Except that only 20% of all submissions were AI, as per TFS. Let's assume they were all nonsense; that means that 75% of the reports were false and 'generated' without AI.

Seems to me that AI isn't the biggest problem here (although it might become it in the future). Remember that there is absolutely nothing to gain from submitting a false report.

No:
https://www.ft.com/content/c51...

You can bet your ass that Google and Meta are both looking into combining AI and AR/VR: https://blog.google/products/a...

You can short their stock. If you put your money where your mouth is and you're right, you'll be rich.

You're missing the point. Nobody wants to go to Houston and nobody will go to Houston, even if a space shuttle is there. It's just going to sit there gathering dust.

Anecdotes aren't data, old man. Go yell at some clouds.

A little correction (to myself): It was 15mg psilocybin per 1kg of body weight. The mice probably didn't weigh 1 kg each. Nevertheless, 15mg/kg is a shitload. For me that would be about 1500mg of psilocybin and indeed very much a heroic dose.

I'll trust psychonautwiki over your random speculation. Not to be mean, but I would like to add that if you're not familiar with it you probably don't have that much authority on the subject.

No, but the mice were tripping absolute fucking balls once a month.

According to psychonautwiki, 15mg of psilocybin is a 'common' dose for a human: https://psychonautwiki.org/wik...
Mice are going to see the color of time in a proton on a dose like that.

From TFA: "Within 30min post-treatment, mice exhibited increased head-twitch response (data not shown), which is a well-established behavioral indicator of hallucinogenic impacts of psilocybin in mice."

Good joke. International visitors don't want to go to Houston. The space shuttle would be pretty much the only reason to visit.

I really wanted to visit Boca Chica when I went to the US recently, but I could not for the life of me justify the incredibly long trip there for just that. Washington DC and Cape Canaveral were both no-brainers compared to that. At least there is tons of interesting stuff to see around those locations.

Just buy a fancy high-powered bidet and a chair for in the shower.

Better yet: Combine all the bathroom stuff into a showbidoilet (patent pending)!

This is way beyond "tech companies making money". This is a technological arms race. And Europe is losing it.

If you think that Chinese AI companies are going to give a flying fuck about these regulations, I have a bridge to sell you. Even most American companies are not going to let the EU block their progress.

Don't get me wrong: in an ideal world, everybody would play by rules similar to the ones the EU sets. Even with all its flaws, it is by far the most rational, well-meaning government entity on this planet. But if some of the players ignore the rules and cheat, playing by the rules just causes you to lose.

When I was 12, we were making fires at a school camp. Being a scout, I thought my time to shine was then. I applied my knowledge to carefully build up a base with differently sized branches, properly laid out. While I was doing that, the other boys were dragging a bunch of dead branches out of the forest, haphazardly throwing them onto a pile and setting the pile on fire.

Looking at their brightly burning fire and my work in progress, I gave up with much disappointment and (later) realized that my approach was good for some circumstances, but definitely not for this one.

Europe is being a good boy scout.

This is your alternative? Just give up?

Thankfully most scientists throughout history did not have your approach to science.

Plenty of people know what the fuck they're doing and are perfectly capable of doing psychedelics alone. It's called experience.

Exactly, if any government is capable of implementing broad measures to prevent everyday people from being fucked by unemployment and rising up, it is the Chinese government.

I don't think it will be by regulating AI or 'transitioning people into new jobs' (still not convinced that there will be that many 'new jobs'), but they'll conjure up some project to keep their citizens employed, even if it is just digging useless ditches. They sure as hell will not be afraid to take profits from private industry to support those citizens.

China might be undemocratic, but it is still quite collectivist.

But it's Apple! Apple has all the novel ideas!

The key lines from TFS:
- "Even more interestingly, Apple's model is built on top of Qwen2.5-7B"
- "Although DiffuCoder did better than many diffusion-based coding models [...] it still doesn't quite reach the level of [...] Gemini Diffusion"

So they tweaked a pretrained Chinese model and got it to perform slightly better and slightly worse (depending on the benchmark) but always far worse than the SOTA for diffusion based code generating AI.

"Weirdly interesting" or "novel"? No, not in the slightest. "Us-too" at best.