Techstrong Gang – March 7, 2024

Hey everyone. It's AI Thursday here on the Textron Gang. We've got all kinds of good stuff around AI and sustainability, and green, and what's going on in tech.

You don't want to miss it. You're watching Textron Gang. Hi everyone.

Good morning. It's Alan Shimmel from Techron Group and welcome to another edition of Textron Gang. It is a, an AI powered well AI themed edition here for this Thursday.

Uh, we've got our ai, uh, heavyweights on here, experts, and, and we're gonna talk a little bit of green sustainability as well. Let me introduce you to who we, we have on gang today with us joining us as usual from Colorado Techstrong CTO and Principal Research Analyst Mitchell. Ashley Mitchell.

Welcome. Good morning, ai. Thursday.

Yeah, baby. Okay. Can't have an AI day without this next guy.

It's our AI resident gang member expert, mark Inco. Hey, mark, welcome from the, well you could say where the company is, mark. Hey, good morning, Alan.

I'm calling from the Outer Banks of North Carolina. Mm-Hmm. A mobile, uh, mobile studio.

Or I could be an AI powered avatar. You'll have to decide for yourself Anytime, anywhere, everything. Alright, thanks Mark for joining.

And, uh, by the way, Mark's newsletter and stuff is the Artificially intelligent enterprise. And, uh, check that out. Joining me here in studio at Techstrong headquarters in Lovely but raining today, Boca Raton.

To my immediate left, our, uh, green expert. She's a meteorologist, a author, uh, as well as sustainability. Bonnie Schneider.

Thanks. Great to be here. And then to my far left is our Chief Content Officer here at Techstrong one and only hailing from the Bronx, New York.

Mike Baard. Mike, what do we got for today? All Right, well, let's jump in.

'cause tongues are wagging. I guess. Dell had a, uh, call with the analyst folks in Wall Street and referred to a upcoming GPU that has a thousand watts per GPU generated.

I think that's pretty hot. What are you hearing, Bonnie, from folks about what's going on with AI infrastructure and wattage and climate? Because it seems like, you know, these things are just gonna generate boatloads of carbon.

Yeah. And I think that they're hotter, literally than previously thought. NVIDIA's next Gen AI is reporting to be taking 40% more energy than originally thought.

So now that we're seeing, okay, this is causing a lot more energy use, how are we gonna keep things literally cool? And I think that's where Dell is coming in with some innovations in design, whether we are creating ai, um, that not only is obviously gonna tap into the power demand, but it's going to be able to cool at the same time. And I think that's gonna be a key technology going forward.

I don't know, mark, is AI killing the planet? You're into this pretty heavily, but there's a lot of folks, especially in Europe, asking tough questions. Well, I think it's the, the key is how much work do you get for those watts?

So when we look at the wattage right now, we're looking at how much power it's consuming. If you look at said, who's the CEO of, uh, Nvidia and their earnings, they were talking about a 20 to one efficiency in the amount of work for A GPU. The problem is that, um, you know, at the rate that we're consuming ai, it may be more efficient than a CPU, but it may, we may be just increasing our appetites.

Uh, another company out there, gr, GROQ, um, they, they power the inference. So the sort of, when we ask Chachi PT a question that's inference, they are supposedly Dwayne, um, 20 to one, uh, efficiencies versus Nvidia. So it's not so much how much power they're burning is, you know, how much power per amount of work.

And that's, um, remains to be seen. So I have a few thoughts on this surprise. Um, number one, I think the cooling argue, uh, you know, line of questioning is a red herring.

We'll figure out the cooling, right? There's always this, you know, how much processing power versus how do I cool it versus where do I store all the data it needs and generates and all of that stuff that'll be figured out. And I'm not worried as much about the cooling.

As a matter of fact, there's an argument to be made that the more cooling you have, the more energy you're gonna need to produce the cooling as well. But however, I am really concerned just about our grid, about our capacity, right? This is what's slowing down electric vehicle production, right?

I don't know how many of you have electric cars and have taken road trips. It's a pain in the, in the butt, right? You, you wait online for few, too few chargers that don't work.

And, you know, it, it's kind of becomes a self-fulfilling prophecy here. The more successful AI is, the more capacity we're gonna need to power these processors. Whether they're 20 to 1, 41 40 to one, or a hundred to one, we're going to need that capacity.

And so between that and electric cars, fundamentally we need a better mouse chaplain. It comes to electricity, right? And I don't know if the answer is building these new micro nuclear fission, uh, uh, plants, or maybe we go to fusion or, or we, you know, we'd line up windmills from here to there.

I, I don't know what the answer is. But we need to do something about the electric, uh, uh, generating capacity and the grid, not just here in the US but around the world. If we're gonna empower these new technologies.

You know, per what I was gonna add is we're, we're in an era where it isn't just, you know, Moore's law and how many MIPS is this chip going back least here. But, um, if you look at the new designs of the chips, and this doesn't alleviate the the power problem, but you have to also consider these are really systems on a chip. They're not just a CPU or A GPU.

They've got onboard memory and, and applications are built to leverage the fact that the architecture of the chip has changed. So we might be care comparing maybe a small orange to a, a small apple, or a big orange to a big apple, I'm not sure. But it isn't as simple as saying it's this fast, because now you're, you're essentially comparing a computer that's really in a chip.

So Mark Elon Musk is saying, well, we're gonna run outta GPU capacity anyway, so maybe this won't be as big a problem 'cause we won't have enough GPUs to run these workloads anyway. And who Cares what Elon Musk says. Go Ahead.

Mark. When did that Make, I just wanna put that in here. So, yeah, I mean, GPU, um, is the sort of supply chain checkpoint right now, along with what you just said, Alan, is the, you know, power considerations for data centers.

Back in the olden days when I got started in the, you know, mid nineties, the big choke point for data centers was bandwidth. Now the choke point is, you know, uh, power for those things. And I think you're right that, that that power that is reliable near the data center is now the new, the new need.

Um, and I think, you know, Elon Musk is right. I, Sam Altman is right, is that we need to, to come up with a global supply chain for GPUs, and not just a supply chain, but a supply chain of efficient GPUs, which, you know, every single bit of technology as, as, um, Mitch just said, has, has benefited from, you know, miniaturization. So Moore's law is part of a bigger trend that was sort of kicked off actually before the space program.

But the space program is what made us make that leap forward is we needed micro, you know, um, micro sized processors that could go into space. If I'm hoping, and I'm an optimist some days, is that this need for, um, AI technology will be the, you know, moonshot for us to start reducing power consumption and improving efficiency, which, yeah, you know, we're seeing it now. That's, that's exactly where, where I think this is going.

We're seeing, uh, a greater demand for clean energy solutions and how to power all this. And that, that is something that I've seen from the interviews and people I've spoken to, is, okay, we know we're gonna need more energy, we need better energy resources than what we have right now. So there's a, a greater drive towards that.

Also though, with data centers, right? Some of the more forward thinking data center operators, Google, Microsoft, uh, you know, so, and some of the dedicated just pure data center players, they've anticipated this, uh, uh, cap electric capacity issue in cooling as well. And they've been building data centers above the Arctic Circle piping in cold air from outside to cool things down.

They've been building data centers next to rivers so they can create hydroelectric, you know, dams and turbines right outside the data center. So we, we have been, this is not a new issue, it's just, I think the, the amount of electricity or the needs of this whole new generation of, of processors, you know, is making it more acute. But this is, this is something that's been staring us in the face for a while.

Can I just say a quick word on an Elon Musk though, too? So, in addition to saying we have a GPU issued, he, he's now threatening, or maybe he is suing Open AI because they went and went commercial on him. And, and it's come out now that when he originally wanted to get involved in open ai, he actually wanted to combine it with Tesla.

And, and I can guarantee you, like everything else, Elon Musk does it, it ends up in making a buck for Elon Musk. And so, you know what I, I respect what he accomplished and all the money he made, but let's face it, the boy's off the tracks now for a while. Well, you know what the, uh, OpenAI just released a bunch of emails from early back in the history of the company when he's involved in starting it.

And there's all these email trails of you've gotta have a commercial offering. You can't fund this just by raising money. Mm-Hmm.

This, this can't, you can't stand alone on that. You've gotta have a profitable company here. So, you know, it, it's, he's not the first person to speak both out both sides of his mouth, but I think it's the money talking, right?

He thinks He can influence. So By doing, that's talking Regardless of Elon's sanity or not. Mark, we hear a lot about AMDs coming down the pike with GPUs soon.

Is that gonna alleviate the supply chain crisis, or what do you think? Is that a real thing? I mean, it seems like they talk a lot about it, but no one's seen one.

I think the, the big, it's a short term fix. I mean, I think that as much as that sounds crazy about Sam Altman's $7 trillion, uh, chip industry around GPUs, I think a MD coming in to the fray helps alleviate that. Um, because it takes, I think 10 in six or seven years to build a chip plant.

So, you know, that's gonna be at the scale that we need. So anyone that can get to, to market faster is gonna relie alleviate that pressure. I mean, only time will tell when you see the performance of these GPUs under workloads.

And, you know, there's two types of workloads. There's, uh, training and inference. And right now the training is exponentially more resource intensive than inference.

And, um, you know, we're taking all our data and training these models on it. And so, you know, we're getting hit from both sides of, you know, huge demand for more AI and a lack of GPUs. 5 trillion company.

Wow. They were, yeah, But we only used GPUs because they excel at parallel processing. They weren't really designed for these AI workloads in the first place.

So my next question is, do we need new infrastructure? And I'll start with you to drive these new applications. I mean, some folks are talking about photonics or photonics and how, and optical systems and all kinds of things.

Quantum, Quantum, quantum, you gotta mention quantum two, It's, it's unclear how we're gonna manufacture these things at scale. But, you know, you talked about space. Do we just need a new approach here because the existing stuff isn't gonna make it?

Yeah, I think a new approach is good, but I think that as we take these new approaches, one of the key, uh, factors, we'll be going back to what you're saying earlier is cost, is, is that we are able to do this in a way that's cost effective and also minimize the use of energy. Uh, we were talking about that previously as well, when, when architecture changes, we're able to do more with less, create more processes that are automated, that's great. But, um, what is the cost of it?

And will it, in the end, um, bring about reduced energy use? I think some of that will be remained to be seen as we, as we move forward. Look, you know, we not driving model.

I'm sorry, go ahead, Mitch. Uh, I think the other thing is, is we can't use the intel model is, is how innovation's gonna move in this era of AI chips. I think we're gonna see architectural changes very rapidly, like the current HBCE, the, uh, H 100, H 200, uh, era of chips from Nvidia.

I, I, I'm not in the inside of what's happening in their product planning, but I gotta imagine it isn't just gonna be an 80 86 to an 80, 88 to a two, blah, blah, blah, blah. I think we're gonna see, uh, real architectural changes in innovations of what's happening on that chip, not just packing more into it and thus consuming more power. Right?

I I, I, I agree, Mitch. The, the Intel architecture was very much linear, packing more, you know, uh, uh, gates into a, you know, thinner and thinner and smaller and smaller footprint. I think in ai, though, you gotta remember, we're still early in the game.

Yeah, yeah. Right now we're using a, a, a butter knife instead of a screwdriver. Right?

Eventually, we're gonna develop a screwdriver that works better on screws than using your butter knife outta the kitchen drawer that your wife yells at, don't ruin my butter knife. Um, so I think that sort of innovation's gonna happen and it's coming and it's coming quicker. Will it alleviate all these problems?

No, because it'll be incremental. But five years from now, as Mark mentioned, when, you know, some of these new foundries are on and everything else, this'll be an early days problem. Before we end this chat though, I do wanna note that AI isn't always the bad guy here, right?

We have folks in Abu Dhabi's oil company is talking about how they're taking one to 2 million tons of carbon out of the environment because they used AI to optimize everything else they have. So, you know, there's, there's No, I mean, there's a reason why everyone's excited, right? It's supposed to.

And this is, this is the promise of it. That's why we're, you know, otherwise who wants to listen to us talk, right? AI is supposed to be doing these kinds of things.

Here's some proof of it. There you go. Mark, last words on this.

Yeah. You know, the one thing that I, I think we should think about too is, and, and you touched on it, is maybe the paradigm shifts. I honestly think that in the long term, Nvidia may not be a hardware company.

They may be a software company. 'cause the thing that they did for, you know, this parallel processing has come up with something called, uh, CUDA, which is the compute unified device architecture. This is the sort of the VMware for GPUs that lets it run across different architecture.

So, you know, I think Alan's great in the fact that, and I hate being that agreeable, but you know, he is a smart fellow. Thanks, Mark. It's a little something extra there for you.

Well, rv, you don't have to do that. You'll get to come back, mark, we'll invite you back. But We, we, we can argue, we can argue about open source kuda when you, when you come back next time.

How's that? Yep. But you know what?

Right now we need to take a break here on text. Drunk back gang. We'll be right back.

All right. We're back here on, uh, Textron Gang. Moving on to our second block, as we call them.

Mike, you want to, uh, introduce this one? Yeah. So there's a report from, uh, cast AI did an analysis of, um, how Kubernetes clusters are being consumed in the cloud.

And surprise, surprise, they discovered that there is a lot of over provisioning, which leads to lots of extra carbon being generated for no apparent purpose. So let's just get started here with, you know, level set us a bit, Mitch, but why do developers constantly over provision? This has been going on for decades.

It's kind of an open secret. I mean, what is it that they don't seem to care about the environment, or is there something else at work here? Well, I, I find it interesting that we don't have more studies about this, given sort of the numbers coming out of this cast AI report.

You know, they, they, they quoted numbers like from a thousand to 30,000 processors. They're only using 17% of capacity. That's a pretty big range.

Um, so I'm not sure what to make exactly of the numbers. I, I suspect it's partially, um, it's partially knowing how to optimize. I mean, Kubernetes is complex to use and set up, et cetera.

And oftentimes people are using the additions, if you will, that come through the cloud service providers or others. And, uh, you know, developers don't spend their time on efficiency. Most of the time.

They're working on capabilities and, and fixing issues and security. And with platform engineers and SREs tend to more for focus on that. So maybe that's where the opportunity is to really develop some great Kubernetes chops in other parts of the team to help make kind of a holistic answer to how do we better use, utilize those resources.

I'm a little more cynical about it. I think developers don't want to be woken up at two o'clock in the morning. So they just get as much memory and storage resources as they can get.

'cause they want their application to always be available and they just, you know, they, they turn the button to 11 every time I say stop blaming the developers is an ops issue, right? And you know, the developers, yes, there are shadow IT where developers are setting up the infrastructure. But in today's enterprises, where a lot of this waste is happening, there are platform engineers, the new darlings, right?

Platform engineers and ops folks who are setting up the infrastructure that the developers code runs in. They put in the guardrails. I think the bottom line is, is this, this is why there's a whole FinTech industry free.

Get Kubernetes for a second. The cloud and on purpose makes it far too easy to spin up wasteful instances and over, over, uh, produce, not over produce, over provision. Yeah.

Your, your resources. And then makes it far too hard to say, wait a second, I got all these instances that I'm only using 17% of, or I'm not using at all. Shut it off, shut it off, shut it off.

To me, this is a FinTech problem. I think you mean finops, Finops, excuse me. There you GOs different.

Alright. Excuse me. Well, let me ask you, I mean, are the cloud service providers somewhat complacent in all of this from a green perspective?

Because, you know, they're like, yeah, we want just more workloads. Well, I think that there's also a shift of education to, to to, to show how to do, uh, green coding, green software. I speak to the Green Software Foundation a lot, and they're actually growing quite a bit in, uh, membership and, and organizations wanting to fix this problem using less resources and having, um, more efficient coding.

So, um, that's something that I think we're just talking about AI and as learn as we go. But I think the green, um, computing green cloud, that's also something that's evolving to, to better use resources and, and waste less energy. You know, there's another approach to this too, and that is, um, I'm full disclosure, I'm doing some work with AWS serverless organization and that has taken this, the Lambda type approach where you don't provision things ahead of time that's controlled through the infrastructure and dynamically provisioned for you.

Yes, there's some things involved in, in configuring that. And of course, they have a, their Fargate offering, which is their Kubernetes based to run workloads, um, you know, containerized workloads. So there, there are other approaches too than just sort of throwing raw Kubernetes at it too.

And maybe those will be, we'll find the, the best applications to use for Lambda and other serverless options from Google and, uh, AWS and, and, uh, Microsoft. Mark, we have all these AI models with all these parameters. Do they need energy ratings?

Can we figure that one out? Uh, I think it's way too early, Mike. Things are changing on the weekly basis for that.

But I would like to go back to the whole Kubernetes thing. Think about this. We started out, it's a, it's a cloud infrastructure.

It came outta Google and it was built at a time where Google couldn't grow fast enough to handle the demand for that Amazon. All the other cloud providers adopted Kubernetes, but they are, they're the utility. Then we realized that we are overspending on cloud.

We started to repatriate workloads in the data center. They were running heavier full virtualization, and now they're going to containerized. So they're using the same infrastructure and the same provisioning model with something that is five times more eff efficient.

So, you know, Charlie Munger always said, you know, show me the, uh, um, incentives. I'll show you the, uh, result. And the result for the developer is, I don't wanna be called in the middle of the night, as Alan said.

So let's, let's overprovision keep the data center up. But the developer isn't on the hook for, for the, uh, power compute bill and over provisioning when it comes to AI models. I think that, you know, they're gonna run in some kind of virtualized workload.

I think PDA from Nvidia is the way they'll do it. But, um, you know, for now, AI is, is probably not, not as big a player in that as just regular workloads running in containers at the application ladder. I think there's a large number of AI workloads running on those Kubernetes clusters already, but maybe not as many o other workloads.

But Mitch Kubernetes was supposed to scale up and scale down. Seems like it scales up, but never comes down. Why is that?

You, you still have to kind of define what the, um, parameters are for the cluster to be able to scale up and scale down and how rapidly it does that, right? Because it's not, it's all, all sitting there to instantaneously, instantaneously use, there's this thing called the Cold start, right? Especially in the serverless world.

Um, which is that time it takes to spin up the infrastructure to catch up with the workload or pre-provision a certain amount ahead of that. Maybe we're just over provisioning. How much is sitting there ready versus what would dynamically be spun up as workloads?

It's a good question. I, you know, I'm not running a big Kubernetes instance, uh, or workload to, to share any personal experience with it. But it seems like, uh, it's a place for definitely a lot of optimization left to, to, uh, go after.

I think why make Kubernetes the whipping boy here? This is the problem with the cloud in general. You know, this whole elasticity thing is one way, right?

Unless you're really using, uh, uh, fin finops. finops Yeah. Solution.

No one turns their stuff down. No one, and, and the cloud, the cloud providers like it that way. What do you want 'em to send out?

Reminders, right? So this is, this is what it is. It is what it is.

And, and again, I think that's why there's a whole cottage industry around that. But can we build that into Kubernetes, perhaps, right. To have that two-way elasticity.

But it, it's been a problem before Kubernetes and it'll be a problem after Kubernetes, I'm afraid. There are companies that specialize in this, you know, storm Forge and you could listen list to a whole bunch of companies that focus more efficient use of Kubernetes. So it, it's kind of driving its own cottage industry.

Yeah. You hear this conversation. Do you think we're serious enough about all this stuff?

'cause I don't see that little energy consumption button on the console when I'm monitoring my environment just yet. No, I don't think so. I think there, there's, um, there isn't, uh, um, that is not happening.

But there are more companies that are providing tools for, um, organizations to monitor energy use, especially within IT departments. From, from what what I I'm hearing is that they're getting more pressure to reduce costs and, and monitor the energy at the same time, which isn't always something that's innate that they would know what to do. So, Cool.

Are we done on this one? Can we, uh, I think we can move On to our next topic. Hey, we're gonna go to our next topic here in just 30 seconds.

com is the number one online destination for DevOps education and community building. com covers all aspects of DevOps, including DevOps, best practices and tools, DevOps culture, DevSecOps, business impact, continuous testing, continuous delivery and more. com has the largest collection of original DevOps content featuring breaking news, blog posts, podcasts, and more.

com to learn more. com where the world meets DevOps. All right, we're back with our green theme and now we're moving on to a, a lawsuit that's floating around.

Can you imagine that there was lawsuits involving energy and it, I'm shocked and appalled, but apparently the Texas Blockchain Council is suing the Department of Energy because the Department of Energy had the temerity to ask the local blockchain operators in Texas, how much energy are you consuming? And I guess the blockchain council basically told the DOE, it's none of your business. Um, and then they went to court over this and lo and behold they got a restraining order.

And so here's my question to you. What is that fine line legally here? I mean, if I'm running a data center, who do I gotta tell 'em?

Why do I gotta tell 'em this? Because it seems like there's no case law here that says, yep, you're responsible for this and the DOE is authorized because it doesn't look like they are. I think that there is, there is a fine line of what's considered privacy to, to run your, your, your, your crypto, uh, operation that you're doing.

But I also think that, um, there's a reason why the Department of Energy is trying to keep track of this, this energy surge for other reasons as well. When you have so many different resources tapping into the grid, especially you mentioned in Texas where there wasn't too long ago where we had major outages during a winter storm. Um, I think transparency is key.

We, we have to know what's, what are these different demands on our energy, um, grid that's happening? Because in the worst case scenario, it could be too much and they're not prepared for it. I don't disagree that we need to know, but I don't think we have any legal right to know just yet because all this stuff has not been put into any legal document from Congress.

It's been a agency seems to be extrapolating its authority a little bit and saying, we have, therefore we can collect. We have seen agencies. Are you, Are you calling for smaller government?

I I am just pointing out the fact that the Supreme Court has made it pretty clear that they don't really dig this, uh, agency overreach, shall we say, whether you agree or disagree. But, um, do we need Congress to show up here? Uh, maybe Mark or Mitch, you want to throw some thoughts in here, but, um, do the people who make the laws even understand what we're talking about?

I think you hit it right on the head. Um, Mike, I mean, it's a public utility. They should be able to instrument their power usage from the grid.

I mean, that's, that's, it's a public utility. Unless these crypto guys have their own little fission plants in their backyards or, um, maybe they're hanging out with, uh, Alan's buddy Elon Musk and doing Solar City installations. Yeah, he's in Texas too.

That's right. Um, but I mean, it's a weird, it's a weird thing 'cause I don't understand why they wouldn't just instrument their grid and, you know, they should be doing that anyhow to make sure that the grid doesn't go down while their, uh, senators are partying in Mexico. This, you know, that Was during that storm actually.

Absolutely. I was talking about, so there's a few things at play here though. First of all, who the hell is the Texas Blockchain Council?

Yeah. I didn't Know there was one Trying the hell out. That was a good idea.

Right? The Texas Blockchain Council. And look, if they were just grabbing electricity from the local utility provider, we got these things called meters Mm-Hmm.

It would make it real easy to, to see, uh, what's going on. I suspect like most crypto mining, uh, operations is they're, they're getting energy off grid somehow, or someone else's energy on someone else's meter. And maybe that's what's really a play here.

But the other thing in this instant case, it wasn't like they were served with a subpoena or they were put under arrest or anything. They were asked to just fill out a survey so that the, the agency can better ascertain what are the energy needs in a state that frankly has had energy needs. That's right.

Deliverability issues. Uh, you know, all kidding aside, so this wasn't a case of, you know, people peeping in the bedroom and, and bringing potential sodomy charges or something. We just wanted to know how much energy using so we could provision for it.

Potentially. It'll be interesting to see now the courts in Texas, so you can imagine which way this is gonna go. But, um, eventually, hopefully it gets brought up to some federal appeals court level and it'll be interesting to see which way it goes.

Well, there's other uses of that information. So I'm gonna take a left turn, but it ties to this, yes, I'm from Colorado. I have to bring up cannabis and the cannabis industry.

One of the things that used to detect illegal grow operations is energy consumption. Um, both heat coming out of, out of a particular location like someone's home or their basement or someone stealing energy from their neighbors, right? To tapping into, uh, a place to plug in and go run all their grow lap grow lights or whatever you do to run a cannabis operation.

So this, this data could also potentially be used for saying, now what's going on here? This cryptocurrency company, you know, everybody elses uses about this much per these many transactions, but these guys are doing something different. Why are their energy so consumption going so high?

Maybe that's where the Russian and the whatever and the, the, the money is flowing off, off continent and black money is being exchanged on the black market or whatever. So not that, not that I'm saying that's happening, but this, this information is more than just the, you know, altruistic. Oh, let's see how much consumption everybody's using could be used for other purposes too.

Do you think? And that was a hundred percent my point, Mitch, is that, you know, asking for the information and monitoring it, you know, it's, it's lawmakers that don't have the technology understanding to know. And I was thinking the exact same thing you were.

'cause where I grew up, that thing is, you know, the DEA would come in to places and they would track, you know, the frequency of those grow lights is communicated across power lines. They could be doing the same thing for GPUs and crypto mining, and it would be a better indication of the trends than asking some middle manager to fill out a survey on, you know, and, and Alan hit it on the head, what the hell is the Texas Blockchain Foundation? Whatever.

I mean, it's just, it's ridiculous. Like, yeah, You don't think they have From the random committee generator algorithm. Lemme ask Bonnie here though, do you think that because of what's going on with AI and what's going on with blockchain is the very reason why he can't charge his car?

Oh, I don't know. I don't know if there's a direct correlation to that, but, uh, but there definitely are issues. Now, this is just a, a weather aside.

Sometimes, uh, you know, in colder locations you can't, uh, you have trouble with that. But I know we're in Florida, so that's not the issue here. Um, but, um, you know, I think that it does, it all ties together because we're, we just, were talking about how we're using AI is using even more energy than expected.

So monitoring what, uh, as Alan calls from the crypto guys are doing, uh, makes sense to, to be able to gauge for these outages, particularly as we said in Texas. So I'm not sure of the connection to the actual charging, but I think that nobody wants to lose power for any reason, especially if you depend on it to drive. So, um, just a better understanding of, of as these things evolve, what energy is being used, how much, and keeping track of it, um, even for this organization in Texas could actually help them when they wanna charge their cars.

But there's also, and, and, and thank you for bringing it up, Bonnie. And for those, I'm gonna assume most of our audience out here is familiar with how crypto mining works. But basically you, you are, you are running the algorithms to figure out the value of a, of a particular blockchain, of a particular Bitcoin or whatever you're mining.

And if once you get it, you get that, that coin, right? 20,000, $40,000, whatever Bitcoin goes for these days. Now there's always like this ratio of how much computing power it costs you to figure out that entire algorithm and get that Bitcoin.

And notoriously, notoriously Bitcoin miners don't want to pay the vig over here for that electricity. So they're always looking to grab some free juice to, to do their mining and, and lower their cost. And that's really my sus and I'm not, I'm not familiar with the good folks at the Texas, Texas Blockchain Council, but I am familiar enough with Bitcoin mining to know that that's probably what's happening.

This leads to crypto jacking where people steal compute resources in the cloud. And too many it people think that that's a nuisance crime, but it all adds up to these. It absolutely it's a nuisance until it's not.

Mm-Hmm. You know, so did that, you know, these Texas Blockchain Council are not angels. I'm sure you know, poor, poor woe is meek helped me from the big bad government.

But anyway, as I said, the courts will decide it. I I have decided one thing. The people who are gonna make a lot of money off all this are lawyers.

Don't they always? Of course. They always, no doubt uhhuh all.

Alright, I think that's what we have on our, our, uh, call for today. Mark, before we wrap up though, I'm gonna embarrass you a little bit and ask you to give us a little promo for this great virtual event we've got going on in late May. Yeah.

Yeah. So, uh, um, May 21st, we are going to have the 24 hour, um, artificially intelligent enterprise online conference. We're gonna have over 70 speakers, uh, executives, uh, uh, practitioners.

We have, you know, analyst firms like Gartner coming from public companies. We have a few surprises, uh, you know, CEOs of a couple public companies are gonna speak, um, as well as people that are actually using Gen AI to solve real business problems. And I think this is a first of its kind.

It's, uh, you know, how to use AI for practical purposes other than, you know, you know, heating up the planet and making, uh, making us hotter than we should be. But, uh, um, yeah, I'm, I'm really excited about it. So Tuned.

Where can people go? Stay tuned. ai, um, homepage shortly.

So, uh, Yeah, no, there'll be links off Techstrong ai. com, but it's the AI enterprise AI enterprise dot what Online? Online ai Enterprise Online.

You can go right there as well. We're looking for full disclosure, we're partnering with Mark on this. We're helping on some of the backend productions and so forth.

But the speaker lineup is amazing. You don't want to miss this one. Uh, May 21st, you said?

Yeah, May 21st. And it's all day, man. 24 hours.

Follow the sun. Don't miss it. Yeah.

Very cool. That's gonna call a wrap to this edition of the Textron Gang. Mark, thanks for joining us from, from the beach there.

Uh, Mitchell, Rocky Mountain High Live Power on Bonnie Michael. Thanks. Joining Party on You, baby.

Yeah, thanks for joining us here in studio. We'll be back on Monday, right? We, we usually take off Friday.

Those guys in Colorado. It's hard to get 'em in on Fridays, but we'll be back on Monday with some fresh tech strung egg. Don't miss it until then, this is Alan Shimel.

We're out.