elcomet

7 ms·

31.
by elcomet 2y ago
You're mistaken. The perceptron was invented by Rosenblatt, a psychologist. This field has deep roots in neuroscience.
32.
by elcomet 2y ago
The ghost is insanely impressive, it's the example that gave me a "wow" effect. The cloth physic looks stunning, I never thought we would reach such a level of temporal coherence so fast.
33.
by elcomet 2y ago
It just means x/10
34.
by elcomet 2y ago
I understand it as " the speed increases by 97%"
35.
by elcomet 2y ago
IRC and distributed protocols un general had a big issue : you loose history every time you disconnect
36.
by elcomet 2y ago
What games are you playing? I'm guessing FPS don't work so well with latency
37.
by elcomet 2y ago
You can try it on hugging face
38.
by elcomet 2y ago
It's a similar approach to OpenAI's o1 model ( it's not cited, but there's no available paper for o1). I don't see any mention of weight release unfortunately.
39.
by elcomet 2y ago
yeah this is simplified a lot. What it truly means is that some of the energy emitted at the core takes thousands of years to get "out". The energy is converted multiple times, getting released as photons, then absorbed again, etc
40.
by elcomet 2y ago
Yes it is both, photons are emitted everywhere in the sun. The photons emitted on the surface will come directly to us, but those emitted in the core will take thousands of years to get out.
41.
by elcomet 2y ago
But this choice has very limited impact; as you are already in a very narrow window of color
42.
by elcomet 2y ago
Yes you can
43.
by elcomet 2y ago
You need to scan the QR code only the first time using the desktop app.
44.
by elcomet 2y ago
Even whatsapp does it now
45.
by elcomet 2y ago
They could publish it with public key encryption, so you can verify existing links without seeing all the links
46.
by elcomet 2y ago
Because you could just print the ticket, then sell it, and still enter the show with it ?
47.
Kyutai unveils today the first voice-enabled AI openly accessible to all [pdf](kyutai.org)
25 points by elcomet 2y ago|10 comments
48.
by elcomet 2y ago
We prefix our comments with "minor:", and all employees know that this means it's something that would be nice but not necessary to merge
49.
by elcomet 2y ago
I think you misread the parent. He didn't say Photoshop is bad for painting. He said Krita is bad for photo edition.
50.
by elcomet 2y ago
Or alternatively, people will need to move to colder regions.
51.
by elcomet 2y ago
I still use screen, is tmux better?
52.
by elcomet 2y ago
They rely heavily on what we call residual or skip connexions. This means each layer does something like x = x + f(x). This helps the training a lot, ensuring the gradient can flow nicely in the whole network. This is heavily used in ResNet
53.
Emulation Is the Sincerest Form of Flattery(fpgacomputing.blogspot.com)
3 points by elcomet 2y ago|0 comments
54.
by elcomet 2y ago
Why do you say it's a terrible idea? I'd say it's a pretty common idea today to ask chatGPT for help in complicated commands. Putting it in the shell directly is smart and helpful. Maybe the implementations has some flaws (it
55.
by elcomet 2y ago
That's not completely true. The architecture must behave well for scaling, which is not trivial. Basic multi-layer perceptrons do not scale well for example, the gradient will vanish or explode deeper in the network.
56.
by elcomet 2y ago
It's not designed to output the most likely séquence. It's designed to output the most likely next token. That's quite different.
57.
by elcomet 2y ago
I think you're not looking at this from the right perspective. A LLM is designed to sample text, that follows the training distribution. It is not designed to tell you the "most likely" text that follows, and we don't ac
58.
by elcomet 2y ago
But not necessarily high or low income.
59.
by elcomet 3y ago
But there are much less of those people. I'm not convinced by what you're saying
60.
by elcomet 3y ago
They have their own data centers, they don't use AWS

More ›