Techstrong Gang – May 16, 2025

Hey, everyone. I love the smell of m and a in the morning, you are watching Textron Gang. Hey everyone, it's Alan Shimmel and Happy Friday.

Wow, this week's going out with a bang. All of a sudden, the deal makers are coming alive. That's what we're a little boost in the stock market will do for you, right?

All of a sudden, everybody's excited. Everybody's looking to spend some money, American confidence at its best or worst, or whatever you wanna say, but we've got some interesting deals to talk about. I also want to talk about a conference I was at this week up in Orlando.

We've got a core gang today, just three of us. Got people on the road, we'll get right to it. First of all, not a topic, golden Gate Bridge, but he is back home out in the valley.

He's feeling relaxed. Our, uh, Silicon Valley editor, John Swartz. Hey, John.

Hello. Um, yes, I am a little bit north of San Jose, so I'm kind of right in the middle of Silicon Valley, and, uh, I grew up here and it's, uh, yes, nice to be home. Yeah, I grew up in San Jose.

It's nice to be back in, in the cradle of Silicon Valley. Absolutely. Good to have you on.

And then, well, he's not where he grew up. He's probably about 20 miles north of the Bronx, but he, he's, it's close enough. We call it upstate on Long Island.

Um, he's our Chief Content Officer that's sporting a Yankee shirt, Mike Ard. Yeah, it's a big sports weekend. We got the Nick Celtics game tonight, Friday evening, and then it's Subway series.

Matt's Yankees all weekend. And I will be at the game Sunday. So, Good for you.

I I need my Knicks to win tonight, Mike. I don't want to go back to Boston on Game seven and without Tatum, I really should win this game. We'll see.

Anyway, though, sports is probably a on Textron TV later. Right? Now, let's focus on today's stories, Mike, as I said in the, you know, I love the smell of m and a in the morning, and there must be something in the air.

What do we got? Well, m and a is a sport in the Valley, right? So we got, Databricks is trying to buy Neon for a billion dollars.

Neon is a provider of a serverless database platform, which kind of fits into their whole, uh, data, I guess, I don't know what we're calling them these days. Lake houses, whatever, warehouses, whatever. That seems to be the vogue term of the day.

But John, I know you wrote about this and there's something odd about this particular deal. So explain. Well, so Databricks and announced its intention to buy Neon for a billion dollars.

And just to put it in context, within the last year or so, Databricks has made $3 billion acquisitions and in this area, so Databricks intends to combine neon serverless relational database with Databricks data intelligence services so that the customers tend to deploy AI agents more efficiently. So this is basically, um, a big push we're seeing in the m and a activity among companies buying AI startups or companies that can help them make it easier for AI agents to work together, because this is becoming a growing, growing concern among, uh, enterprises. So essentially, developers and AI bots use this, uh, neon's cloud-based database platform, which is based on, please, please, uh, uh, bear with me.

I cannot pronounce it. Is it Postgre Post Grays sql? No, it's Postgre.

Okay. Postgres. It's one most popular databases in the world.

So yeah, it's, it's been around forever, right? It's been around since maybe the eighties. So this would be built, used to build apps and websites.

So Neon would presumably function as the underlying database for customers that create AI agents with data they store in Databricks platform. That's the, that's the idea. Um, the, I just wanna mention the two other, I wanna mention just the two other deals.

So last year, uh, Databricks, I spent nearly $2 billion for Tabular was the data management startup. 3 billion for a Mosaic ml, which is an open source platform for training large language models and deploying AI tools. So it's, it's just the latest activity, um, by Databricks, which has also just raised a huge fund, one of the biggest, um, investment funds in the Valley.

So they're putting some of that money to use by buying or snapping up these companies. All right, I'm gonna start on this because first of all, I have a serverless database, and just because somebody threw the phrase AI agents into the press release does not make this an AI agent move. It is basically a serverless database that drives into a data warehouse.

Nice piece of tech, been around forever, not exactly net new technology or a net new concept. So I'm already, you know, we're already started and I'm already sick of all that ai Well, Mike, it ain't got ai, it ain't got that swing, Right? And then the Right, and then the second thing about this, and I, and I'm not saying it's not a good deal or an important deal, but a billion dollars under what math, what valuation, who stood up and said that this thing is worth a billion dollars Because You don't need, you don't need math.

It's strategic. I'm, I'm thinking, I'm thinking this is, this is an Elon Musk like acquisition, right? You know, a bunch of guys sat in a room with Inspire everyone and decreased the valuation by 75.

Yeah, Exactly. EE exactly. So there's nobody out there who says that Neon is worth a billion dollars.

I just don't See Well, but in today's world, and you slap a little AI on it and a little chacha agent here, agent there, and it's a bargain at a billion. But, you know, I agree with you. It's, it's serverless Postgres.

But here's, I do think my favorite Yankee fan that you're missing one key piece of this one, this is about orchestration for a agentic ai. That's what this is about. I'm gonna talk more about it when we get into our, uh, thing on the automation, uh, anywhere.

But the race is on, not just to come out with Ag agentic AI to come out with AI empowered agents. And what's the difference between an AI agent and an API call, I don't know. I think it depends what PR agency you spoke to, but nevertheless, people are realizing that we're gonna have all of these agents running around here, right?

And we need, we need an orchestrator. We need something that's gonna manage these agents. I don't doubt that for a second.

Neon is server litigation. Engine Neon is Neon is a database access serverless. Well, no, But they, they're gonna make that, what did, what did our friends at Service Metrics call it last week?

The control tower. The Control Garage something. Oh, That was ServiceNow.

The I Control tower For it. You know, I totally agree with the requirement. I'm just saying Neon doesn't do that.

Neon is, well, They, I, I think, you know, don't, don't, don't tell me where I am. I want you to, I want to tell you where I'm going. Uh, I think that's what this is about, is that they, they have I ideation of using a serverless Postgres database that could then run, you know, serverless anywhere, um, to be the, their AI orchestrator, their AI garage.

The other part of this that's also kind of nonsensical is Postgres is a relational database for structured data. And everything that drives AI is unstructured data. So again, you know, people are throwing around AI like it's magic and, but, But they're not, I don't think they're looking to host LLMs or anything in here.

They just wanna host a, they just want an orchestrator. They just want a database that scales up and down on demand. Full stop.

I, I agree with the need, but all this other nonsense is just that nonsense. Well, I mean, we live, I mean, everything, everything is, is through the prism of, of ai. I mean, this, this is what they do.

It's all about, and Alan's right, it's about orchestration of AI agents, and they're gonna, they're gonna twist and mangle and reinterprets what they're doing to, to appeal to the market. And I mean, for all, was it neon is revenue, is is minimal. Um, this is almost like the, Yeah, no, this, this wasn't a multiple deal.

You know, this wasn't based on revenue multiples. Here's the deal. I don't even, I think, look, data Databricks basic business, they're real business that allows 'em to do all these billion dollar deal deals.

You referenced, John is, they're a data lake manager, right? They manage data. And using AI empowered agents to manage data is not a dumb thing to do.

It's probably a smart thing to do going forward. So if you are gonna develop all these AI agents, and some of these AI agents are gonna be ephemeral, they're one time users, they're like disposable. Others will be persistent, and you gotta house them somewhere.

So it makes sense to say, okay, we, what, what database are we going to use to house these agents and to manage these agents now? But let's get away from that a second. You mentioned John, I Postgres, you were a little unsure how to pronounce it.

Yes, I've been following re market. You know, at one time there was MySQL and there was Postgres, two more popular open source databases you never saw, right? The both of these were huge, huge.

Now, MySQL managed to get Larry Ellison White to take his checkbook out. Martin and the rest of the guys over at, uh, open, uh, MySQL made out pretty damn good. My SQL's still a force in the market.

But you know what? If you watch what Postgres has done since MySQL acquisition by Oracle Postgres filled that vacuum of people who said Oracle and open source databases are, they don't fly, right? That that's oxymoron.

And so the Postgres ecosystem blew up. Neon was one of those players who basically had their own Postgres distribution. Their, their shtick was it's serverless.

There are a lot of Postgres distributions and a lot of companies that make their money hosting Postgres servicing Postgres, you know, offering it as a SaaS kind of solution, or databases a service. Neon's one of those in, in, in, in serverless. So look, to get a billion dollars for basically someone who, who is porting an open source database onto a serverless platform, I don't care what their revenue is.

Bravo, you know, Bravo to the neon people, whatever they sold, whatever bill of goods they sold these people about agents and AI and anything else, God bless them. A billion dollars is a billion dollars. So here's the part I do like about this deal, right?

We've been talking about trying to connect databases and DevOps forever, and it's always been a disconnect. But if the database is serverless and you can call it through an API, then it becomes a lot easier to manage the database within the context of a DevOps workflow. So I think that's gonna be kinda one of the bigger benefits of that.

And is somebody gonna put an AI agent on top of Postgres and make it easier for me to do that? Probably. But you know, that's not quite the same thing as saying I'm going to be the center of the AI agent universe.

I don't know if it's gonna be the whole AI agent universe, but I think it's gonna be data brick. Every one of these companies is developing multiple agents, multiple agents, as I said before. Some will be disposable, some will be persistent, but, and they're not, they're not like a robot that does 15 different tests.

They're, they're, all of these agents are basically, like, they do one thing. So Databricks may, yeah, Databricks may see them selves, uh, you know, developing dozens of different agents for dozens of different tests that you do as part of your data management. And they'll, and they'll, you know, they'll, those might be in Postgres.

I'm sorry, go ahead, John. Oh, no, That's okay. Um, ServiceNow kinda hinted at that when they announced the, uh, AI control tower.

They basically said is as you, as you alluded to, uh, Alan, there are gonna be agents that are gonna do specific tasks. Maybe they are temporary tasks and they're gonna be slotted in and out and be popped in and out. And those who don't, and, and the ai AI agents that don't perform as well as other, uh, will just be plucked out and replaced by something else.

Look, you know, how, what's the lifespan of the average container in a cloud ative de uh, deployment? Seconds, seconds. They, they're, they're spun up and spun down and spun out, and new ones take their place.

They're ephemeral. And that's why you need something like Kubernetes to orchestrate and manage these things. It's the same thing here.

I think it's gonna play out more like this, the agents that ServiceNow is creating, or Salesforce or any of these folks, is gonna have to invoke a large amount of data to go take whatever action they're gonna be. And they're gonna go get that data from Databricks and they may talk to another agent to pull that data. And that's gonna be crucial because you need to do that at a, at a really low latency.

'cause these processes are gonna be running in near real time. And so that part of the thing makes a lot of sense to me. I just think, you know, the database company should just stand up and say, we need a lot faster databases instead of the a AI wash and everything they Do.

Well, yeah. Thi this, I think going forward, Mike, I think we're gonna see more instances of this. We already have.

We're just gonna see an acceleration of deals and they're gonna slap the AI moniker on the deal regardless, you know, what the underpinnings are of that deal. And it's gonna only, I think it's only gonna escalate This is, this reminds me when, you know, CRM was the rage and somebody came out with a CRM laptop, right? It was just easy, right?

It's just, I mean, it's history repeating itself really. I mean, in terms, and also in terms of these deals, in terms of the money, the, the, uh, dollar signs, uh, allotted to some of these deals. I mean, some of 'em are just kind of fictional in my opinion.

I, I, I always go back to Stargate. That just to me, they're having problems raising enough money for that to even reach one 10th of what they want to re what they wanna raise. Absolutely.

So how much of this deal, do you think Databricks is still private? Is that the deal? Right?

Yeah. Value. I think they raised market.

Yeah, they just raised like $10 billion. So they're valued, valued like at 62 billion, but they're still private. Oh yeah.

No, they, you know what, I, I'm not a hundred. I'm now all of a sudden I'm thinking there was a Databricks IPO at one point. Lemme just gonna give a quick look here.

But I mean, I, I think that they've signaled that their intent is to go public, but I wonder if they're gonna be like the next big rollup vehicle in the valley where all these startups are just gonna get rolled up into Databricks and then they all go public, quote unquote, together. Well, you know, as long as the stock market is as choppy as it is in the IPO market doesn't appear to be open to, to tech companies, one may say, you better off private if you're, you know, and use, use your, you got an outrageous valuation, right? And use that, uh, use that as currency to, to buy these companies, right?

Right. Was it billion in cash or a billion in stock? Did they say?

No. And you know, the, the, the thing that's interesting too is that, um, a lot of these startups, this is your exit strategy, right? In this case of neon, this is your exit strategy.

Oh, sure. I mean, you're neon market is, is, yeah, it's a no brainer. Like the IPO market is just not moving closed for, for tech company.

It's not closed. So it's for both sides. Yeah.

So for both sides, look, If you would've told the founders of Neon, if you would've told the founders of Neon seven, eight years ago that, Hey, all this messing around with Postgres and, and a serverless version of Postgres is gonna get you billion dollars, they would've ask you to pass it. Oh, pass over whatever you were smoking. Let's face it.

You would've told them, and by the way, that billion dollars is gonna be based upon your capabilities around AI agents. They would've looked at you like you were crazy. But that's the world we live in today.

So congratulations to them, as I said. All right, roll up in database aisle eight. Yeah.

The, the year of AgTech AI cash. So it strikes again. But let me, let me just strike one more familiar tone that I always take, which is the lesson I learned from Brad Feld, if you're not in the top three, get out.

So if you're in that Postgres database market, here was number one neon. Just hit it for a billion bucks. Take if you are, you want to be number two or three or get out.

That's where you are. Let's take a break here on the gang. We're gonna come back and, and we're gonna talk.

Are we talking Automation Anywhere next? We are. All right.

I'm excited. I hope you're excited. We're watching Text Pro Gang.

Hey folks, we're back with one of our field reports. Alan was up in Orlando at an event hosted by Automation Anywhere. They're talking about the future of what we used to call robotic process automation, but it's got a whole new spin and a whole new vibe.

Alan, give us the deal. Thanks, Mike. Yeah, you know, first of all, I love going to events in Orlando.

'cause we just get in the car, we drive up there, it's easy peasy. I I drove up Tuesday, spent the day, spent Wednesday, stopped in on some of our fu pals who were over at a, a a, a click connect, uh, event at home late last night. But Automation Anywhere, look, for those of you who aren't familiar, automation Anywhere is a 20-year-old company.

You know, I had a chance to sit down with their CEO, Meher, Meher Shukla. And, uh, he's co-founder and CEO he's been obviously there all 20 years. Meher has had several exits prior to starting Automation Anywhere.

But this guy's truly a visionary for all intents and purposes. Automation Anywhere invented the robotic process automation, right? We, we had business process automation BPA, they invented sort of, or at least put it on the map, RPA and, and the, the B-P-A-R-P-A market.

It's kind of aligned with that whole low code no code thing. It, it's about automating processes, but now it, it, it, I don't know how many of you remember the old TP commercials when we were little, put a little STP additive in the gas tank, and boy, you, all of a sudden your car's a muscle car. Well, you put a little AI agent AI in your, in your RPA and all of a sudden you're driving a 1968 GTO with a six pack sticking out the hood.

You know, this thing takes off like a rocket. And and that's what they've done. They've re you know, credit to the Automation Anywhere folks, they've reinvented themselves.

RPA is, so 1999 or 2009 or whatever, it's a P ag agentic process automation. And I remember when we, I remember when we talked about RPA was like the next great ai, but what happened in time was that people discovered that it was really good for automating a, a closed loop process. But everybody had a process that had, you know, more exceptions than rules.

And then the RPA stuff started to fall apart. Gen AI though, allows you to handle the acceptance. Uh, You hit it.

Now, all of a sudden, I could deal with the exception, I could be a little more autonomous. I could bring a little more intelligence to this. And, and that's exactly what it is.

You know, I, I, in addition to Mahar, I spoke with, uh, ADI Kudi, who's their CPO, chief Product Officer. I spoke with, uh, Micah Smith, who's their chief, uh, community developer community. And then I had a chance to speak with a key customer of theirs.

A company called Alight. Alight works for many, many health insurance companies. You know, when it's open enrollment, we've all been through this, right?

It's open enrollment season, season. You gotta go click on, you know, what, what coverage you want, what dental you want, which plan you want, what this is what that is. And Mike, that's exactly the closed loop type of, of, you know, business process that was made for RPA, right?

Another one, I have friend Martin Logan when he was at, uh, guaranteed rate mortgage in Chicago doing a mortgage application. Also kind of that closed loop that's perfect for RPA, but now with a PA, you know, and, and, and all that that brings, this is opening a whole new vista to how we can automate work. And you know, Maha has a, a, a great vision for this about freeing a little star trekky, but freeing humans up to do things that humans like to do and are worthwhile and are gratifying.

Um, here's another, and it goes back to our previous discussion here, guys. They are creating something that they are calling straight out, calling an orchestration layer because they say and told me this, and their chief product officer told me this, they don't care whether you're using some agents from Salesforce as well as some agents from ServiceNow, as well as maybe some agents for Automation Anywhere and agents from any other apps you're using as part of your business processes, right? All of these agents are gonna need to be orchestrated.

And, you know, is ServiceNow going to, ServiceNow is gonna orchestrate all of the ServiceNow ecosystem agents, right? From our report last week or earlier this week. Salesforce is gonna have its own stable of agents.

Automation anywhere says we'll manage any bank agent, right? Because no company's gonna have just one company's agent. We're all gonna have dozens and dozens if not hundreds of agents.

And so someone has to orchestrate that whole thing as it relates to business processes and as it relates to business process automation. And it is big, I think, I think there are, I'm, I'm think they're honest. Something I'm becoming a little dubious of all these claims where somebody says that they're gonna be the orchestrator, I think that there will be multiple orchestrators and gateways between them that hand off different processes.

'cause you know, everybody wants to be the boss, but they can't all be the boss. But, you know, nobody's gonna let one company become the dominant orchestrator. So I think we're kind of kinda, Do you think anyone's gonna let one company do it any more than anyone Let Kubernetes do it, right?

But yet Kubernetes became the dominant orchestrator in cloud Native. Yeah. But most of what people are running still is monolithic apps that Kubernetes has no control over whatsoever.

So, you know, Oh, Mike, um, you didn't get the news. Everyone's modernizing because Mware licensing. Yeah, I, I, I was just up here in New York on that is the separate topic altogether.

But I asked somebody, what percentage of your apps are cloud native? And they went 20%. And what percent are mission critical?

And they went 2%. So There, there is that, there is that. Um, but you know, nevertheless, you, you look at this company 20 years old, they, they kind of have, are reinventing themselves right before our eyes here, you know, jumping, uh, as we talked about later, if it ain't got that ai, it just ain't got the thing.

And, and so they, they clearly are bringing AI into this. And, and in their particular business case, it's not pie in the sky. It's not slick selling.

I, I do think that AI in the form of AI agents especially, um, really are the STP fuel additive to RPA. 'cause it gets you out of it. It allows you to have a wider swatch or of, of of business processes that you could deal with that are in all that closed loop, you know, A to B2C to D kind of thing.

If you gotta branch out this way, branch out that way, you know, whatever the, this sort of intelligence really makes a huge difference. You know, what I find interesting too, and maybe, I don't know, John out in the valley, usually those startups that say that they're gonna be the next greatest thing around this AI agent orchestration. And yet the legacy players have all seemed to like close down their flying pretty quickly and say that, you know, they were gonna be the orchestrator.

So I don't know, are you hearing any noise in the valley about orchestration as an for agents as kind of the next startup? Or is it just something that Yes, I gotta be established to drive? Yeah, that's a good, that's a really good point.

Uh, Mike, I I was, I'm starting to hear that. Yes. Um, so yeah, what you have, as you said, and, but Alan said you have Salesforce, ServiceNow, Informatica go down the line, right?

They, they all, they all wanna be the end all, be all orchestrators, which I believe no enterprise or a few enterprises in the right mind are gonna be beholden to one company for that responsibility. So they're gonna be looking for multiple types of, or orchestration and maybe somebody who ties it together. And I think that as you're hearing murmurs that in terms of orchestration of like, if the possibility of a startup or someone filling that void, which is an incredibly important void, which will be the next wave.

I mean, we're going through all these little mini waves and, uh, and AI in general. Now we're in the ag agentic wave. Now we're drilling down the orchestration, but it's not gonna, it's gonna be one company offering all the solutions.

I I, I think there is an opening for someone to, to kind of fill that gap. You know, I, I asked their CPOA very pointed question, what's the difference between an, uh, a AI agent and an API call right now? Look, the internet runs on API calls, right?

Something like, I forgot what it is. 57%, 62% of all traffic on the internet is a API to API kind of call. Um, I mean, and I'm not trying to be cute.

I don't do cute, but what is the difference between gen and API calls, Um, hardly anything. And, uh, and I'll go, you one better. If you take apart the model context protocol, you will find an implementation of GRPC with JSON attached to it to do the description.

And, but that wasn't sexy. So they called it MCP and made it out to be like this great advance for integrating all those AI agents that are essentially just gonna wind up making API calls to legacy applications, Especially these efe ephemeral single use, relatively simple. Just, Hey, do this.

There's not a lot of intelligence built into it. I wanna bring up one other piece of my conversation with Meher though, because I enjoy, I enjoy listening to me, I, I enjoy listening to smart folks. We, we spoke about, um, general artificial intelligence, right?

GAI, the sort of holy grail, the singularity that we're all, you know, supposed to be waiting for. 'cause that's when skin comes in and decides carbon based life forms are no longer necessary, right? Um, he says that may maybe pie in the sky, not be pie in the sky, but there is sort of a, a, a, uh, general AI type of AI that is gonna be used in business processes.

Automation. And Mike, to your point about what held RPA back, it's not that RPA was a failure, mind you, right? I mean, a lot of industries use it extensively, but what held it back was that ability to kind of think outside the box, to make a left when it had to make a left and make a right when it had to make a right and, and know when, right?

You can't rule out, you can't just have rules for everything. It's gotta have some autonomy and intelligence. And that to Mahesh is sort of general artificial intelligence as it relates to process automation.

And that's really what they're shooting for. They make another announcement around that up in Orlando, and I think that is something that Bears watching. Yeah.

I think other people are kind of coming up with a, a flavor of general artificial intelligence where they're not saying it's full boat a ga g Well, that's, that's what he's doing too. They're, yeah, they're saying that within the confines of a narrow set of tasks, we have achieved some general intelligence. I don't know, I'd like to actually see that work.

I've seen, you know, people talk about it in demos, but business processes are funky that way. You know, they, they tend to be, need to be done the same way every time. And if there's an AI agent that is, you know, probabilistic, they may not do it the same way every time.

So you gotta really understand how to implement that. I think as we go along, you're gonna see a mix of probabilistic and deterministic processes that people are gonna have to mesh together. And this requires, you know, enterprise expertise and a lot of work.

So it's not gonna happen overnight, but it's Gonna happen. I, I agree. I agree.

Um, let me close out this segment with just two quick announcements. So, uh, the, the interviews that we did up in Orlando, we didn't, we didn't do those live. com, but, um, the videos will be up, I believe, on Monday on Textron tv.

So you'll be able to see them there. They'll be part of the Monday show, as well as available on our OTT channel and Techstrong tv. And I did you, I did mention the, uh, uh, click Connect conference where I, I ran into some of our Tech Field day, uh, brethren from Pucher as well as Guy Courier, um, that I believe will also be up next week.

They, they didn't broadcast live from there, but it'll be up. We'll get, we'll, We'll get a report from Guy next time he is on the show, and we'll make sure we, Oh, we'll do it justice. All right.

So from Automation Anywhere to text on TV here, let's take a break. We'll be back with, uh, more m and a news and more m and A news. Hi.

Oh man, I feel like it's Apocalypse now. You're watching Textron Gang. Discover Textron Group, the epicenter of tech innovation.

We are your go-to for reaching IT leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more.

Join our satisfied clients. Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Techron Group.

Hey folks, we're back and we're talking about open source and robotics. 'cause there was a move by hugging face that maybe didn't get the attention it deserves, at least in my opinion, but they bought a company that's gonna allow them to share with other folks open source hardware and presumably software that will be used to build various classes of RO robots. Now, John, this is your story and you've covered it, but I think something big is a foot here.

Yeah, no, no, I thought it was very interesting. So, hugging Face bought this company our plans to buy a company called Pollen Robotics, which is based in France. They're the people who developed something called CHE two, which is a $70,000 bot that's used for academic research at Carnegie Mellon and Cornell University, as well as, as, uh, they test, uh, embodied AI applications.

The reason why it is interesting is hugging faces purchase of pollen is, is i, I, I think arguably marks its first step into selling hard hardware. After it, it enhanced or boiled, its its robotic software capabilities last year. And there was a key hired by hugging face.

They hired, um, former researcher on Optimus Humanoid in March, 2024. His name is Remi Kade. And they also later launched, um, lab robots, an open source robotics library code.

So I think as in the words of hugging face co-founder Thomas Wolf, he said that robotics is gonna be the next frontier that AI will unlock. Um, he said that having AI embodied robotics might help solve challenges to achieve human-like a GI or artificial general intelligence. So I think it's very interesting and, and, and, and kind of moving forward what the applications could mean in terms of use of robot humanoid robotics in, in various, various capacities.

Mm-hmm. Alan, I think we've seen this play before. Proprietary folks start something and then the open source community kind of jumps in behind it.

And the next thing you know, um, stuff is everywhere because, well, the total cost of actually doing something dropped through the floor. So, um, is open source gonna do it again? Absolutely.

And I, you know, there's always a place for open source models within the various tech, you know, silos if you will. What, what couple of interesting things. Number one, this is a move from plug face, not just into robotics or hardware, but into, as we call it, or as Jensen won calls it, right?

Physical ai, right? Which is after authentic ai, it's the next great frontier. Physical AI mar marrying AI to physical devices, whether you want to call 'em robots, IOT devices, what have you.

They're, they're, they're real, you know, uh, device, it's physical, it's much software. Now you wanna call that hardware, okay? I think in the case of a robot, it's hardware, but it'll, you, you know, is your refrigerator gonna be hardware, but because that's gonna have a, you'll be AI enabled.

Now, the intro, the other interesting here is this is a, a case of a company that has open source hardware. You know, open source hardware is not as popular as open source software, right? A lot of people have a hard time getting their heads around the concept of open source hardware.

What the heck is open source hardware anyway? Well, basically they're printing schematics, they're printing the blueprint, or not printing. I'm, I'm so old we don't print anything anymore.

Um, you know, they're, they're making available the schematics and the blueprint so that you could build this hardware yourself. We, we see it with some CPU, you know, type of, uh, semiconductors, their open source design. And this was a company, I remember when they were first launching, you know, they, their idea was if they can get everybody building this robot because we open source the design for it, and then they offered better software for the robot or additional functionality, maybe a, a removable arm that does some specific test that's not in the, like open core, open source hardware, right?

The, the, it's the same principle as the software now with the hardware. So now one has to ask themselves, okay, so hugging face is buying a company that makes an open source hardware robot that anyone can make. Theoretically, does hugging face think they're gonna make it cheaper, better, faster?

Maybe, but why are they really doing it? Because they think they could put better software in it. And now the question becomes, look, I own this open source hardware ip, but anybody could make it.

But anybody who makes it should use my software. It's kinda like the arm processing model, right? Basically Very much is the R model.

That's why I said semiconductors. You get the design and you can manufacture it, and then you can use any model you want from the AI side and put it into the robot. It's interesting to me, you know, I was having this chat with this fellow who was testing some robots in a warehouse.

He was saying, you know, it kills them because, um, you know, it's a half a million dollars to get the robot, get it all set up and tested, and it falls down a flight of stairs and it A, doesn't get up, and b it breaks. And so they're like, we need a cheaper way to do this. 'cause um, you know, physical AI is far from perfect and it's still trial and error.

Trial error. Yeah. I mean, but again, you know, much like back to our previous conversation on AgTech process automation, you know, you, we, we've been playing with robots for, for decades, right?

We've all seen the latest advantage advances. The door runs really fast, it doesn't fall over as much. I mean, most of our factories, Teslas, iPhones, they're built by, you know, very single use, single dedicated type of, uh, of functionality.

But they're robots. Not all robots have to be humanoid. This, this particular one is right.

And, you know, they'll, they, we will get to the point where we need to discuss how human do we want our humanoid, humanoid robots to be. Right? It could get a little freaky unless, you know, they'll be an adult industry that of course will want 'em very human.

But, um, but such as, such as humanity. Unless, unless these new pornography, uh, laws come in being, and then you're not allowed to mention that stuff anymore. But I digress back to humanoid robots.

Um, um, you know, the, the key here is what, what, you know, who's really gonna manufacture them? How much does it cost to really make a quality robot, even if I already have the blueprints? Mm-hmm.

I'm assuming that, um, well, I guess we'll have to have a made in America robotics campaign, but right now Chinese are projecting that they're gonna turn out these robots in the millions. So I think they're building the plants are in it. Yeah, they are.

Yeah. They're, they're, they, they're, they're far ahead of us. We, you know, it's, it's inter I was talking to somebody at Amazon, a couple, it was about a week or two ago about this concept.

And, uh, and they, they came out with something called Balkan, which they're gonna use kind of a robotic arm on the assembly line. But he, he was talking, this is an executive, he was talking about the challenges of implementing humanoid robots. So I imagine something like this, and then the overarching comments of, of Jensen about physical ai, I mean, we're all kind of moving, moving in the direction of how do you bring the cost down?

Because I think Rishi too is, like, I, I mentioned 70,000, that that's a major issue here, is bringing the price on and bringing the efficiencies and specialization. So it's, it's at the start. We're far behind the Chinese, but at least it's a start.

And it's, that's why I think it's interesting and significant. Sure is, sure is. Look, I think, we'll, you know, I, I see a robot in your future, Mike.

I do too. I'm just Not batting cleanup for the New York Yankees. I just can't figure out if It's, I just can't figure out if this is a robot that's helping me or chasing me.

'cause I have both drinks. Again, you know, that might be illegal under project 2025. Don't go too far into it.

But can we have robotic boxers? MMA fighters, robotic baseball players? Yeah, but, and how do I bet on robots?

'cause if they're both equally skilled, then they'll just beat each other to a pulp, and then there'll be a draw. So like, FanDuel out, It was a movie. Wasn't there a movie with Sylvester Stallone managing like a beat up robot?

I can't remember the name of it, but it, yeah. Robotic football Players who Don't Get Injuries. I think, I think, I think the remake of that is gonna be any day now.

It's gonna be called Robo Rocky 27. Robo Rocky. How about Rocky Robo?

Either one. Maybe we could sell that for a billion dollars right here on Textron Gang. All right.

We're getting silly. It's Friday afternoon. As long as you mention AI in the first, uh, paragraph of the press release, I think you're gold.

Well, the robots can't work without ai, otherwise, they're just dumb robots. You, you, you're, you're killing me, John. You're killing me.

Sorry. All right. Hey, John, Mike, have a great weekend, Mike.

Enjoy that Subway series. Let's hope Judge keeps hitting like he's hitting. Let's hope our next win tonight.

John, it, it, it's, uh, no joy in Mudville for the dubs out there. Silicon, Silicon Valley. Oh, know, I know.

It was, it was, it was, it was hard to watch What's Over. And when you're so dependent on what player this is gonna happen, It happens. Listen, we've got a lot of text on TV coming your way right after today though.

So this, this week isn't over yet. Stay tuned for that. But we will be back Monday with probably more m and a news and more AI news and gens and robots, and Yankees and lions, tigers and bears.

Oh my. Have a great weekend, everyone. This is Alan Hummel.

We're out.