sp332

35 ms·

421.
by sp332 3y ago
But as noted in an article published 16 minutes earlier by the same person on the same blog, Fortinet did say something. https://grahamcluley.com/tooth-be-told-toothbrush-ddos-attac...
422.
by sp332 3y ago
We know that the power is wasted because the results are just thrown away. Only potentially winning results are even presented to the network.
423.
by sp332 3y ago
I know it's unhelpful, but it's not new. "Core" inflation leaves those out because they change too rapidly to be useful as trend indicators.
424.
by sp332 3y ago
The Stockfish installer is ~45 MB. At 16 bits per parameter, the 270B model would be over 500 MB. The 9B model would be smaller than Stockfish, but you could probably find a smaller chess engine that achieves 2000 ELO.
425.
by sp332 3y ago
The CLI for llama.cpp is very clunky IMO. I put some kind of UI on it when I want to get something done.
426.
by sp332 3y ago
Well, sometimes. Sometimes not. https://arxiv.org/abs/2310.17567
427.
by sp332 3y ago
Llama models were trained on Books3.
428.
by sp332 3y ago
It's cheating because it has memorized the answer to the puzzle instead of using logic to solve it.
429.
by sp332 3y ago
This seems like the opposite of how Game Pass works.
430.
The video game funding gap: How investors are failing marginalized developers(polygon.com)
2 points by sp332 3y ago|2 comments
431.
by sp332 3y ago
To be more pointed than logicprog, Raichik just lies a lot. Example from the article: Libs of TikTok’s post said it happened “in front of children.” Anthony said no children were present.
432.
by sp332 3y ago
Well, we haven't stopped.
433.
by sp332 3y ago
Sounds cool, do you have any examples? Why is it xtraordinary? What's the prompt format?
434.
by sp332 3y ago
https://www.rnbo.gov.ua/en/Diialnist/6714.html "Total mobilisation may follow the 2024 presidential elections."
435.
by sp332 3y ago
Yeah, it's been big news for a couple of years now? https://www.cbc.ca/news/world/russia-military-conscription-a...
436.
by sp332 3y ago
Fair, I am just posting the other side.
437.
by sp332 3y ago
The point is stated around 8:00 in to the video. But if that sounds interesting, I think it is worth watching the earlier part of the video with examples.
438.
by sp332 3y ago
Not only that, but so the decoder knows how much memory to allocate.
439.
by sp332 3y ago
https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2...
440.
by sp332 3y ago
Won't the fungi eat the structure of the building?
441.
by sp332 3y ago
A collection of links from the Twitter thread: https://serl-robot.github.io/ "SERL, our open source software framework that enables learning manipulation skills in 15-30 minutes from raw images" https:/&#x2F
442.
by sp332 3y ago
Physically maybe, but knowing that someone is getting off to images of you as a young kid is still traumatising, especially if you still are a young kid when you find out.
443.
by sp332 3y ago
Cough drops usually help me enough to get to sleep. Get the good ones like Fisherman's Friend or numbing ones like benzocaine lozenges. And there are cough suppressant drugs. They are central nervous system depressors though, so don&#x
444.
by sp332 3y ago
I don't suppose there's any way to actually get this?
445.
by sp332 3y ago
OpenAI has a huge amount of people working on training data. https://time.com/6247678/openai-chatgpt-kenya-workers/ https://www.nbcnews.com/tech/innovation/openai-chatgpt-ai-jo...
446.
by sp332 3y ago
That one was from UK BioBank data and published in April. This one is from John Hopkins in the USA and was published in July.
447.
by sp332 3y ago
Interesting that another Lancet paper on this topic was recently retracted. https://news.ycombinator.com/item?id=38894636
448.
by sp332 3y ago
Seconded. The way to minimize prep time is to set aside a couple hours and make at least one week's worth of meals. The planning and shopping lets you control what kind of food you are eating instead of depending on pre-made pre-proces
449.
by sp332 3y ago
They were planning to get a couple billion dollars from investors. Now that that deal is off, they don't seem to have any plan.
450.
by sp332 3y ago
The main benefit of a GPU in that case is much faster prompt reading. Could be useful for Code Llama cases where you want the model to read a lot of code and then write a line or part of a line.

More ›