magimas

5 ms·

1.
by magimas 4d ago
You even notice that with the recent opus and fable models by Anthropic. If you give them a wide open problem statement, they'll start talking a lot of semi intelligible gibberish. My guess is that this happens because that's not
2.
by magimas 20d ago
I think in general these smaller applications of AI are seriously overlooked even though they will actually add lots of economic value for companies and lots of quality-of-life features for users.
3.
by magimas 20d ago
At least my QUERTZ layout also works as expected in the simulator. You use WASD to move the duck, Q and E to kick left and right etc. I think it's just the splash screen they didn't properly localize.
4.
by magimas 1mo ago
openworm: https://github.com/openworm/OpenWorm
5.
by magimas 1mo ago
oh please... lesswrong was full of idiots already in 2010 when people outside of that community were laughing about "self-taught expert" Yudkowsky's bullshit physics takes. They are not "further along" the AI discus
6.
by magimas 1mo ago
what are you talking about bro, I think you're really getting carried away here. Nobody cares about the direction of writing on the spine of a book. And your thing with sentence vs title case doesn't make sense because German does
7.
by magimas 2mo ago
but this seems quite overkill no? I currently run pi agent in Lima on a Mac with only the code project folder mounted and an extension that prevents pi agent from reading the contents of .env files directly. Yeah, there probably are some fr
8.
by magimas 4mo ago
I don't know, I don't think this "effort for effort's sake" is a very convincing argument. In particular, I think it's very much affected by recency bias in a way? What we perceive as "effort worth taking&
9.
by magimas 6mo ago
I think this could be an interesting read for you, I read it last week and it kind of argues the same points: https://shakoist.substack.com/p/against-time-series-foundati...
10.
by magimas 6mo ago
we did some internal tests. The quality isn't bad, it works quite well. But it's essentially on the same level of an ARIMA model trained on the data just much bigger and slower. So in my opinion it currently falls into a kind of v
11.
by magimas 8mo ago
this completely misses how crazy word2vec is. The model doesn't get told anything about word meanings and relationships and yet the training results in incredibly meaningful representations that capture many properties of these words.
12.
by magimas 10mo ago
mh, maybe it's cheating because it's still a STEM degree but I have a PhD in physics without any real computer science courses (obviously had computational physics courses etc. though) and I managed to 100% solve quite a few years
13.
by magimas 2y ago
yeah that's actually what I was thinking about. I have a PhD in physics, so I easily notice when ChatGPT just keeps agreeing with me even though we're on very shaky ground. But I worry about the times it does this when we're
14.
by magimas 2y ago
one of the more important ones in my opinion would be a prompt to make ChatGPT much less agreeable. Unless explicitly asked for, it never really challenges your observations and will just keep being supportive and telling you how well of an
15.
by magimas 3y ago
To my knowledge the original authors promised they would send out samples in the future but want to wait till they have a finalized paper approved for publication. Whether that's reasonable or not is very hard to tell. IF they are corr
16.
by magimas 3y ago
replication happens over time. For example, when I did my PhD I wanted to grow TaS2 monolayers on a graphene layer on an Iridium crystal. So I took published growth recipees of related materials, adapted them to our setup and then finetuned
17.
by magimas 3y ago
horrible take. Taking the LK99 situation as an example: simply copying and adapting a well described growth recipee to your own setup and lab conditions may take weeks. And how would you address situations where measurement setups only exis