Sharp Tech with Ben Thompson - Biden’s Executive Order on AI, Innovation and the Steve Jobs Approach, Balancing AI Risks and Upside

Episode Date: November 2, 2023

The White House executive order on AI, the accidental lesson of all the analogies to nuclear technology, what Steve Jobs and Bill Gates can teach us about innovation, and a look at the competing systemic priorities that create AI questions without clear answers.

Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp and on the other line, Ben Thompson. Ben, how you doing? I'm doing well. I'm doing well. I've already given up on the buck season. The coaching approach is a disaster. So, you know, it's very important to face reality in the face, accept it, and then move forward.

So that's just a statement I want to put out there. And I'm ready to, ready to progress. Yeah, you know, we should pivot from the buck. as soon as possible. It's been a bit of a roller coaster early on. I will say it's been a pretty unwraquitted disaster. It has not been a disaster just for the record for objective observers. It's been all right. When it's looked great, it's looked really great. And man, oh man, when it's looked bad, it has looked really, really bad. So we shall see. We'll continue to monitor the progress. It has been a sobering first week or two here, though. I think it's good to be clear-eyed that

Nothing is guaranteed with the remainder of this buck season. But as for the remainder of this podcast, Ben, we're going to be talking about the Biden administration's executive order on AI. And we're going to start with what might be my favorite news brief in the history of this podcast. I'll read from the Associated Press. Biden was profoundly curious about the technology in the months of meetings that led up to drafting the order. Quote, he was as impressed and alarmed as any. anyone, Deputy White House Chief of Staff Bruce Reed said in an interview. He saw fake AI images of himself, of his dog, he saw how it can make bad poetry, and he's seen and heard the incredible and terrifying

technology of voice cloning, which could take three seconds of your voice and turn it into an entire fake conversation. The issue of AI was seemingly inescapable for Biden. At Camp David one weekend, he relaxed by watching the Tom Cruise film Mission Impossible. Dead Reckoning Part 1. The film's villain is a sentient and rogue AI known as the entity that sinks a submarine and kills its crew in the movie's opening minutes. Quote, if he hadn't already been concerned about what could go wrong with AI before that movie, he saw plenty more to worry about, said Reid, who watched the film with the president. So, Ben, as an initial matter here, have you seen Mission Impossible Dead Reckoning Part 1?

I did. And it was, you know, the AI was at some extent widely botched. I enjoyed it. I thought it was an interesting presentation. Have you seen it? I have not seen it. But that's interesting. You're not really a pop culture guy. I'm surprised you actually saw this movie. I mean, on Wednesday, you wrote an article expressing skepticism about the Biden executive order on AI. And I found myself wondering, maybe Ben hadn't seen Mission Impossible and needs to go watch that movie. So we understand. the scope of this threat. Maybe you didn't watch Mission Impossible closely enough, and that's what happened here. No, I see movies.

I saw several this summer. I saw Oppenheimer, I think, which is arguably perhaps relevant to this conversation as well. It's interesting because actually Mission Impossible, this particular movie, came up on Strateree in a Stratory interview that I did with Nat Freeman and Daniel Gross, the most recent one. We'll put a link in the show notes. And I sort of mentioned that, that I, you know, the vision. visualization of AI in that movie was ridiculous. But there were some sort of like underlying sort of philosophical questions that were raised by it that I thought were actually pertinent and interesting.

And it's interesting because Daniel Gross sort of jumped in right on that and expressed concern in that were the Malil that were the cultural Malil. Malou? Am I, yes, Maloo, whatever that we really struggle with the foreign words here on Star-Tec. It sounds French. Anything French is problematic. So no apologies to the French, though, just for the record. Yes. Is that... USA. USA. We eat freedom fries here.

But is the cultural representation of AI, represented by this movie, are going back to things like the Terminator, has historically and consistently been pretty universally negative and focused on sort of these destructive outcomes. Right. Dystopian outcome, sure. And the concern Daniel expressed is that this could be a war you lose on the cultural front before anything substantive actually happens. And it was going to be important to communicate and talk about the positive potential outcomes as well. I mean, you know, where AI actually is super beneficial, where AI actually progresses the sort of human race. Ray, I actually helped solve these problems we have that seem intractable, but if we could apply

massive amounts of intelligence to them, could be, you know, really move things forward. And we've talked sort of broadly, philosophically about the reason to be optimistic about technology is it is the only way to grow the pie, to actually make things better for everyone. And it's just so striking that this anecdote actually came up in the context of this sort of first pass at sort of regulation because it sort of I think gets to exactly what what Daniel was worried about, which is, look, we're operating in a cultural environment of broad-based skepticism of technology, particularly over the last five to six years for all the reasons we talked about in that podcast two weeks ago. I think that's very pertinent to this layered on

with this longstanding sort of AI as bad guy. And it's interesting because AI has been, it's been a stand in now that we have like a globalized movie scene we can't have movies about the Russians anymore right or you know a great point oh my God every villain is going to be AI so that we don't offend various

important foreign markets over the next that's right because the obvious movie villain is China right like I mean it's sort of like a geopolitical rival if the movie market was limited to or or if Hollywood was banned from China tomorrow we would have a lot of movies where China

And sort of the villain, right? But in this environment, you need something, right? And I remember this was a, you first started to see this, I think, I remember a discussion. I can't even, I have no idea where to cite this from. This was years and years ago. But in the context of the old sort of independent state, there was a discussion where they wanted to do a patriotic sort of movie, but in the modern environment, who can you cast as the enemy?

And so, oh, what if we did an alien invasion? And I think that was actually the flow in which that idea came about. We want to do an old school military sort of movie, but we need an enemy. So let's get aliens. So it's going to be aliens and AI. Nice. Inoffensive enemy. The whole world can agree on that.

Yeah. That was a key plot point in Independence Day with the world banding together and building its resources. Yeah, I mean, listen, that was a very thoughtful response to Biden's Cape David mission impossible. I really just wanted to note it for the record because I could not believe that that was a real article. I saw a screenshot on Twitter and was like, that's a kind of a funny satirical account to put together. So good job for that viral tweet. And then I see it's a real associated press article. Like, oh, my God. I guess this is how we're making policy now. I do trust that the Biden administration was not crafting this policy as a direct response.

to Mission Impossible, dead reckoning. But I enjoy the deputy chief of staff relaying that anecdote for all of us. But it does matter. Again, how do you say this work? It's very important. The cultural milieu. Yeah, the cultural Maloo. We're definitely both butchering it.

It matters. It matters massively. And I think this is going to get to a lot of broader issues around this executive order. Why I am skeptical and concerned about regulation. A lot of it has to do with the cultural, is it Maloo? I've never, I actually, that sounds so odd to me, but maybe it's right. You know, the blue in which we are operating and it's, I think, frustrating for technologists broadly because it is so ineffable.

And it's something that you can't sort of measure and you can't see, but it matters. It matters arguably more than anything. Right. Well, to anchor this discussion, we have talked a lot about AI. regulation at various points over the past year. For anyone who's not familiar with the broad strokes of this week's news, so on one hand, you have policymakers from 28 countries who are currently meeting at Bletchley Park in the United Kingdom, which was made famous by Alan Turing and UK codebreakers during World War II. And they are there discussing the dangers of AI and the

importance of global partnerships as this technology matures and evolves. And then on the other hand, there was the executive order that was issued from the White House on Monday. A read from ABC News. President Joe Biden issued a wide-ranging executive order on Monday that aims to safeguard against threats posed by artificial intelligence, ensuring that bad actors do not use the technology to develop devastating weapons or mount supercharged cyber attacks. The move stakes out a role for the federal government in a nearly half-trillion dollar industry at the center of fierce competition, between some of the nation's largest companies, including Google and Amazon. The Biden administration also calls on Congress to pass data privacy legislation,

an achievement that has eluded lawmakers for years despite multiple attempts. The executive order exerts oversight over safety tests that companies use to evaluate conversation bots such as chat GPT and introduces industry standards like watermarks for identifying AI-fueled products, among other regulations. The batch of reforms amounts to, quote, the strongest set of actions any government in the world has ever taken on AI safety, security, and trust. White House Deputy Chief of Staff, Bruce Reed, said in a statement, my favorite Deputy Chief of Staff of All Time. Directionally speaking, though. Why is it your favorite deputy chief of all time?

Because he relayed the Camp David Mission Impossible. Oh, got it, got it. He's quickly propelled up your rankings. Got it. Just rocket it up the power rankings. Directly speaking, I do have some concerns about all this, but can you lay out in broadstrokes what your concerns are and what you were writing about on Wednesday as you tackled all this? Well, let me go in the opposite direction. And to your credit, this was a prompt from you sort of last night. But I think it's worthwhile to start with why this might be a good idea and why it might be useful. and why my coming out pretty starkly opposed to it,

and frankly being a little bit more polemical than sort of analytical about it, why I thought that was necessary. If I'm going to articulate that view well, I need to be able to sort of state the other side. So I think that is actually a good and valuable place to start. And there is, I think the issue of existential risk is a real one. Like there is some tangent of truth.

in these movies and, you know, what happens if the AI is smarter than us, right? Like, we've been smarter than everything else in the world and have bent it to our will. What happens when we're no longer the smartest? So there's like a species argument here that this is something to be concerned about. And even if you want to argue that governments, you know, there is such a default lien to get the government involved. the government can help. We're here to help, right? And generally speaking, that has not worked. It certainly in technology, the government absolutely played a role as we can talk about

or discuss in some of the underlying technology, underlying investments. But to carry it to where it is was not a function of regulation. This was a function of spurring things, of pushing things forward, not trying to put limits on. But you could argue that, look, everything until now was not an existing. risk. This is an existential risk and therefore putting some starting the process of putting frameworks in place is appropriate. And in this view, you could argue the executive order isn't that bad because there's not that many. I call some things prescriptive. I think I probably overstated that. There is a lot of like putting things in place that are clearly going to be

prescriptive down the road. But for now it's just like, let's set this up to weigh out what rules we should have. Let's set this up to do sort of X, Y, Z. The prescriptive part was some specific numbers around compute and model size and all those sorts of things, and I think that's a mistake. But if you want to take the positive interpretation, I think that is that, which is, look, this, it is different this time. And usually you can dismiss, you can't say it's different this time, but I myself make that argument, right? We had the, the whole antitrust thing is my insistence that it's actually different this time. The fundamental market forces that, you know, that make these companies larger than any company's ever seen before by a massive margin

are because the fundamental economics and network effects and scalability of technology companies is different than we've seen before. So I am not one to dismiss the It's Different This Time argument. And I think that is the positive interpretation of that is like, look, the risk is real. It is different this time. And therefore, being more proactive and starting to get the frameworks in place, to address potential issues is an appropriate response to existential risk. And so just to drill down on the risk element then, to circle back to Oppenheimer,

basically what you're saying, if I'm understanding correctly, is if you look at AI like the nuclear technology of modern day, then it makes sense to be really careful and aggressive about regulating this and be cautious about the way you allow this ecosystem to evolve. Is that correct? Well, I mean, I'm trying to make the devil's advocate position here. But the problem is we're quickly running into issues about why I personally don't accept that position. Specifically, I think our approach to nuclear has been an abject disaster. We should be swimming in energy abundance. We should have like there was the proposals on, you know, in past administrations that we should have thousands of nuclear reactors all over the U.S.

Power should basically be free, which would be transformational. Energy is the foundation of all this. And we don't because the government got involved for existential risk reasons. And the problem is that the way it actually came out in practice is, you know, we have the Nuclear Regulatory Commission, which has never approved a nuclear reactor since it's come along. And they set standards that are like, it's like cost plus, but it's like safety plus. whenever you meet a particular standard, the standard goes up.

And so you have this situation where people say, oh, it's, well, it's way too expensive that. Why it doesn't happen? The being so expensive is a choice, right? And we take this sort of view of, well, imagine, you know, three mile island. Well, you know, yeah, let's actually calculate how many people actually died from that. Let's compare it to the cost of global war. It's compared to the cost of smog. Let's compare it to the cost of the downstream effects of not, you know.

And we had this whole thing with Fukshamaya where, you know, in Japan, No one actually died from that. The actual deaths came from people fleeing in the evacuation, but the actual, you know, it turns out a lot of the safety systems worked well. And even in an isolated spot. And the discharge has been handled well and all of that is perfectly safe now. But that's also, I just want to jump in. That's another great example of where the cultural conversation has distorted the reality of what that technology is and what it can mean for the future of society. and has led to bad policy.

Yeah, with a big mistake we make, and we did this during COVID, is making policy based on imagine bad outcomes. And when I say imagined, that doesn't mean it's not plausible. We, of course, you need to think about what might go wrong and work around that. But there is a tremendous risk that we've seen again and again, like with COVID. What if COVID behaves unlike any other virus and has all these potential. bad outcomes. We have to worry about that. Well, no, we should probably start with a Bayesian prior that it is a virus that operates like other viruses, make policies based on that. If we discover

things that operate differently, then we can adjust. Had we done that, we would have had a drastically different outcome from COVID. Instead, we were governed by our worst fears of what might happen, all of which is exacerbated by social media, is exacerbated by a broad lack of accountability as to who is making decisions, deferring to experts who do not have the holistic view, an abdication of responsibility by politicians, first and foremost. That's their job is to get feedback from a broad array of experts

and to make choices and then to be accountable to the people for whom they're making those choices. That system has fallen apart and we end up in a cover your ass system, which always defaults to the most sort of extreme sort of thing. That's what happened during COVID. That's what happened to nuclear. and it's one of those things where it's in the water. We don't even realize the price we paid for our utter and abject failure with nuclear.

Like, no one even talks about it or thinks about it. Global warming could have been solved. Like, we could have infinite power, which would unleash so many productivity, sort of possibilities and applications. This isn't a fairy tale sort of thing. It's something that actually could have happened, and we limited it because of our fear of existential risk. Now, hold on.

I just wanted the other aspect of the nuclear conversation, and we can return to AI shortly, if we were to move forward with a nuclear approach to solving the energy crisis, I would want government to still be very strict about the regulations imposed on who's building nuclear power plants and what safety precautions are used as you build those plants. I mean, that's one of the reasons the Fukushima water release. worked out great is because it was supervised closely every step of the way. And all those safety procedures have sort of evolved over the years. So it's an example of where some regulation is important, but you don't want so much regulation that you have paralysis. Yeah, that's a 100% correct and fair pushback. And I think the the challenge and concern that I have to bring this back to AI is it's fair to question how well we can actually walk that middle road.

Can we get the sort of just right amount, right? And, you know, and maybe your argument is that, look, the, the, maybe the probability is very low 1%, or 0.1% or 0.001%, but because the outcomes are so bad that it is worth embracing an approach that once it's in, you know, implemented, right? You don't have the, the, the, the, the, the, A folks being ultra-rational and implementing these policies. These policies are implemented by politicians, answering to interest groups, answering to their voters, are going to have all the complexities and challenges that come with that. Stuff never happens as neatly and cleanly as you want, which means the

bias matters. You're going to be biased towards allowing it or bias towards restricting it. because you're going to, which way are you sort of going to tilt? And again, just to say the devil's African-Marvoo, look, the potential downstream, if everything goes wrong, is so bad, it's okay if we're biased in the other direction and forego potential upside. That is, I would say, that's the, that's the sort of, you know, steelman argument as to why even going in this direction is a good thing. Right.

And that's logical and honest. I would say also that's not really the way it's presented, which is we're going to have the best of both worlds. We're going to have all the innovation we've all dreamed of, but we're also going to do it safe and responsibly. And I can understand why people roll their eyes at that. The executive order does not mention existential risk. That is the number one red flag and concern that I have. I just presented to you what I think is a defensible, logical argument. as to why it's worth running the risk of screwing this up from a regulatory perspective

because there is this outsized concern, you know, bad outcome that could happen. And yet we have all these words that does not mention this specific point. It vaguely obliquely refers to its safety, you know, blah, blah, blah, just like it obliquely refers to the potential upside and benefit of AI. But it makes me question the underlying motivation and concern where it's not just that we're going to be biased in an over prescriptive fashion in a way that limits growth. But actually, if the goals are not aligned with the real risks, we're probably not going to even address the real risks either. And I worry that is the direction that we are headed from a regulatory perspective. And it's downstream from this cultural point I've talked about.

what is in the air? Things like let's stop bias. Let's have equitable outcomes. It's not to say those aren't important, but if that is your motivation for limiting this technology, you're not doing anything about the downsides because you're focused on the wrong thing. And implicit in any sort of limitation approach is you're limiting the upside. And that's concerning to me.

Yeah. Well, I have two concerns of my own that I wanted to run by you. I mean, number one, we talked a few weeks back about the disinformation obsession that gripped elite institutions. And I will say that a good amount of overlap seems to exist in terms of the type of people who will broadly talk about how dangerous AI is and the people who five years ago were telling us that disinformation was the biggest crisis we had on the planet. Like what I read about this AI summit at Bletchley Park, I think back to like disinformation summits that were held at the Aspen Institute. I think like Prince Harry may have been involved at some point or maybe Prince William, some prince. And it's just like, you know what? A lot of those conversations didn't age all that well.

And so I just wonder, the AI risk conversation is one where everybody just sort of nods their head gravely as you talk about how dangerous this is and how important it is to be responsible going forward. And I think it is potentially dangerous and it is important to be responsible. But it's also a lot of the actors here were dead wrong like five or six years ago. So I just want to proceed with caution before putting too much weight into the concerns that are expressed in like White House press releases. This is actually a critical point. These are the folks that have been wrong about everything. And if you go back through the intellectual heritage, it's the same folks that stop nuclear. And the problem is you do have this group of very committed, very intelligent, extremely utilitarian in their thinking, very systemic in sort of thinking through what might happen, what might go wrong, and who do make, I think, valid intellectual arguments about how this could go wrong. Now, I am optimistic that their interpretation understanding is wrong in that I believe there is a distinction between like the

intelligence layer and the sort of motivation will layer when it comes to sort of humanity. I think there is a bit of over anthropomorphizing of AI that happens, but I might be wrong. I will freely admit that. The fundamental issue I have with this contingent, though, is I believe they fail to understand the real world and are ultimately aligning themselves and putting the best possible spin on efforts of an entire apparatus that is demonstrated over the last 50 years, but definitely over the last 10, that their goal is top-down control.

And, you know, particularly when it comes to information, when it comes to, like, the disinformation thing is 100% relevant for all the reasons you noted. And if they screw up, well, sorry, we were just doing our best. If they can do a Google search on Twitter and find a bunch of tweets that say, that say something is wrong, it's a global crisis. And it's indistinguishable from their desire for political power. Like, I am generally another thing I try to avoid is I try to avoid the overly cynical interpretation of motivations, right?

I think like money in politics, for example, is a great example. I think the vast majority of money in politics goes to politicians who already agree with the person giving the money. I am skeptical that money changes a politician's mind. They would have supported this thing, but they got a check for $5,000 or $100,000 or everyone I've been, and then they changed and then they did the wrong thing. No, they already agreed and they were getting support because of that agreement.

That's my general interpretation of how that works. And I think that there is some interpretations that refuse to accept that some people view the world differently and just disagree. And so they blame it all on money, right? For example, that's just one example of where I think you can have an overly cynical approach that actually misses what's going on. But you can also go too far in the other direction and give people too much credit. I touched on this article. It's awfully notable that the companies that claim to be the most worried are the most ahead, are not slowing down their

investment and just so happen to be supporting regulation that is going to kneecap potential competitors, particularly when it comes to sort of open source and broadly available models, which by definition are always the biggest threat to the integrated, centralized sort of entities. And I think that applies to these questions of, you know, why is it that the New York Times is so concerned about like the power and influence of tech and social media? Because that means they're less influential and less sort of have power. Again, the tricky thing about this. And look, this applies to me as well. I try to be cognitive of my biases, but for sure they exist, is that you can 100% believe what you say and believe what you write and be totally blind to these

motivations, but that doesn't mean the rest of us have to be. And I do worry that this group that is genuinely concerned about existential risk is being incredibly naive about the reality of the folks that they're allying with. Yeah. Well, let's be clear on one point. As far as money and politics are concerned, the way policy is actually set is the president of the United States kicks back at Camp David and watches a movie. and decides to get active on AI. The real problem is Tom Cruise in politics, not money in politics. Exactly.

Alibaba. I'm going to play two clips that you included in Wednesday's article. The first is from Bill Gates talking about the future of smartphones and mobile computing. And the second is Steve Jobs. This conversation happened in 2007 at the D5 conference. So first, here is Gates. So what are the core functions? of the device formerly known as the cell phone, whatever we want to call it, the pocket device.

What would you say the core functions? Like five years out, what are the core functions of that pocket device? How quickly all these things that have been somewhat specialized, the navigation device, the digital wallet, the phone, the camera, the video camera, how quickly those all come together, it's hard to chart out. But eventually, you'll be able to make something that has the capability to do every one of those things. And yet, given the small size, you still won't want to edit your homework or edit a movie on the screen of that size. And so you'll have something else that lets you do the reading and editing and those things. Now, if we could ever get a screen that would just roll out like a scroll, you know, then you might be able to have the device that did everything.

And now here is the Steve Jobs answer to that same question from Walt Mossberg at this conference. You five years from now, what's going to be on that pocket device? I don't know. And the reason I don't know is because I wouldn't have thought that there would have been maps on it five years ago. But something comes along, gets really popular. People love it, get used to it. You want it on there. So people are inventing things constantly. And I think, the art of it is balancing what's on there and what's not on there is the editing function. So can you explain how those competing visions are relevant to the AI conversation? Because I thought it was a pretty profound part of the article on Wednesday and just a profound

observation on how truly transcendent technology ever emerges. I mean, I'm glad you had the second part that you thought it was profound because when you ask me to explain the connection, I worry about, I hope the connection. landed that I was trying that I was trying to get across. It was a broader commentary on how we invent new things, how we have these new technologies that increase the pie that come about. And the way they come about is through openness, through humility, through not knowing what might come next. And what was so striking to be about that conversation is if you step back, particularly with the, you know, looking back 10 years. But even at the time, Gates was filled

with so many facts and so many observations that it's almost like he was, he couldn't see anything because there was like too much of a jumble sort of in front of him. And like he was right that navigation computers and and music players would be consolidated as one, is one device. He was correct and he starts talking about,

oh, but can you edit a movie on there? Blah, blah, blah. You need an unscroll, a scroll out screen. That ended up being totally wrong. Like what, what of the policy? challenges we face his TikTok, guess where those get edited? On the phone. So one question I have about the Gates quote, is it possible that he was just in a situation where he is the CEO of a public company and can't be up there on stage saying, yeah, who knows, you may not even need your computer.

He may not even need PCs in the future. Probably not his sole driving motivation there, but it occurred to me in the aftermath of the article. It's a fair question, except for the fact that Microsoft continued to pursue the PC-centric strategy for years after that interview. And I don't think he was CEO at that point. I think he was like chief technical office or something. He had some sort of special role. But obviously he still drove a lot of these sorts of things. But the point of going back to Gates is just to reestablish what is the mindset that actually leads to new and better things.

And to my mind, the number one thing is it's humility. It's doubt. It's understanding we don't know, right? I mean like Steve Jobs who had just invented the iPhone in conjunction with lots of smart people at Apple, which was a total reimagining of the way what a phone should be. But like obviously it should not carry over paradigms from the desktop computer. It needs to be something new. Who is more qualified than anyone to speculate on where it would be in five years, have the humility to say, look, who knows, right? Right.

And I don't think that was fake humility. I think that's inextricably aligned with how Apple got to where they were at that point and how they've gotten to that going forward. If you start to prescribe the future, if you start to be like Gates and say, well, this is probably going to happen and that's going to happen. And then in his whole comment, some of which was right, some which was speculative. Oh, we heard about this technology. He would do this. He's like, but for sure, people will definitely have a PC and a phone. That actually ended up being totally wrong

And that was actually the core problem he had He couldn't imagine a world without PCs He couldn't imagine a world where there'd be 2 or 3 billion smartphones in the world And only like a billion PCs Which by definition means there's like 2 billion people That don't have one at all And the reason the application to AI

Is I worry that there's way too much confidence About what is going to be useful And what is not going to be useful It's like we You mentioned it sort of at the beginning Like there's a lot of language, well, we want to make sure we allow innovation and we restrict this. That entails a high degree of confidence to which I would call arrogance about we know what the upsides are and we know what the

downsides are. And we, your benevolent government with a horrible track record are going to thoughtfully impose guidelines to make sure we get the good stuff and to make sure we don't get the bad stuff. And the issue here is not, it's not an issue of intention. It's an issue of structure. Once you start going down this road, by definition, you don't know what's going to happen. And you're going to cut off stuff and you're going to end up in a situation where take my Buck's analogy.

I feel bad for Damien Willard. He is going to get blamed for the Bucks having a bad defense because he's a bad defender. And that was the concern coming in. My argument is that their, their defensive philosophy is stupid. It does not play to their player's strengths. You have Brooke Lopez running around like a crazy person. And what happens is that poor Damien Lillard is going to get blamed for the Bucks's defensive troubles when it's not actually his part. We haven't actually seen what it would look like in a normal defensive philosophy that actually plays to our player's strengths.

Only then could we determine if he's the problem. Okay. This is the same sort of issue that comes about being too fast to put rules in place. There is stuff that will not be invented. There are things that will not happen. There are possibilities that will never manifest themselves. And there will never be a moment of accountability that says, look, we would have invented that but for these laws we put in place because we don't know what could be invented.

And we end up shutting off things in the future. And we don't even know what we missed out on. Well, and that's why I love the Steve Jobs quote because to me, that neatly encapsulates everything I find exciting about the AI. landscape right now. It's like the entire world is focusing on this little ecosystem and nobody knows exactly what it's going to turn into and what platforms might come to dominate in five years, in 10 years, in 15 years. And that's a pretty exciting place to be with the entire world and a lot of really smart people paying attention to this space. And to the extent that we do get

breakthrough transcendent technologies, I think it's more likely that it's going to come from sources that are beyond that centralized four or five companies that are going to be doubling down on the advantages they already have. And also, to the extent that there's risk, like, I mean, my first reaction, and I think I've expressed this elsewhere on this podcast, like, I don't even know exactly what we're regulating right now, because the AI ecosystem is so amorphous. It's hard to do this in any coherent, targeted way. And if the craziest risk are going to materialize, down the line, I don't trust today's policymakers to know what levers to pull to prevent that from happening. And it actually goes back to your discussion with Greg Allen about the chip ban last week

and something that Bill Bishop has said a number of times on Sharp China talking about the chip ban, which was the problem with the first version of the chip ban, was that we wound up incurring 100% of the costs associated with an aggressive decoupling strategy in the semiconductor space but the language of the first ban was porous enough so that we weren't realizing all the intended benefits. And so we were truly like cutting off our nose to spite our face. And I think something similar could happen here where I don't necessarily know that any amount of regulation is going to guard against like the true existential risk that is presented by artificial intelligence. And And in the meantime, if we're foregoing some of the innovation upside, then I'm not sure where that leaves us.

You know what I mean? Yeah, I completely agree. I think it's a very good articulation of the challenge, which is even if you believe in the risk, you have to consider your capabilities as far as enforcement goes, right? This applies to the chip ban as well. It's like, sure, we're going to stop China from developing AI. Okay, great goal. How are you going to actually do that?

Right? And then you have to actually get down into the details. And you're going to say, oh, try to develop seven nanometers. Well, yeah, of course they did because they already have all the capabilities. If you understood the space, you would not be surprised. So what you have to do is actually find a point of leverage. A law is worthless if you don't have the means to enforce it. And by the way, if you choose the wrong lever to enforce it, you're going to have all these bad outcomes that you sort of didn't intend. And I think this is the mistake a lot of these folks make. I am not denying the problem of AI risk. think it is a legitimate thing. Again, I would say that my view on the probability is quite a bit lower than a lot of these folks, but that may be motivated reasoning because I generally, because that more broadly aligns with my argument. I will freely admit that that is the case. I'll put that on the table. My concern is that I think that risk is inescapable. I don't think we can fix it with government regulations. I don't think they, and not just because government isn't the right tool. Just I don't think particularly in our current environment, it is a viable, it's going to remotely

work. And so if we're going, if you accept that as the premise, then why would you want to run the risk of all the downsides that go along with that sort of attempt to limit the risk? That is probably, that at a core is my view. I am accepting the risk because I think it is an inevitable risk, an inescapable risk. Once you take that as an assumption, then I think you should be terrified. of the government being here to help of sort of coming in to sort of put these limitations on. Yeah. And to elaborate further, when we talk about foregoing innovation, foregoing upside,

it's not just like foregoing an AI pin or something or new products here and there. It's foregoing like a key driver of economic growth and potential strategic advantages over the next 20, 30, 40 years by seeding an advantage in this space that we, the United States. States. Shout out to the listener who complained about me being too pro-America. Beceding an advantage that America has had as sort of the leaders on the forefront of technological innovation over the last 30 or 40 years. We're giving up on humanity. This is how we, again, it's just a fact that the only way you grow the pie for everyone

is through technological progress. And technology broadly speaking, not big tech companies. We're talking about the wheel, right? We're like just increasing productivity. is the, otherwise we're just fighting over a shrinking pie. That is not a world you sort of want to be in. If we want to actually achieve, if we want to achieve sort of, you know, this massive amounts of energy, the sort of thing we talked about.

If we want to achieve sort of incredible sort of breakthroughs in sustenance or capabilities or if we want to explore the stars, like having a massive amount of intelligence, if we want to solve diseases, having a massive amount of intelligence at our disposal has these huge upsides. And it absolutely, I am not, again, I'm not at all dismissing risks. It's just that I think there's a tendency to avoid the question of implementation. And the reality is if you could theoretically control everything and there's a certain commonality in these ultra-smart people that are super obsessed with these risks, it's sort of the same failure state you get with communism or something like that. It's like, look, if we could actually just make it work,

then it would be great. But all implementations have to deal with the reality of humans. That's to do the reality of governance. It has to deal with the reality of coordination. And the problem is if you start getting into rules, the good guys are the ones that follow the rules. The bad guys do not. The other countries that are not governed this by do not.

And the real spur in sort of investment and motivation is going to move on to illicit open source models. that are best operated locally or on relatively contained compute so you don't have to register your data center with the government, which to the question we had on the podcast last week is good for China. So you're actually structurally shifting the market to a more distributed sort of and compute. Like broadly speaking, if the dominant sort of application of AI is in large data centers by large companies by and large. And ideally there's competitors and we're progressing. But if that ends up being the structure, that's good for the U.S.

government. That's good for the U.S. because they have the tools and stature and position in the world to sort of influence and exert control. If you start limiting and cutting that off in pushing everything else to under the table, to more distributed, to more local, you're sort of like encouraging. It's like, like, to go back to the nuclear example, to have nuclear visible and above the table is better than having it sort of like under the surface, right? But at least nuclear you can track uranium. You like there's like a signature or plutonium or whatever the sort of things are. We're dealing with bits with software. We're dealing with chips. These are things are hard to track and hard to keep track of. And the overapplication of the nuclear analogy is, is to me, a big mistake. Yeah, and to be clear, practically speaking, this is going to be so much harder to contain. Like, I understand the best case in error for the U.S. government is that it is a highly centralized industry where you have to have incredible amounts of compute capacity to really be on the forefront. But realistically, there's just as good a chance that it is more widely dispersed, and the open

source model of the future of AI comes to predominate. Is that right? like whatever the, whatever the U.S. does. Well, I mean, I think there's path dependency in these sorts of things, right? One of the things I've talked about in the past is the sort of path dependency of the Chinese manufacturing versus Western manufacturing, where China had a labor advantage for years and years and years. And so they have come to dominate in the industry that is labor intensive.

And so the natural response of Western countries was to be capital intensive. Now, over time, China's labor is quite expensive. now, right? They actually don't have a labor advantage, but because of the path dependency that got us there, they are utterly dominant in any sort of industry that requires huge amounts of labor, okay? And because they have the whole infrastructure around it, all these sorts of things. So I think there's a question about the long term how AI has actually been in the world. Right now, to be super large and be super centralized is advantageous, as we would expect with any new technology magnified by the compute requirements of these things. To run a big data center with a 32,000 Nvidia chips all linked together, right now, is the best way to sort of push things forward. And if you followed a path dependency sort of model and just let this develop, there's a good chance that approach would become the dominant approach to delivering AI services. And having large centralized entities providing these services is from just a geopolitical perspective and along from the U.S. government, if you're the top dog that can basically

reach your hands and do anything in the world, as we see through the financial system, for example, that's actually a good outcome. If we tip the path dependency such that actually the cost and complexity and latency and issues of a centralized provider is actually it's better to have it local. Apple just released a laptop chip with like 190 gigabytes of memory. Like still not enough to run a huge model. You can run a 13 billion parameter model, maybe even a slightly larger one now. If we keep going down this road such that you get good enough AI in a completely ungovernable fashion because it's on the, these open source models that you got on tour or whatever it might be.

You're running it on your local computer. And whatever sort of concerns they have that you're going to generate recipes for a bomb, which, by the way, you can get on the internet right now, you don't have control over that. You actually don't. And I don't think it's for sure that we would, that that's inevitable, which way it's going to go. If you start putting real fundamental limitations on innovation on those centralized things, that you have to get a license and you have to be with the government, you have to, like they have this policy board, that's like,

23 people spread across multiple departments. Who can actually afford to do that? Microsoft can afford it. Google can afford it. Can startup X afford it? No. Just to simply navigate the bureaucracy is going on. So any sort of innovation and push is going to go towards in this other direction.

You might end up on a different path where it's now fully distributed. Now, maybe you're uber skeptical of the U.S. government. And you're like, good. Like more regulation because I want to push the path dependency in this other direction. You're going to start sounding like a Bitcoin absolutist, but I mean, it is sort of a valid sort of point in this case. And that is a risk that is being run where you're actually going to accomplish the opposite of your overall goals because you're trying to exert control too early. Yeah. Well, and is there a disincentive then to like in terms of this executive order,

it feels like they're making it harder on any new entrant who might be starting from scratch and might emerge from this open source ecosystem, like, and taking advantage of, for instance, the models that Facebook has open source. And is it going to be harder for a company to do that in America after what's been issued this week from the White House and the signals that have been sent? Right. Maybe that was issued this week. But if we keep down this path, I mean, to be fair, a lot of the executive order is just setting up committees, setting up studies. We need to start prepping rules. There's something like in 270 days we have something in place, XYZ.

And so, but again, I think that to think it's going to stop here is misguided. Like, like, this has been one of my biggest learnings over the last sort of decade. This ties in, ties into like the free speech discussion. Slippery slopes are real, right? Like, I like, it just once, once you, in particularly in technology, once you develop the capability of doing something, something, the pressure to leverage that capability becomes overwhelming. So once you set up the ability to sort of, look, we need to stop this clearly bad misinformation that is going to be pushed until you're reaching way further than the original

positive motivation was directed towards. And so I am concerned about that where once we start down this road, the ratchet only goes in one direction. And that's why I'm concerned about the ratchet being placed at all. And just I just want to double back on my point about centralization. The tech industry famously on sort of not regulated, right? The most substantial piece of regulation about tech was section 230, which was actually like anti-regulation. It basically shielded tech companies from lawsuits that they probably would have won anyway in the long run on free speech grounds.

But like we're just like super problematic. It actually did lose something like we'd get into the history of 230 at another time. Well, it would have been unbelievably expensive to litigate lawsuit by lawsuit by lawsuits. It's not viable. It completely shut out like like. You want to talk about a barrier to entry. Yeah, exactly. That's right.

A brilliant piece of legislation for, for lots of reasons. It's completely misinterpreted and misrepresenteded what it is by both sides of the aisle for different reasons. But it actually was a very positive piece of regulation in my estimation. So what happened was because of the openness and lack regulation, there was no government. thumb on the scale as far as

as its path dependency. And what happened was this is aggregation theory is my, the whole point in like the foundational business trajectory is everyone misunderstands the implication of how the internet works. Everyone thinks that because it's fully accessible,

we're going to have fully distributed. Everything is to be distributed. There's going to be like no more big players because anyone can have as access. And the whole point of aggregation theory is, of aggregation theory, is no. When there is no friction, you have an overabundance of information such that someone who can marshal demand, who can serve customers to help them just find what they're looking for, whether that be friends, whether that be information, whether that be shopping, they will actually become massive and massively larger than any company we've seen before.

And they will flip over market power instead of controlling supply gives you power of demand, controlling demand will give you power over supply. all the things we've talked about the context of Amazon, of Google and Facebook. And that is a downstream from there being no regulations in a frictionless market. And so what happened in the long run is the U.S. government exerts tremendous power over the internet because the internet is controlled by these very large entities that are in America. Which, again, it goes back to the long-term strategic advantages inherent to all of this. That's right. And so, like, just, I just want to sort of double back on that path dependency point because we've already seen it happen. Yes. Well, so I agree with a lot of what you've said on here and all of the concerns that I expressed earlier are sincerely held.

But I have to say, you know, I read your article on Wednesday and I would have been 100% on board like a month or two back. but now I'm not 100% sure what the right answer is. And I go back to the disinformation conversation a few weeks ago where I feel like that was a pretty good episode, but I also feel like it was a little incomplete. Like one of the things I kept coming back to on that show was the notion that the constant panic over misinformation on social media is a red herring. And misinformation from institutions can do exponentially more damage. and I'm essentially recapitulating various arguments on Stratory.

But in thinking about it more, particularly over the last couple weeks, I actually think that focusing on the relative impact of one piece of misinformation on social media or a couple pieces of misinformation online versus a piece of misinformation in a traditional outlet, like the New York Times or the CDC or something like that, it's the wrong way to measure the relative impacts on the ecosystem. system where what actually feels like has reshaped society is the constant exposure to media, the way it's tailored to your interests, how quickly it spreads one version of the news or one framing, and the way we've just seen like social consensus becomes so fractured in all sorts of

harmful ways over the last 10 or 15 years or so. I don't think we fully comprehended the impact of social media. Like, I don't know exactly what TikTok is doing to change the way young people see the world. And you made this point that part of that story is the way social media has changed news organizations. Yeah, exactly. And so all of this is difficult to quantify. But I think in the grand scheme of things, while the focus on disinformation on social media

may be a red herring, I think it's possible that what it's distracting everyone from is the real problem, which is just social media itself or social. certain forms of social media, TikTok being exhibit A, but you could cherry pick examples from anywhere over the last 10 or 15 years. And I only mentioned that as an example of technology and innovation that hasn't necessarily been a net positive at a systemic level and a data point that was lingering in the back of my mind as I was reading about the possibilities of AI that might be sacrificed as a result of the attempts to regulate it now. I mean, if we could go back in time and just not invent TikTok and cut TikTok out,

cut it off at the knees like five or six years ago, that probably would have been a good thing. And there are tradeoffs inherent to all of this, but I think it's worth considering that even if it's not about preventing like the existential risk that's downstream from AI, 20, 30 years from now, it's also possible that, you know, regulating this carefully because AI is going to be so much more powerful than a lot of what social media has done to society, being super conservative about it, even if it means sacrificing some upside in terms of innovation, isn't the craziest thing in the world. Does that make sense to you? it makes sense as an argument. I'll tell you why I disagree, but first I have to just express sadness that you, you know, so blatantly

attacked TikTok. You know, very, very sad. You said, you wish it had never been invented. We've been cut off immediately. Did I say that every feet? Oh, my God. It's happened to be on Sharp China, too. It's what I deserve for inflicting that name on the mask.

I wanted a different name. I can't remember what the name I wanted was, but yes, it is your own fault for the record. Okay. Fair enough. But you... What's worse than a mispronunciation? Just a blatant attack on my parenting advice.

I know it's hurting our brand equity here. We've spent, you know, blood, sweat and tears trying to build up the TikTok empire. Anyhow, your point is a correct one. And I am not here making this Pollyanna argument that tech is an unallied good. I've been consistent in not saying that. Let's take the social media a bit. If you want to control...

the issue of social media and ensure everyone has the same frame of facts, then we should go to Xi Jinping and ask to sign him to a consulting contract. That is the answer, is you have to legislate things, you have to control things at the ISP level. You have to punish people for breaking the consensus. You have to have a top-down censorship regime that is most effective when people start to internalize it and they know not to even say things that aren't right.

And again, I am not, I'm not going to stand out here and say that's 100% bad. There are real benefits that come from societal cohesion. There are real benefits from troublemakers knowing to not lift their head up too high. Like, like, if we're totally honest and are willing to let go of deep rooted assumptions that we don't even realize we have. And this is perhaps easier for me because I am in Asia, because I'm close to China, I'm in a Chinese society. it's shocking the things that you believe about life that are just totally unshared by the people around you, right? And that ought to instill doubt and humility that, man,

I might be making assertions of fact that are actually opinions grounded in my cultural background, my cultural milieu or whatever it is. And so that is beneficial. But the whole point is, yes, that is a fair point. But it's why implementation matters. Are we going to implement a China-style system to have a consistency of sort of belief and a shared set of facts? Is that going to happen?

Do we want that to happen? If your answer is no, then I think your concern should shift to not getting stuck in the middle. To make sure, look, if we're going to bear all the costs of everyone having an opinion, everyone being able to publish, how can we best leverage that to reap all the benefits? And I would say the same thing applies to AI. If you are worried about existential risk, there is a logical thing to do. And to Elisor Yukowski's credit, he has said this, destroy all the data centers. That is the answer.

Don't even go down this path. Break out the bombs. And if you're going to pursue it, this idea that you can straddle the line to me is a mistake. like, if you're going to go down this path, if you're going to bear the risks, then at least don't sacrifice the upside. And the mindset I am concerned about is everyone talks about risks and no one is talking about upside. And that is understandable because we don't know what the upside is. What we need is Steve Jobsesque humility that I don't know. But what I'm going to do is I'm going to make sure we put in place an approach and a structure that we, that acknowledge

the risk that we are taking by doing this. But let's make sure we don't end up risking the upside because we're trying to have it both ways and be obsessed with the downside. Yeah. No, and I think that's well articulated. And I don't know is, again, what's most exciting to me about the short term future here, medium term future. I just, it occurs to me, it's possible that AI is elemental to the strategic future of the United States.

It's also possible that the long-term strategic advantage will go to the country that best understands the downsides of this technology and decides to get aggressive about it to mitigate the potential negative impact on its population. Exactly. There's a broad point where that is reality for everything. There's been, like, we live in such richness and such relative safety. We've forgotten there are real risks in life and that progress happens with people dying and with bad things happening. That's just sort of reality. And the concern is that we become so obsessed with safety.

We become so upset. Like, again, not to go back to COVID, but like being obsessed with safety at the cost of children's future, was that actually, in retrospect, a risk worth taking? the kids have been back in school as soon as possible. I would argue yes. And like that's the sort of, the reason we go back to it is it's such a beautiful object lesson. And it's both sides will go back to it.

It's pretty funny. But like to my mind, that's my takeaway. But what about TikTok, not TikTok. Do you think it reflects an obsession with safety to say, all right, in 2017, would it be a net positive to just ban this and prevent it? You make tradeoffs. No, having a foreign entity control such a massively important media channel is stupid. Like, that's the long and short of it.

That's why I wrote, I supported the band. I wrote it. And I say this as someone that has been on here talking about free speech, blah, blah, blah, blah. Look, life's difficult. We talk about this time. You're going to be hypocritical. You're going to have paradoxes because life is a series of tradeoffs. And in every decision, you have to think through what are the tradeoffs.

And so I will sit here with this AI bit. That's why I wanted to start with the devil's avic. get position, which is like, look, there is an issue here. There is an issue of existential risk. And it's worth sort of acknowledging that. Now, I think in reality, it's inescapable, so let's make sure we don't miss the downside. I am worried that all this regulation ends up being in service of political goals that are about controlling information and what people see or whatnot. Right. Political and corporate goals and your power in whatever facet it appears, financial or cultural or whatever it might be. But let's let's try to not.

do that. But tradeoffs exist. With TikTok, I think it's insane. We let China have such an influential platform, which with documented influence on it. Again, like I was, before I wrote this should be bad, I talked about it in the context of the Hong Kong protest. It was clear the thumbs were on the scale, at least for me here in Taiwan. And it's nuts that we allow it. And it's nuts that we're issuing executive orders about this stuff when that stuff is so clearly a You know what? That is a fair complaint. And it's funny because over the last couple of weeks, I've wondered, you know, is someone in China putting their thumb on the scale to determine what news is served to young Americans?

I mean, they removed Israel from, they removed Israel from online maps. Like, I think they pretty signal clearly their point of view. And again, who knows if that's the whole problem. We can't trace it back. Who knows what thumb is being put on the scale? That's what occurred to me. I was like, you know, it's an incredibly black. box and we just have no idea, which is part of the problem with all this.

This was back to the whole thing, right? Like, look, maybe you're just a anarchist or a libertarian or you think the U.S. government is just as bad as China. Like, I'm not going to argue. We both have put it on the table that a fundamental premise of this podcast in our world views is that we are Americans. We are going to be biased in that direction. We are going to look at and recognize the shortcomings of our government, the shortcomings of our culture, and make the explicit choice that when push comes to shove, I'm going to be pro-America in these various policy discussions. So that's sort of on the table. And I understand and respect if you disagree with that, even if you're an American disagree with that. That's fine. You're free to do what you want.

We will put the bias sort of on the table. In that sort of context, would you feel better or worse about TikTok being owned by a U.S. company or being owned by a Chinese company? That's just sort of it. Well, and not only that, TikTok could just cease to exist because I'm not sure, even if it were owned by a U.S. company, it would be good for society. Right, but there's a market, there's a market demand there, right? So Reels becomes huge. And then like Facebook, Marks, I work is that much more power, right? I think there's, there is an angle where TikTok existing is one of the best thing that's happened to social media. It's certainly spurred Facebook to not be lazy, to get back on their horse, to really sort of invest and push forward.

It would be disingenuous for me to sit here and act like it's all negative. That's part of why these questions are really difficult because there are, different considerations. There are positives and negatives inherent to all of this. And striking the right balance in terms of regulation is incredibly difficult in its own right. And also realizing there's a difference between legislating the past and saying like, what if we had gone this direction and sort of being honest about where we are and what possibilities we have in front of us and how to progress in that regard. Would we, would it be better if we had just, you know, three newscast and that's where I haven't got their

news? Maybe. But that the ship has sailed. So, So what can we do going forward? Yeah. No, and the only reason I raise it is as a data point to underscore that, you know, tech and innovation isn't necessarily like purely good in every circumstance. It's a thing. It's a thing that exists. Humans decide if it's good or bad.

And I think this is, you know, this is where I get to the, you know, if we want to get philosophical, the anthropomorphization of AI that I think that happens, is imputing goodness or badness into it when it's people that are still making the decisions. And maybe, again, maybe one day they won't. I'm skeptical, but I acknowledge and accept that risk. But we need to think fully through about the possibilities. It's hard. And we all need to move with humility.

You know, I think there are certain people who have been involved in tech for the last 15 or 20 years, for whom the emergence of this new economy has been nothing short of miraculous. And there are also people in the rest of the country for whom this has been sort of a mixed bag. And it feels like everybody has whiplash after the last 20 years because in some ways it feels like everybody's gone crazy. Well, it's a mixed bag because they don't actually account for all the consumer surplus they have. They don't account for the fact that access to the world's information. They have access to group chats around the world, one of our favorite past times that you can talk to. Can't wait to get group chatting.

account for the fact they have access. You can watch a million more sports than you could. We can become F1 fan, something that was not possible when there's only a broadcast channels. They don't account for the fact there's a gazillion TV shows online and things we can watch. They don't account for the fact that they can, they basically, their entertainment value is drastically higher and the cost is incredibly low. You can pay $60 for a video game and play it for 100 hours.

Like the consumer surplus of tech is astronomical. And the reason we can bear things like ever increase. deficits and all these sorts of things is because we have this massive deflationary force that is actually making people's lives materially better in a way they don't have to pay for. Now, is it making their lives spiritually better? That's a whole podcast, right? Is society better? Systemically, there are questions about are we in a better place now than we were 25 years ago?

And how do you measure that is its own question. I talked this before. Like, maybe there is some value in this struggle. Maybe there is some value in remembering humans die. You have to fight for things. You don't just get to have these beliefs for free. Go out and interact with the world. Yeah.

So I mean, yeah, but that's a philosophical discussion, which is fair to have. But you have to be humble in that discussion as well, which is understand where we are and what capabilities we have and have that framing what your prescription is for the future. Right. And all I'm saying is that I understand where the impulse is to regulate some of this stuff. If you've lived through the last 20 years and you're just like, holy shit, what just happened to society? And now this new technology is going to be 10 times crazier and transform society in ways we can't even imagine right now. Like, I understand the impulse to say, okay, let's be really careful about this. But I also am a little bit worried about the practical implications of any of this.

and how you actually regulate any of it effectively. But I appreciate the humility on your end of the line here. And, you know, thank you, Steve Jobs as well. I think you can agree. Thank you, Bill Gates. I was a little rough on it today, but hey, what are you going to do? Hey, you know what I always say? The best takes are the worst takes because those are the takes that you remember for years on end

and can laugh about for decades to come. So thank you, Bill Gates as well. but then until next week, I wish you the best with the Milwaukee Bucks. The in-season tournament is tipping off this weekend. I'm so sorry. He was so forlorn even before we hit record on the podcast. So I'm genuinely worried.

I'm not lying. I'm not lying. I'm good. Well, I know. I guess he's accepted it, which is the final stage of brief. I'm insufficiently foreworn, which is making you concerned. You're concerned I'm a psychopath. I get it. that's what I'm concerned about.

I'm not all that worried about the AI robot from Dead Reckoning, ruining our futures. But I am worried about my friend, Ben. Exactly. Put Brooke Lopez back and drop coverage, God damn it. All right. Well, on that note, we'll circle back next week. If you have questions, email at sharp tech.fm. And Ben, have a great weekend.

Touch some grass. And I'll talk to you soon. Talk to you later.