, I agree with you, Mike, I think quantum goes that way.

I don't really know. But yeah, we obviously have to, we can't get more and more brittle and, and longer supply chains. That'll all all fall apart anytime somebody sneezes.

Yeah. And I think quantum goes beyond ai, right? When it comes to the broader impact to humanity, right?

Can accelerate scientific discovery and health outcomes, right? So enabling faster drug discovery, more effective treatments, breakthrough in, uh, clean energy, batteries, medicine, there's so many applications for quantum that, uh, in terms of research in these areas, it's, it has a humongous impact humanity, right? Which would take classical computing many, many, many years to do, right?

So I, we shouldn't forget that part either when it comes to quantum. So Forget, let's not forget that there are other countries in the world that are investing a lot more in this than we are at the moment. Well, thankfully though, as Ellen pointed out, it's something, you know, something.

So thank you. There is, it's the world we live in, but the world we live in. So we're outta time for today's texture and gang guys.

Um, Chris Kay, Sid Danno, Mike, thanks. Thank you for joining us. Hey, we had a few more comments.

Say we appreciate them. We're still kinda working out the kinks of, in incorporating them into the flow, so stay tuned for that. Stay tuned for Text Drunk TV immediately following Today's Gang.

com. Go check it out. It is a, toward the force of, of our futurum groups, uh, feelings towards 2026.

You don't want to miss it. Also added bonus, R-S-A-C-R-S-A conference, their analyst team is giving us their view of what they see big insecurity in 2026. And I'll be announcing the DevOps dozen winners.

So go check it out. Predict 2026 until, And I believe we have a lot of quantum computing stuff coming up this year too, right? We do have some quantum computing stuff.

Yes. Oh, so, but worry about that later. com.

This is all shovel. We're out.