AI Regulatory Debate Meets NVIDIA’s Chip Doubling Bet

Hey everybody, welcome to Techstrong Gang. It's Monday. It's also Yom Kippur, which is a holiday for many of our friends.

So we have a shortened group of folks to talk to you today on Techstrong Gang, but let me introduce them to you. Yvette Schmitter, how are you, Yvette? Hey, Mighty Mike.

Mike, Mike, Mike, Mike. You're supposed to say that for Wednesday. All right.

And Steven Bosket, how you doing? Hey, I'm doing better now that I saw Taylor's cat. All right.

Those of you who don't know Taylor, who's the production person behind all this, has a cat. So we've been all hanging out with Taylor's cat for the last 10 minutes in the green room. And of course, pets are always welcome anytime during the show.

So if you are a guest, then you just bring whatever dog or cat you want. It should be fine. All right.

I swear though, you cannot make this stuff up. So today is Yom Kippur. It is the Day of Atonement in the Jewish calendar.

And guess what's happening here in New York? Just about every AI executive in the world is coming here to New York for some United Nations conference. There's going to be a dinner with the leaders of the US and the Chinese government, and hopefully all this will come together.

But it's all about, at some point, regulation. So let me get started here with Yvette. What is your sense of all this debate about regulation?

" And then I got other folks saying, "Well, it's only a matter of time before these next generation models destroy society as we know it and chaos will ensue. But what's to worry about? " Well, Mike, I don't know what to say, but I do know what to say.

Here's one thing. So one idea I think survived every fight in the news this past week is independent verification, right? So the presidents, Obama and Trump, fought over it.

" Europe wrote it into law about for kids. And then California, not to be outdone, made certification a requirement. So when all the noise clears, I do believe the question that's left standing is going to be always the same.

Who gets to check the work, and are they on your payroll if they're doing the checking? And I think that's it. I do think at the end of the day, there needs to be some level of regulation.

I know Jensen don't want none, but he's never wanted any, right? But this is not the same social contract like us getting driver's license and driving a car and freaking seat belts. There's a Department of Insurance in every freaking state.

You have to get your car inspected. There's people sitting on top of stuff. You will get tickets.

You will lose your license. There's people looking at it because we all believe if you're driving a car, you need to have a license, insurance, and if you don't, you get hemmed up. There needs to be that assumption that people need to recognize that is wiped away with AI.

There is no social contract. You can bot code some s**t on Friday, launch it on Saturday that pwns anybody, and you walk away scot-free. That is not the social contract.

It's like I build something, it's made commercially available, and I'm responsible and accountable for the outcomes. That doesn't exist with AI. And I do think if enough people die, I guess it's going to have to come to fruition.

So Steven, there's been a number of people in the Trump administration and elsewhere, including NVIDIA and also in the Valley, going, "Heck, we can just police ourselves. Don't worry about it. " Do we believe these people?

No. Thank you. I'm not Steven.

I'm not Steven, but no. Go ahead, Steven. I think Yvette speaks for both of us here when she says no.

No. Here's the thing. As a wise man said this weekend, the more serious problem than killer robots and AI on the loose is what they can do in the hands of bad people.

They can be weaponized. That's what former President Obama said this week. And I feel like that's really the case.

Somebody was asking me over the weekend, a normie was asking me over the weekend about the AI escaping and hacking into other things and everything and what I thought of that, because they know that I'm, well, here. And what I said was, in every one of these cases, it was the people's fault, not the AI's fault. Well- They literally got shocked to find that it did.

And the problem is, as you pointed out, that there is no social contract, as Yvette said, that would make it a standard social thing that if your machine does bad, it's not on the machine or some abstract. It's on you. Look what's happening in legal communities.

If a lawyer-- Again, there was a great example of this, we talked about it previously, where a lawyer submitted a brief full of AI hallucinations. And the presiding judge said, "It's not about AI. This is not an AI discussion.

This is a tool discussion. You used a tool. The tool did a bad thing.

You signed your name. " And that's what we need. That's what we need as a society.

" No, it's your fault. You did a thing. Yeah.

If you're behind the wheel of a car that smashes into somebody, you're the responsible party. That's what we need. Yep.

But again, we live in a time, I won't say a time, because it's not the time that I grew up in. I grew up in a time where, look, if you were playing three-on-three basketball in Riverside Park, right, and you got clucked by the back of the pole in the back of the thing, that's on you. That is not a foul.

You misjudged it. You own that stuff, right? And the same thing with what's going on today is just like no one wants to be responsible or accountable for their stuff doing stuff.

I say this all the time. Just because you can have kids, doesn't mean you should. And this is the same thing.

Have you been in a restaurant where you go out to dinner to spend good money, and baby kids is swinging from the rafters screaming, and their parents are sitting there like nothing's happening, right? The whole entire restaurant is in disarray. They could get up and leave, but they don't.

Accountability. Same thing with AI. HuggingFace was not part of OpenAI's test.

They got got. " They had no idea. And you mean to tell me y'all expect a good outcome to come from people to say, like, "Oh yeah, that was us.

" And it's just like, okay, then y'all going to fix it, you going to clean it, and you're going to write me a check. But of course, we'll never know if that actually happened because NVIDIA bought them. But anyway, there has to be a level of accountability.

You just can't throw s**t over the wall and then it breaks stuff in somebody's life and be like, "Oops. Hope and prayers. " You mean I can't do donuts on your front lawn and then just drive away and say sorry?

Okay. Steven- No, I will find you. Mike, I will find you, and like, "Oh, you chose violence today.

" Here's the thing. Fundamentally, there can be no responsibility without accountability. So do I think this industry could regulate itself?

No. Yes. But only if there's real accountability and real penalties for what they do wrong So, the problem is that they want to regulate themselves without any responsibility, without any accountability, without any repercussions for anything that happens.

Well, you're not going to do it then. Absolutely not. " If the kid is crazy enough that you've went, "Johnny?

Johnny? " Well, and it's the same here. Why should normal people know what HuggingFace is?

The only reason is because they're bad actors sometimes, or because they've been incompetent, or they have not had the responsibility, they've not had the accountability for their actions. So here's the thing I'm really worried about, though. I'm almost more worried about what kind of regulations will actually be created, because there will be pressure to create them, but they will be, I'm afraid to say, the usual half-assed thing that comes out of Washington.

And so we will have regulations with no accountability anyway. So Yvette, how do we go and get something that's meaningful here? So the last week, every day, sometimes twice a day, three times a day, something is kicking off.

And so let's just watch the calendar, because that's the tell. So on September 12th, Dario published his essay about how the industry delays the frontier in slow capability and models. Right?

All his bobble heads nod, said, "Yeah, yeah, yeah," within hours. Three days later, Senator Hawley kills, like dead on arrival, the antitrust waiver that would let them coordinate legally. Okay, so then a lawsuit lands calling out, allegedly, the slow down talk as a cartel itself.

Then, not to be left out of this debate, our President of the United States calls AI safety warnings a hoax. Oh, God. The hoax.

A hoax. The hoax. And rolls out an AI force on, and we're going to get a new AI tsar.

Then on Friday, his antagonist of antagonists, Governor Newsom, signed a kills switch. Well, it is read as a kills switch order, but it's really not a kills switch order. I think the PR in the news kind of look for clickbait, and kills switch was like the word of the day, and they used kills switch, and not a kills switch.

And we'll talk a little about that. So here's what my tea leaf read is. Underneath all that, while the White House drama and the Sacramento theater, they're fighting over headlines, the real move is this in itself.

California's, I think it's called SB813, already certifies independent verification organizations. Amenai's whole plan puts outside evaluators inside the lab. Welcome, Accenture, to the play.

And for the love of all things freaking Dolly Parton, I'm going to leave that whole Accenture Anthropic thing for a complete different conversation, but I digress. Everybody keeps arriving at the same answer, and the answer is a referee who doesn't work for the team. And until we can get to that point, we're going to have Accenture cooking the books for Anthropic, allegedly.

Right? And then we're going to have any regulation that comes out that doesn't have any teeth. Because when you read the tea leaves, everything that's coming out on the state level, Mike, has a kicker.

It has to be 100 people got to die before something happens. That's how we get there. But it gets better than that.

Other folks are saying, "Well, Steven, if we have these regulations, we will fall behind China. " I'm not sure what the goal of the race is, but we're in a race, apparently. So, is this a real conversation?

Or is this just noise? Because, well, last time I checked, it looks like the Chinese Communist Party is just as scared of this stuff as everybody else is. Oh, yes.

Got to stay ahead of that competition. Okay. I would say that, yeah, honestly, it's more likely that China would come out with regulations that would actually control this stuff than we are because, as was a front page story in the New York Times over the weekend, Trump and his family are so embedded in the AI industry and invested in the AI industry that they have financial reasons not to.

And I don't want to sound like a conspiracy theorist or some kind of anti-government propagandist or something here, but it's hard not to sound like that when you've got, as Yvette said, you've got the people responsible for making the decisions literally investing hundreds of millions of dollars in the outcome on one side versus the other. And when you've got Trump saying that the financial outcome of this stuff is so lucrative, I don't think they care about competition with China or anyone else. I think they care about making money versus- Money ...

China or somebody else. That's what it's all about. And honestly, ultimately, I think some of this problem may work itself out since as these models get better and more refined and so on, they work in smaller and smaller environments, and we may end up with an AI future that doesn't necessarily need trillions of dollars worth of investment in hardware, and isn't a big competition between the US and China because everybody's just got it.

Well, we will come to that one in a minute. But at the end of the day, Yvette, why can't we just set up the, much like we have the Nuclear Energy Commission, we can have the AI Regulatory Commission. What's so hard here?

Oh, Jesus, Joseph, and Mary. Mike. Okay.

Managed by AI Force. Sorry. Right.

The United States AI Force, I might mention. So that could work if we didn't have literally schoolyard fighting happening. Literally, who gets to determine who sits on that commission?

When we have governments in power who are selecting who gets access to what, who gets to see it, who gets to test it, there has to be equal footing. And if there is a commission, it has to be equally weighted like the UN. Because everyone's kind of got a seat at the table.

You can't have one block determining who's on the commission and then who sits on it and who gets it, especially when they're cooking the books. And it has to be fair. Yes.

It has to be naysayers and uninterested parties, and people who don't have a- And right now, nobody is naysaying s**t, is it? Yeah. Today, let's put it this way.

Our S&P is run by the same people who build this s**t, sell this s**t, fear this s**t, secure this s**t, and make money off this s**t. So you expect them to not put guards... Megan Thee Stallion's song is a bit like this.

Anyway, I'm just trying to say, at the end of the day, there has to be equal where everyone has a voice, and one block cannot determine if another block doesn't get access to it, or can't see it, or can't use it. " And if they try to do that, they're being labeled as doomerism or hoax or they're going to get tired to death. All right, folks.

Well, you heard it here. Hey, they're all here in New York this week. The trouble is, none of them seem to have brought any adults to the meeting.

Nope. All right, here we go. Switching gears, though, the question has come up, and Steven touched on it.

" And it was also interesting to see what exactly they said they were going to double, because it wasn't GPUs. They were looking at this Vera CPU, which apparently is the thing that they want to build to run inference, which happens to look a lot like a system on a chip with a CPU that I think I can get from Intel and AMD and just about everybody else. So Steven, what's your take on what's going on here?

Well, first off, again, I don't want to question the numbers here whatsoever, but I will say that Jensen Huang cannot go to a conference and say, "You know, everybody, we're not going to grow at all. " So to have him say that it's going to double next year is not a shock, but of course, he is on the hook for that now. He is responsible for living up to that.

And that being said, the chips in question are really, really good. The Vera Rubin architecture is really solid. The numbers that have come out of that thing are incredibly impressive.

But as you mentioned, they're not the only ones, and that's the interesting aspect here. So you've got Apple's finally released the new Mac Studio, and the reviews are out literally today. The reviews are out and thumbs up.

The Mac Studio well outperforms the NVIDIA Spark DGX by a wide margin running local models. Surprise, everybody figured that it would, but it really does. But interestingly, when I was looking at those numbers, I was looking at AMD and Intel's numbers as well.

They're doing really well, too. And so it is not a one-horse race. NVIDIA has a great platform, they've got a great product, but they've got real competition, and that real competition comes from Intel, it comes from AMD, but it also comes from other Chinese companies.

And so you've got, for example, Huawei out there showing some really impressive numbers as well from their silicon. And ultimately, I think that what's going to happen is it's going to come down to manufacturing. Who can actually get the fabs running and who can actually make the chips in order to sell the chips?

But of course, even that is gated on whether there's really ultimately going to be a business case for all of this investment. And, I think that there are opportunities to do a lot of really incredibly productive things with technology related to generative AI. But that being said, I'm also extremely worried about the amount of money that's been invested in these AI data centers and the relatively low revenues that are being reported by a lot of these companies so far.

So ultimately, yeah, I think that NVIDIA probably will be able to double their sales next year. I think that we're going to see a lot more investment from other parties and in other parties as well. " Like, "Where's the customers?

" And to your point about that, the semiconductor industry itself is saying that they're short about 150,000 workers. So it's not just about the manufacturing facilities, unless we think robots are going to show up there tomorrow and take care of all this stuff. But we don't have the people to actually drive that, and maybe AI will change the way we make semiconductors, but I don't believe that that's happening in the next year or so.

So Yvette, is all this stuff built on a house of chip cards? Yes. " So we all know Jensen loves his black leather jacket and comes out with these audacious numbers.

And so let's just say that he's saying doubling that, that is a fun, unaudited fact. He has to say something, like Steven says. He needs something audacious.

We're going to double. Okay. Triple.

Whatever. It's going to be grand. But then he also has to live up to that number.

So it's a confident forecast, but it's not fact. So two, the word that I want you all to think about is this, and Steven started talking about it a little. So Vera is the CPU feeding the Rubin GPUs, and that placement tells you where the constraint kind of moved.

So raw GPU power stopped being the wall. So getting data to the GPU fast is the new wall. And as Steven said, Huawei, they read some tea leaves too.

So they saw what NVIDIA was doing, Intel, Apple, right? So it pulled its chip roadmap forward to early 2027, and it built this whole pitch around interconnect, linking hundreds of thousands of processors, da, da, da. But okay, let's get real.

Per chip, it's still years behind. Yeah. So instead of racing the calendar, and it's racing the cluster.

But here's the gut punch. The gut punch is this. No one's putting this s**t on slides.

The same AI boom building these fad labs, right, is draining the workers who staff them. So the question is just like, who's going to do this s**t? Like, has manufacturing really worked in the US?

You can't pay people to kind of keep automobile plants open. Like, the plants that are being here, what about Tyson Foods? They're shutting down plants throughout the US.

So what makes you think that we-- Where are they going to build this s**t? And then who's going to do it? Who's really going to work the work?

Make it make sense. Somebody please pass me a drink. I know it's like science fiction out here.

It's crazy. It gets better than this. But Steven, I've had a number of conversations now in the last 10 days where people are pulling back on AI projects simply because they cannot get enough systems and chips and processors.

" Yeah, absolutely, and we're still facing a crunch. And these things do not ramp up fast. We were talking about building new fabs in the US for chip production years ago.

They're still not online. Why are they not online? Because it takes a long time.

The equipment, the supplies, the employees, you are absolutely right as well. It's very, very hard to staff these things. One of these reports that you quoted here shows over 150,000 worker shortage just in the US for these fabs.

And it's not just, of course, the chips. It's everything around them. It is memory chips.

It is storage. There's another story about a new storage fab that may be built in Upstate New York soon. But again, who's going to staff these things?

Where are they going to find these employees? These are skilled jobs. These are high-paying skilled jobs, and it's hard to find people to fill them.

But then it's more. Who's going to build the rest of the data center? How about the power?

How about the air conditioning? How about the literal construction and plumbing? My brother is an HVAC contractor in Columbus, Ohio, and it's really hard to staff HVAC and plumbing jobs right now because so many of the employees have gone over to work on the Intel plant just outside of Columbus.

If you can drain a major metropolitan area of skilled plumbers by just building one plant, they're talking about building dozens of these plants all around the US, plus the data centers that are draining even more people. It's a real challenge, and that's why it's incredible when you see some of these build-outs that have actually happened, like kudos to xAI in Memphis for getting that thing actually built. I never would've guessed that it could've gotten built.

It has some problems, but they got it built. That's more than you can say for a lot of these things, because most of them are not yet operational. Where are these Arizona fabs?

Where are these extra fabs in Texas? And honestly, the Intel fab in Ohio, they're not up and running yet. So we'll see.

It's going to be a real uphill struggle to try to fill this stuff out, and what it takes is real adults doing real work, not hype, hype, hype. And unfortunately, all we're getting right now is hype, hype, hype. And the same people who are asking questions about data centers will soon be asking questions about those semiconductor fabs that also take a lot of water and have an impact on the environment, no?

Oh, yeah. Yes, but so on the data center thing, in New Jersey, this data center that's being built in New Jersey, they literally had a 5,000-gallon leak of gasoline into the water supply. Right?

So this is not landing well. On top of that, I don't know if you guys ran the numbers. I got my engineering degree from Tufts University.

Only 3% of US engineering graduates go into- ... f*****g s**t making, right? And so make the math work.

Come on, folks. Dude, people can't even afford to go to college. Yeah.

Right? To get these advanced degrees, you still got doctors, and myself, paying student loans still to this day from school. And I've been out school a very long time.

So I don't think that we are all looking at this from a systems thinking, like shift left, shift right, and look at it from end to end to see reality. But everyone's in their phone playing with Instagram and building squirrel videos, so hey, here we are. I'm sorry, I wasn't listening.

See? See? All right.

I'm going to leave this here, but I would note that I think the Pope over the weekend put out a note saying, basically, hurting the environment is a sin. Data centers, by extension, seem to be hurting the environment, and they're being built to help AI, so maybe this AI a sin. There's an interesting conversation.

Who knows? You want to take that one on the text round gang? Yeah.

Well, I love Pope Leo because he's a Virgo and I'm a Virgo, and we're coming to a close of Virgo season, but I'm going to let him keep that one. And I'm sorry, white socks. Really, dude?

There we go. All right. I'm going to shift gears, though, because while there's another new report out and there's a think tank that is starting to look at, well, what is it going to take for us mere mortal humans to live alongside with robots?

And it will address everything from, I guess we got to think about if you're in a factory working with a robot, will the robot be safe? Will the robot go crazy and do something that might hurt you? Or two, running around your house and there's a robot that is now part of your life and, well, I don't know.

Do you want to get dressed in the morning in front of a robot that might upload that to something? I don't know. What is the lifestyle going to be here out of all of this, and how does it come together?

Steven, I'm going to toss this to you because you think about these weird things as well as I do, so how are we going to live with robots? I do think about weird things. And I do live with crazy monsters who might tear me to shreds at any moment.

They have fur instead of metal, but- ... it's the same kind of thing. You know what?

I was actually thinking, first off, we have this bizarre anthropomorphization of robots in our heads, thanks to sci-fi movies. Everybody thinks it's going to be like C-3PO or- No, Hal ... Blade Runner.

Hal. Hal. Hal.

I'm feeling like I'm terrified. Hal. I feel like, you, like Hal.

They like- Wait, wait, wait. There's a pod door I got to go close real quick. Hold on.

Well, and honestly, Hal is a lot more similar to the AI systems that we have. But the other thing that kills me, though, is combine this with that recent story about the LG TVs that are listening all the time and watching what you're watching and reporting back to their masters, and the fact that most TVs are sold at a loss now. Think about the incredible, evil, horrible data that you could steal from somebody.

Like you said, if there's a robot in your house that's watching everything you say and do, they could give those away. They could manufacture them in volume and give them away to people and just steal all their data. I hate the idea of always-on robots.

I hate the idea of this stuff being everywhere in our world. There's no way I would carry around a device that was constantly in front of my face and constantly listening to me and that I would do everything on, because that's just a terrible idea from a privacy perspective. No one would ever do that.

Oh my God. Excuse me. Steven, you need to tell me, what about Meta's glasses?

Never mind. I mean, you- Oh, can I jump in with a story about Meta glasses? Okay.

I'm so sorry to do... I'm not going to say the name. So my lawyer, I show up for a meeting with my lawyer, and he's wearing the Meta Ray-Ban spy glasses.

" I'm reconsidering even working with this guy because how stupid... I'm sorry. Sorry, dude.

How stupid do you have to be to think that those are turned off in a legal situation? Oh, jeez. Yeah.

Just because the light went on or is off doesn't mean it's not on, right? Correct. And people are covering them.

I literally did a training for a hospital system to show them how they can help the front desk staff and the clinical staff, front of office, back of office, identify it because they need to build policies and how are you going to talk to that person so you don't have a meltdown, but also protect the patients and the providers that are providing the service. Because once that data's in their glasses, where's it going to go? And you got to think that through.

So Steven- So- Go ahead ... are we all going to wind up in some sort of classroom setting trying to learn how to deal with managed robots and how to live in a robot society? Is that what's going to happen?

Or are we going to just kind of trial and error our way through this, and if we don't like what the robot does, we're just going to kick it down the stairs and see if it can get back up again? So it's funny that you say that, because there's research happening about what happens in offices and warehouses. But you know vi-coded shit's being created, and they're actually testing these out in rehab hospitals, elder care, dorms, homes.

The coexistent question is landing on, from my perspective again, my perspective, they get you with warehouses and offices. But right now, in clinical settings and settings where the most vulnerable people can't speak up, they're testing this stuff. And this coexistent question is landing in the places where the people in the room who are on their back has the least amount of ability to push back.

A patient learning to depend on a machine that's also learning on them is bad. Just like Disney hiring the new CTO from Character AI. How is that going to work?

You know what I mean? Just tell me, how is that going to work? Is no one thinking through?

The same company that just sacked Character AI a year ago, bad, you hire the CTO? Come on, people. How do you think that this is really going to work?

This raises more than just... It raises real stake, right? And it's just not about social etiquette, like privacy on, privacy off for me.

This is about consent, this is about memory, and it's about who the robot answers to. What we talked about at the very top. Who's controlling the robot?

Who's controlling the data? Who has access to the data? How is that data being used?

And we know where I'm going to go. People don't read the terms of service today, so they're going to scroll, scroll, scroll, click accept to use the freaking app. What makes you think they're going to be able to do that with the wherewithal when they sign in to go to a hospital?

Continue. Well, as I understand it, the automobile companies are going to be making most of these robots going forward. And on the robot, there's going to be probably a little GM OnStar button, and then I'll call somebody from help.

And Steven, what could go wrong with that? Absolutely. Right?

Yeah. The crazy thing is, on the flip side, I think this actually could be valuable. I think it's a thing that there could be incredible use cases.

You look at factory robots. They don't look like C-3PO building the car. It looks like a mechanical arm that just sticks out of the wall, right?

And it's under careful control, and it's not interacting with people. People conflate robotics and AI all the time because movies conflate robotics and AI. But robotics and AI are not necessarily the same thing.

They don't have to be. I don't want to be all movies and stuff, but R2-D2 doesn't look like C-3PO. It doesn't even have arms, right?

I think that the robotic thing that sits in a factory that's well-controlled, we have that for years, and I think we figured it out. I think the next wave of this stuff is these quote, unquote "humanoid robots" that can move around and do stuff, and I'm not sure we're ready for that as a society, but- I'm not sure either We're not. In New York City alone, over 3,000 of the startups are fintech and freaking healthcare, and they're creating these robotics with AI to help offset the challenges in the healthcare workflow, right?

So if you have a nurses shortage, oh, hey, send this thing in to monitor. That's- This is the type of robots I'm talking about. Like, hey, the big robot helping move big parts for and all.

Okay, got it. The robots that Amazon warehouses use, I'm totally cool with that. What I'm talking about is when you start bringing in this into everyone's everyday life, when you bring them into a hospital system, when you put them into freaking toys.

Do you remember Furby? I'll bring it to the next one. I have one from upstairs.

That was back in the day. The battery, it worked really, really slow when his eyes used to move. Right now, here, a machine that remembers your habits is useful and surveilling at the same time, but the person in a hospital or a kid barely signs an appointment contract.

And what I'm talking about, and I think, Steven, what I want you to kind of think about, I know movies are cool, is that these assets being put in scenarios with kids' hands, in hospitals, in elderly home sessions, assisted living buildings, end-of-life facilities. Helpful? For who?

That's all I'm saying. Well, I don't know about you, but when I was a kid, I asked Santa for a Big Track or a 2XL, and he didn't get me either of those things, so I'm still disappointed about mine. Here, let me just explain to people who don't know what I'm talking about.

Just freaking Google it. Okay, go ahead. Yeah.

I know what you're talking about. My cousin had a Big Track. That was cool.

So in healthcare, though, I've been on the receiving side of both people with awesome bedside manners and people with not so awesome bedside manners. So, are the robots going to be better than half of the healthcare providers out there and the people in that field, many of whom are already hard-pressed and in a bad mood because they're being asked to do a thousand different things inside of an eight-hour shift, or probably 12 hours in many cases. So, can it be any worse than it already is?

At the end of the day, it's a license, right? Let's just do a scenario real quick. So, if something happens not so great, and a robot was sent in a medical facility to look at all the patients, because nurses were on a different floor and to kind of offset staffing shortages, when s**t goes sideways, Mike, and there's a not great outcome, who's liable?

Whoever bought that robot and put it in the room. We went back to the same topic from the first one. Exactly.

That is something- Exactly. Who's liable, right? But no one's thinking through this, right?

Because when you end up in court, when someone's loved one doesn't leave the way they entered the hospital, the robot's not taking the freaking stand. It's not, right? So let's just think through that.

It's not taking the stand. You are. So before y'all go out here and get all crazy and buy code and throw robots and workloads and help extend, think about who has to show up and defend when stuff goes sideways.

It ain't going to be R2-D2, C-3PO, HAL, or whatever. It's going to be you. No, because I signed that user agreement that you talked about, and on page 492 that I didn't read, it said that if the robot kills me, it's not their fault.

Right. And- Cut and dried. Yeah.

Cut and dried. So, to the Techstrong gang listeners, read your terms of freaking service. And for the executives that are buying this stuff, seeing this pitch desk, think through how you're going to operationalize that, because at the end of the day, you showing up.

Right. Not the robot. And read whatever legislation is going through Congress very closely, because while it may be worded one way, it may actually be saying- That's true ...

we're going to absolve all these people of any liability in the name of national interest and being competitive in the world at large, and that would be just as equally bad. And maybe spend a little extra on a product that values consumer privacy and values your data, and so on. True, true, true, true.

So, Steven, so I don't know if you were on the receiving end. So, during the war that's happening on the other side of the pond, when the Amazon data centers got hit, and it didn't come back online, that section in your contract about force majeure, well, you ain't getting your data back. I mean, people, think about it.

I saw it. Your data's gone, dusted. It's force majeure, out of our control.

So just think about that. So, that line, how many of us looked at it? Well, when Superstorm Sandy hit, oh, that hit.

We haven't looked at it. And then when 9/11 hit, oh, we all looked at that. It came into play.

And now, when these wars are happening, and you're looking to balance your workloads across the globe for high availability, if your data center gets got bombed or whatever, your data's gone. And they have no obligation because you signed the contract under that section about force majeure. So think about that.

Okay. All right, folks, I got to leave it here. But, hey, anybody shows up saying they got your best interest at heart when it comes to AI, robots, or any of this stuff, think twice, because all of them have an agenda, and all of them happen to be here in New York this week.

So, thank you all for watching the Techstrong gang, and thank you all for being a fan of the show, and please stay tuned for the rest of the replay of the Techstrong TV lineup. And we have a little thing coming up behind us here. We are having a nomination here, or voting, for the people who are the most influential in quantum computing.

We're going to have a little ad behind that, so please stay tuned for that. Until then, we'll see you all tomorrow.