algobaker

The 290 yen ramen at this place doing serious work for the city

Went to like 8 cities on my trip there and Fukuoka was my favorite spot

There should just be like 6 meme formats on arxiv, and when you upload your paper, the title is pasted on one of those as the thumbnail

Bayesian Optimization is just slightly-smarter random search

This is truly one of the spiciest paper's in causality

I took both pills

I'm pretty shocked that he's seen One Piece. Like...what. Feels like totally wrong generation, and he'd have missed it during it's peak when he was knee-deep in paypal

My favorite antichrist:

You should try nvim

His idea ain't it but your rule of thumb sould also lead to... never using AI tools for anything...

New bedtime reading:

Gets even weirder when it's leaked that the demo was tele op'd

Maybe some latent factor: being so obsessed about your field that you're willing to move anywhere to do it, means you will also probably be pretty good at it

Will be watching for this

This is an insane finding. Excited to see what therapeutics come out of being able to do early drug screenings for psychosis using skin cells. Seems like something Recursion could also potentially explore at scale?

I don't think "without profit incentive" is accurate, else there wouldn't be so many artists complaining of fears about being put out of the job

Best I can do is a Star Trek comparison

Finally some redemption for this guy

Any technological or clinically relevant announcement that drove it?

Sent straight to my Kindle

I too would guess it's true that many currently on H1B wouldn't be able to get O1. But I think a very specific subpopulation of them would probably have a high success rate: specific field (for sure AI-related but maybe there are others), with high citation counts and from well-known PhD programs.

This a fantastic goal. You'll definitely be kept busy. Good luck!

I need to find more accounts on Bluesky in AI or Bio to follow. My 'Following' feed here is totally dead

Do you think people might start switching to different visa options, dampening the effect? For example, at least in machine learning, I don't think many PhD grads from top schools would have much difficulty getting an O1 visa, which I guess they could then use to do a postdoc in the US?

Drop the full team surely

Was also shocked to hear Kurzgesagt has 70(?!) employees + freelancers

I do think the way we use the internet and discover truth online is already irreversibly changed

Everything they say about the future is mostly an extrapolation of the same failure modes models have now, which isn't necessarily going to happen, but also totally plausible to me.

This perspective on AI slop was pretty interesting. Their description of using current AI models to do creative tasks is roughly in line with my own experience.