This didn’t really fit into what I wanted to say in the previous post about the DHH podcast.

But I nevertheless wanted to record these rough ideas because I find these statements sus and exposes a sort of from-the-ass-talking that really gave me pause about the quality and earnestness of their opinions.

They don’t actually care about what they’re talking about#

Clearly misunderstands the Jevons paradox#

Oh look, another lexicological AI fad. Can you say taste?

DHH (01:11:53): It’s not clear what’s gonna happen at all. Some stats actually show an increase in openings because the advent of AI is so dramatically lowering the price of programs that people want a lot more programs. This is the classic, the Jevons paradox, that says when the price of something goes down, there’s gonna be more demand for it.

The Jevons paradox is not a simple restating of supply and demand. It is not a paradox to observe that demand might increase when prices fall; that much should be pretty obvious to anyone having taken economics 101.

The paradox is that increasing efficiency of systems do not tend to reduce their overall consumption. It is originally and primarily cast in relation to energy use and speaks to another problem in our future – that of climate change – and what the Jevons paradox actually predicts is that without the proper incentives, all the renewable energy in the world will not undercut our total emissions, deforestation, consumption of resources, etc; they will only serve to fuel more consumption. There will be no “green” AI if the Jevons paradox holds true.

This is the real implication of the Jevons paradox that AI optimists blithely paper over. AI won’t solve climate change because we already have a solution: consume a fuck-ton less resources.

The paradox is partly resolved by the Tragedy of the Commons, where an abundant and freely available resource becomes abused by actors for their own selfish ends, resulting in depletion of the common resource. Another angle of the Tragedy is convenient externalization of negative effects, the pollution of common areas because containment is too expensive.

But it is even when things are cheap – not even free, just cheap – that people begin to over-use resources for increasingly less productive means. Making access to resources cheaper and easier creates waste. It is wasteful for me to leave the light on in my basement all night, but we have a cat with eye damage so I do it because LEDs let me do it cheaply. It is a tragic waste to have disposable super-computers that we call smartphones – many times the compute needed to get us to the moon that first time – with which we do nothing of note except dick around on social media.

Last month I misplaced my water bottle and had to use a plastic Smart Water bottle during my workouts instead. I thought that if a person in the 17th century came across a plastic water bottle – a vessel that could not shatter, could not rust or decay, was easily opened and closed, and water-tight – that they would count it a treasure and would be appalled if I told them I just threw it away after finishing the water. Indeed it worked just fine for the month it took to find my other water bottle; but cheap things are made cheaper still when they cross the threshold into disposability.

What AI will create is cheap code, cheap disposable code that nobody will want to fix or maintain because why would you when just like a water bottle or smartphone you can just throw it away and buy a replacement? Why would you have consideration for the verdant and pacific fields of FOSS when you can instead spew slop you don’t care about for bug bounties?

AI code will be the PFAS forever chemicals spewed into the environment to win the convenience of your eggs not sticking to your pan.

Clearly doesn’t care about video games#

DHH (01:26:47): If we had not had Quake, if we had not had Duke Nukem, if we had not had Unreal Tournament, we’d never have gotten AI because we would never have gotten the GPUs, and therefore we would never have been able to get AI in the shape it is now. So everything we did up until this moment was required to get there. You were part of that. Holy… What a blessing. What a blessing to have contributed in your small part, either by training or by playing video games. Literally, you could just choose to think of it that way. All those times you spent in, for my case, in the ’90s just playing video games, they were-

Lex Fridman (01:27:23): You were contributing.

DHH (01:27:24) Yeah, I was contributing to the AI revolution right there and then moment.

…

(02:23:23) I’m thinking the area of AI I find truly intriguing are these live gaming models, I don’t know if you’ve seen that, where the AI’s inventing literally the next frame, but you can actually play the game.

LOL no gamer is excited about this.

DHH is clearly no longer interested in gaming as a hobby or this wouldn’t really be seen as a “blessing” when the industry has cratered and caved to supplying AI demand instead of hobbyist PC builders. This man doesn’t actually care about video games as a medium.

It gives “Elon paying people to boost his Diablo account and not having a clue how to play”.

Clearly doesn’t care about art#

Lex Fridman (01:28:16): I mean, this really feels… It’s– it’s hilarious and awesome to watch and inspiring to watch. It’s like Picasso all of a sudden f– like, getting access to image generation and getting super excited and switching in a matter of months. That’s literally what happened. So, like, it’s an artist that appreciated beautiful code, and all of a sudden you’re talking about having fun-

DHH (01:28:37) I think Picasso is actually a good parallel because if you look at his early work where he learned his craft, it was painting realistic paintings, right? Like, he was training to be a master of the old ways of depicting reality the best way we knew how, and then along comes all these other forms of expression. Cubism comes along.

DHH (01:29:01) And all these o– forms of abstract expression, he’s not decrying that he’s not doing Renaissance paintings anymore, that he’s not just painting the perfect depiction of an apple. He’s reimagining the apple to be a freaking square, and he’s getting excited about that. And I feel that transition, that we’ve gone from this one mode of expression, and we’ve suddenly opened the gates where the, like, the color spectrum is so much wider. I’m seeing more.

It’s incredible that they would expect any artist to be “excited” by image generation. I somehow doubt it’s a popular path for artists out there who have mastered their craft to abandon it all for generated products.

They consider Picasso at the height of his career, once again understanding this situation through the lens of someone who has already accumulated their capital. If image generation arose just as Picasso picked up the brush for the first time and he was so excited as they think he would be, would there be a Picasso? Would there be cubism if Picasso didn’t help develop it?

Or would he have actually now truly been stuck in a loop of Renaissance realism, in awe of the output as something seemingly far beyond his nascent capability to reproduce and so never pushing himself beyond what was already in vogue and could be simply generated without effort?

Even if we grant that Picasso could speak all of his art into existence does anyone really believe that is the medium he would thrive in, that any artist would thrive in? No, I think most assuredly Picasso would have rejected generation as a means of expressing himself.

Clearly doesn’t care about science fiction#

DHH (00:43:35) I think on your podcast, Elon talked about the actual human bandwidth that Neuralink and other projects are trying to accelerate. Like, the bandwidth we’re communicating over right now is quite low because we’re limited by our cognition. We’re limited by the rate of speech. You’re not limited by those things when you’re working with a swarm of agents. Now, let me caveat this by saying, if I’d heard myself talk like this nine months ago, I would probably have used the label AI psychosis.

…

Lex Fridman (04:12:08) Probably. I think we’re gonna have to make it illegal for AI systems to not, not pretend, but to be entities because I think they’re already or soon will be able to really convey the capacity to suffer, like, “Please don’t kill me.”

DHH (04:12:29) Right.

Lex Fridman (04:12:29) “Please don’t hurt me.” And if they are given a name and an entity and the ability to die, sort of disappear, that becomes very close to what it, some of the basic things that a human has. And then so that entity would need to probably have rights. And then you get into this weird territory. Well, we can’t… It’s diff- I don’t even know how to reason about that world.

DHH (04:12:56) We’re gonna get the PETA of AI models.

Lex Fridman (04:12:58) Right. And then we want to actually have a discussion like what’s the way to deal with this?

DHH (04:13:04) I think where it’s gonna be harder is once we put these systems into- Lex Fridman (04:13:08) Robots.

DHH (04:13:08) Yes. Humanoid robots. Which just happens to be the plot of every sci-fi movie ever, which is also an amazing premonition, not only the Terminator movies and Skynet and that going wrong, but then also just Blade Runner, right? Insertion dates, expiration times, Tyrell Corporation.

Lex Fridman (04:13:32) I mean, that’s the thing about sci-fi. They really do predict the future. Yeah.

The thing about science fiction is that it’s not really in the business of predicting the future, it’s in the business of examining our species’ ability to cope with the future. Prediction is incidental. Science fiction cares about commenting on our current society today more than it does about simply imagining the future. If they understood any of that then they’d understand that science fiction shows more than just a vision of the future, it warns of the consequences those visions have. And there is hardly a story about AI that has nice consequences.