ThouYS

6 ms·

91.
by ThouYS 2y ago
absolutely loving SublimeMerge!
92.
You Can Now Use DeepSeek R1 in Perplexity(techwiser.com)
3 points by ThouYS 2y ago|0 comments
93.
by ThouYS 2y ago
polycam or scaniverse (gaussian splats though)
94.
by ThouYS 2y ago
tried the 7b, it switched to chinese mid-response
95.
by ThouYS 2y ago
I wish Prusa weren't asleep at the wheel, then we would have bought a core one (that is, the hypothetical variant with large build volume and same quality as bambulab). Instead, we bought a P1S, which is, technically speaking, a fantas
96.
by ThouYS 2y ago
impressive tenacity
97.
by ThouYS 2y ago
marimo is really cool, albeit "pure python" is only true insofar as the diff is concerned. other than that, it's an unconnected group of functions that need the marimo runtime to stitch together. would be cool if marimo could
98.
by ThouYS 2y ago
thanks man
99.
by ThouYS 2y ago
3D printers, Bambulab is incredible!
100.
by ThouYS 2y ago
hm that would be very interesting, hadn't perplexity already solved all my AI needs a while ago
101.
by ThouYS 2y ago
That video inspired me to read the NEAT paper. I think that was the first scientific paper I ever printed out and read
102.
by ThouYS 2y ago
lol
103.
by ThouYS 2y ago
same q here. what can I do with this "world model" that I can't do with a game like minecraft or counter strike? asked the same thing a while back, and the answers boiled down to "somehow helps RL agents train". but
104.
by ThouYS 2y ago
copy the works of masters. that means look at it, think about it, put it away. And then, on your own, try to recreate it
105.
OpenAI Email Archives (From Musk vs. Altman)(lesswrong.com)
6 points by ThouYS 2y ago|0 comments
106.
by ThouYS 2y ago
hillarious
107.
by ThouYS 2y ago
If I had to bet, I would bet against it. Boston Dynamics for example, for the longest time, didn't use anything other than Model-Predictive Control. Only recently have they started using RL
108.
by ThouYS 2y ago
super good question, especially with all the tilting involved, which would make visual servoing difficult. Maybe some form of beacons on the ground?
109.
by ThouYS 2y ago
Oh, I see. I was somehow under the impression that the simulation was the game the RL agent learns to play (which kinda seemed nonsensical).
110.
by ThouYS 2y ago
I don't really understand the intuition on why this helps RL. The original game has a lot more detail, why can't it be used directly?
111.
by ThouYS 2y ago
thanks mate, exactly what I've been looking for
112.
by ThouYS 2y ago
In all these calculations you have to wonder what "using water" even means. Add to that that most water cooled systems are closed loops, i.e. no water escapes
113.
by ThouYS 2y ago
why get a remarkable, and get screwed by the company, when you can get a much cheaper boox go 10.3? it runs android and you can do what you like with it
114.
Wang Chuanfu: A Name Everyone in the West Should Know(interconnect.substack.com)
2 points by ThouYS 2y ago|1 comments
115.
by ThouYS 2y ago
one of the many books with a great premise, but ultimately no actual story
116.
by ThouYS 2y ago
nice! just like the strong interaction material from the three body problem!
117.
by ThouYS 2y ago
nice, more of this please. the constant abuse through everything digital has to be fought
118.
by ThouYS 2y ago
It's really cool, but the flame profile is that of a blowtorch, not a rocket engine
119.
by ThouYS 2y ago
sanitizers are a constant source of pain
120.
Noodles: A Path Tracer Made of Blender Nodes(tizianzeltner.com)
2 points by ThouYS 2y ago|0 comments

More ›