I don't think everyone will agree with me, and we might have truly AI generated movie and pop stars, but I think most people won't want synthetic art- it's missing the point of what art is for, for expressing something about humans.

Like programming? Or what else, I'm not sure I understand what that would mean.

And there will be a lot of ego death, esp. with people deep in a lot of kowledge work.

But if you claim that AI is absolutely a waste and bad for humanity I feel sorry for you if you can't see how incredible it is that we have this new technology. -That it will most likely solve some very core issues humans are working on, and that it is a kind of evolutionary moment in human history, creating artificial intelligence. I feel sorry that you can't see past your own nose to see the bigger picture in that sense.

I tend to think big changes will always make things better and worse at the same time. If you look at the long horizon of history (100s of years) it appears that it makes things better on balance. But who really knows for sure what will happen.

There are two different groups I think. One is not having kids because of environmental issues, but I think there's a larger group not having kids because of systemic societal issues that make it hard to have kids. (sometimes it's the same group but not always)

So part of what I was pointing out is that when people say "we need less people" they aren't implying "wow, I'm glad the population is shrinking" they're really saying, "less people need to be born in the future"- which probably isn't true.

> and the population is an obvious lever to scale that impact

Also, it's not obvious, since, with the exception of China, no one actually meant to do this. It might actually lead to a lot of human misery, the same as climate change might do, so it's not clear that having less people in this specific way is a net good- for example countries like Japan completely collapsing in on themselves and going extinct. I think what you meant by smaller population was just have the same exact society but thinned out, but that's not what this next chapter of degrowth looks like. The boomers who insist on their full pensions at 60 are pretty much killing the youngest generation and leaving them with nothing, and to a certain extent everyone is being rational actors- there's no reasonable way out of the problem.

This might also be one of those things that aligns with people's experience of the world, and the de-population thing is inherently counterintuitive unless you are, by definition, outside the norm- that is, you would only ever notice this if you live where people already don't live. (rural Japan, etc.)

It reminded me of this: https://malus.sh/ "cleanroom as a service" - which I guess was kind of a joke, but looks pretty serious now.... (discussion back then: https://news.ycombinator.com/item?id=47350424)

But this seems to hinge on the "clean room"- is this a separate set of case law or is it related? As in, if they had been more judicious about the way the game was implemented could they have claimed it was not under copyright?

I wonder how much of this has to do with possibly not-that-well-trained runners and or gait inefficiencies.

All elite marathons are now won with big cushy shoes. But those runners might be forced to wear them so it’s hard to come to any conclusions from that.

But at the very least if it was way less efficient to have highly cushioned shoes I think we would probably see that at the top level.

The study is also suspect because overall running efficiency changes drastically between trained and untrained runners and it doesn’t seem like they accounted for that.

Has anyone seen the neo-cloud profit margin on deepseek? I wonder if the deepseek served API prices account for the training cost? Because they don't / have not raised the money to fund their future operation / training- and they depend on that cashflow for now? That would suggest a really high profit margin on inference only.

You've somehow summed up the entire debate around AI here, but seemed to have missed the point as well.

Alignment is pretty hard!

There's been endless debates here about moderation vs. censorship- on the spectrum of Hackernews to 4chan. Calling the Hackernews moderation rules "censorship" would also not be reasonable- nor do people want to visit 4chan because it's "free speech".

I think it's actually a really nice workaround that human language and english is flexible enough that we can still say exactly what we mean in a different way and everyone knows what we are talking about.

I fundamentally don't think the platforms want to get rid of the engagement these topics bring, (the last round of criticism being that they want divisive topics on the platform) and if and when they do it won't look like changing "kill" to "unalive"- the algorithm will simply be directed to silence the topic. Many such examples in China. We should be angry about the censorship we don't see and can't detect as easily instead.

It comes off a little bit like they made a lazy decision at the outset, didn't want to reverse course to save face and knew they didn't need to back it up with anything....

In the 1900s this was communism. A lot of people (either disingenuously or not) interpreted this as anti-democratic. But if you look at what actually happened where conditions were right, it turned out to work well- China and Vietnam, and if you just look at authoritarianism, not just communist authoritarianism, then Singapore, S. Korea, Taiwan as well. Some would also argue that the USA falls into this category- that the same arm of the government suppressing corruption is also going after marginalized groups.

There aren't that many examples in history where the ideal balance is actually struck.

For instance a Chinese EUV machine seems like it's very far away. Even if they have (steal/borrow) the necessary IP.

I agree in principle, but it could be more than 5 years, maybe 10. Who knows what things will look like then.

For frontier labs the current compute seems to be driving model progress (in training) at least to some degree, even without true RSI, and this seems like it'll continue to keep any chinese model from drawing even with the frontier labs, at least for the foreseeable future.

Inevitably the chinese government will drive more funding in chip fab technology and the money will come around to build chinese data centers, but who knows how far off that is. A few different things in the tech tree need to fall into place. It doesn't seem like it'll be next year.