Alpha3031

4 ms·

31.
by Alpha3031 1mo ago
I'm in the "post both and let your readers figure out which, if any, they prefer" camp but even if you didn't (and unless you were using one of those super special models that can't be released according to the comp
32.
by Alpha3031 1mo ago
The author giving the slightest amount of shit seems like a big assumption to make, but ultimately even it that were the case I don't see why I would prefer what I read to be curated in that specific manner (having access only to what
33.
by Alpha3031 1mo ago
People having preferences different from you is not a hive mind.
34.
by Alpha3031 1mo ago
Less unlikely, but I would still suggest that it is somewhat unlikely for any recent LLM to place any significant probably on, e.g. "disappointment" instead of "failure" as the appropriate token (or series thereof) after
35.
by Alpha3031 1mo ago
If it's not a problem, then there should be no issue with not concealing the fact, no? Lying about things one considers inconsequential is a useful signal about one's willingness to lie with little benefit.
36.
by Alpha3031 1mo ago
Seems unlikely unless Anthropic invented a time machine, given that the phenomenon predated Claude 1 by two years, and their stated introduction of watermarking (August 2) by 5.
37.
by Alpha3031 1mo ago
In general, yes, you'd probably start at a library's reference desk. (which are still a thing!) Some serial publications have indices where you can find key words (and there are also similar indices published by subject area). If
38.
by Alpha3031 1mo ago
Supposedly people are getting ~40 tps decode at Q8 on 2× DGX Spark (higher for Q4) which is what I assume they're suggesting is just under $10K USD. Prefill is just above 1.5k so TTFT is maybe 2 to 4 minutes? I don't have two DGX
39.
by Alpha3031 2mo ago
Depends on what you want to do. Some task specific models can be trained with a few ten or hundred thousand training examples so you can use a bigger model to produce synthetic training examples and then fine tune a smaller student model. I
40.
by Alpha3031 2mo ago
You said it yourself, model size and hardware. Big models cost more (good optimisation reduces things slightly, but they still need the hardware).
41.
by Alpha3031 2mo ago
If someone says just the word, it's hard to say if they mean just artificial selection, which can and is done to some extent today via genetic counselling by medical geneticists, coercive selection at a society-wide level advocated for
42.
by Alpha3031 2mo ago
The 13,800 number seem like it would be quoted from their annoucement of the $830 million funding round they had in March. Here's CNBC's article on it: https://www.cnbc.com/2026/03/30/mistral-ai-pari
43.
by Alpha3031 2mo ago
Yes for NAND, and I suppose nobody is using mechanical hard drives for this.
44.
by Alpha3031 2mo ago
Didn't the US stop new aid since 2025? Sanctions against Russia sure (at least until they decided the spat with Iran was more important) but I feel like that moral high ground is little enough that proposing to " ahem remove&quot
45.
by Alpha3031 2mo ago
Besides which, I doubt China has sufficient influence over Iran to restrain them, considering the US " ahem removed" the all the people that could plausibly considered moderates. I wonder how reasonable China would feel if the US
46.
by Alpha3031 2mo ago
Per the OSD definition of source code, "the source code must be the preferred form in which a programmer would modify the program." which means an argument could be made (as charcircuit is making) that the weights, being the prefe
47.
by Alpha3031 2mo ago
Jiang et al. (2021) TransGAN: https://dl.acm.org/doi/10.5555/3540261.3541391 In general probably not much of a stretch to get rid of convolutions or recurrence by replacing with attention and see if it works hence
48.
by Alpha3031 2mo ago
Aren't RLVR signals typically based off formal systems not natural phenomena? The formal sciences are certainly useful for producing tools used in natural science, but it's not entirely clear they alone are sufficient to associate
49.
by Alpha3031 2mo ago
Microsoft surely has enough resources to keep their fork going if they choose to, so I see no reason why Google would try that given they would likely end up with less control over the web.
50.
by Alpha3031 2mo ago
There are a lot of applications like that though, only they're all being rewritten for electron.
51.
by Alpha3031 2mo ago
It's more typical to use parsecs in practice for astronomy.
52.
by Alpha3031 2mo ago
Disney is able to jackboot anyone violating their IP because the DMCA imposes liability on big tech companies if they do not comply with a valid request. The comment you initially replied to proposes imposing such liability for fraud also,
53.
by Alpha3031 2mo ago
I want the mob boss to face legal consequences. If they choose to get out of the business of facilitating crime because of that then so be it.
54.
by Alpha3031 2mo ago
Normally, if there is sufficient evidence, both the mob boss and the ground level gangster are liable and criminally responsible for their crimes. You are free to disagree, but I don't see any reason why big tech companies should be co
55.
by Alpha3031 2mo ago
That's very interesting. Does that mean you can reduce say, a 30B class Q8 from ~30 GB down to 10 GB or less?
56.
by Alpha3031 2mo ago
People's right to make stupid decisions end when they reach public roads and/or borrow children to make a point. That's why we have things like driver's licences.
57.
by Alpha3031 2mo ago
As I understand it, it is the position of several DPAs that denial of access entirely (i.e. "consent or pay") could contravene the "freely given" requirement of GDPR in most cases, though this has thus far not been teste
58.
by Alpha3031 2mo ago
Lots of species are r-selected, I don't see why that would be considered inefficient. In fact I think there are probably more r-strategists than K-strategists.
59.
by Alpha3031 2mo ago
A lawsuit is not the same as a criminal prosecution. One is supposed to remedy damages, the other is supposed to be because it is in the public interest that people follow laws and we don't devolve into anarchy (or more dangerously, so
60.
by Alpha3031 2mo ago
Basically, they shot someone in public to promote their cool new gun.

More ›