Speaking of economies, anonymous asks: how much of a role do the big AI companies have in selling their vision to people? I personally think that a lot of the coverage they get is unfair, but instead of whining about it, what can they actually do to counter it? Or is there no way for them to do that? You have people attacking Sam Altman's house, data centers are getting protested across states, across political lines. The sentiment toward AI by the general public is swinging wildly negative. What can and should the AI labs do about it? And SB asked a similar question and sent in a variety of polls and news stories indicating that AI is hugely unpopular among the American people today. I have some thoughts here, but what do you think of that question?
Well, you go first. I talk a lot.
Well, you do talk a lot, but hey, like...
I got this question and it's actually super interesting and I might write about it on Sharp Tech next week. I'm of two minds answering because on one hand I think there are definitely things that the AI labs can do. And first and foremost, I would start with aggressively policing some of the misinformation about like data center pollution and water use and even energy to some extent. Like there's just a shockingly pervasive idea that AI infrastructure is uniquely harmful to the environment, which as far as I can tell, is not really supported by reality. And I don't know what they can do to police that, but I would be very aggressive with their messaging in that regard. And then on the other hand, I think one problem with AI messaging is ultimately going to be reality. Like it's hard to win on messaging when observable reality is telling people the opposite and most of the people who oppose widespread adoption of AI are probably right that it's going to be pretty disruptive and there may not be clear benefits to the current generation of workers. And all of this is happening in an environment that's already been disrupted by tech for the last 25 years. So people are already frustrated. So it's a real uphill battle that may not be winnable for companies like OpenAI and Anthropic. It also does not help that Anthropic is constantly talking about like the real doomsday scenarios in the midst of all this, which people who are naturally skeptical latch onto. So a lot of thoughts there.
Sorry, you said the reality is going to be against AI. Can you give me some examples of that? How AI is making...
Well, thus far it has not played out that way, but I think if there's...
What is... So what has happened thus far? It's been...
Well, there's been some belt tightening in tech. I would worry that there's going to be job loss and entry-level work is going to become much harder.
You are identifying the issue here. Okay. It's hard to fight a rhetorical battle when one side is continually living in hypotheticals.
And the other side is trying to live in reality. So this is a perfect example. AI to date has by and large made employees more productive and the job losses are theoretical.
Yeah. I'm not trying to like say there won't be job losses. I've like said like obviously it makes sense there's going to be job losses. And you know, I compared it to the analogy that I still sort of hold on to is the first wave of enterprise computing, which wiped out entire back rooms of people that used to write down numbers in books. Like that's how you did accounting or that's how you did ERP systems. And no one looks at that as a problem. That was arguably in like the very short time scale, the biggest widespread wipeout of jobs by tech. And no one views that as a problem because of course we use computers to keep track of ERP systems. And of course we use computers for accounting and keeping track of numbers. Do I want a bank that has a bunch of people in a back room writing down numbers? No, the computer should do that. And so this cuts both ways, which is, okay, we didn't... Number one, let's see what actually happens. Number two, historically speaking, like we actually absorbed and adjusted fairly well. And it's embedded in this that there's no change in demand. There's no new problem to be solved. It's just everyone just doesn't have a job. And like, again, there's just no historical precedent really for this unless you want to get super political and say, well, actually the vast majority of people today don't really work. Which is, I mean, we could have that discussion, you know, fake jobs, email jobs, whatever it might be.
Invented hamster wheels across the economy. This is doubled down by the fact that all this rhetoric is driven by people who need to touch grass.
They live in this entirely virtual world and the virtual world is eating itself and you talk about plumbers and electricians, they don't count because they're not in the virtual world. Oh, actually maybe the real world counts more than anything. They're great people. Live events matter more and more. Like actual things, social interactions. All those things, like you go back to why don't we build beautiful cathedrals anymore? It's like the people who are capable of doing that can get much better jobs doing something else that pays way more money. Like you have a shift in what's valuable and you just don't have the resources to build cathedrals because of the comparative advantage of actually should use this somewhere else to get much more of a return than doing this. Back in the Middle Ages, if you don't build a cathedral, what are you going to do? Plant more potatoes.
I think we've discovered the first talking point for all these AI companies. We're building cathedrals again in the future where all the boring white-collar work is automated.
No, but the problem is that... So number one, everyone's living in hypotheticals. Number two, the only hypotheticals that get traction, whether it's the human condition or whether it be the motivation of the people propagating them, are the negative ones. All the positive ones get waved away. Oh, and then number three, you have this issue, even if the AI companies were to come out with all the positive hypotheticals, they do this all the time. That's why they talk about like curing cancer and all this sort of stuff. Now, are they bad at it? Possibly, but also no one listens to them because they're like, oh, that's all self-interested. It turns out if you actually want to get traction, being negative and saying this is all terrible is the only one that people give credence to. So the question is, you want to criticize the people delivering the message, the problem are the recipients. The only thing that people... So demand problem, people want to hear doom.
I was going to say it also may be kind of a reflection of the tech platforms that we all consume news on these days.
Well, this is number three. Number three is what's different today as compared to past tech waves is the total capture and relentless negativity of tech media. I do think that's a real thing. I was talking to someone today, like where is the... where's the like... there's a whole category maybe because I've been in the Mac world and Apple world of like the Apple enthusiast site, right? Where they're just like there's a whole area of cool things you can do with the Mac. Those people are amazing. Let's...
The news they seize on as they cover Apple every single day never fails to delight me. Just they're endlessly reviewing new apps and look at this new thing you can do, right? And there's not... that doesn't seem to really exist in AI, right? Like who's out there like actually... if there's things on Twitter and stuff like that.
