krackers

20 ms·

451.
by krackers 9mo ago
I guess they are at a difficulty where it's not too hard (unlike millennium prize problems), is fairly tightly scoped (unlike open ended research), and has some gravitas (so it's not some obscure theorem that's only unproven
452.
by krackers 9mo ago
Even with flux? Shifting the white point is far more comfortable than dark mode for me. In fact it also matches with the ambient lighting, 4000-6000k during the daytime then 2000-2700k after sunset.
453.
by krackers 9mo ago
You could equally say "we're evolved to hunt during the daytime, where you scan the environment when the surroundings are bright."
454.
by krackers 9mo ago
Although revisiting this, you have to be a bit careful about the argument. Basically what you're doing is breaking down p(correct) = p(correct & agree) + p(correct & disagree) where former is 0.8*0.8 and latter is 0.8*0.2. Expl
455.
by krackers 9mo ago
I found https://daedtech.com/defining-the-corporate-hierarchy/ has a good comparison >The Peter Principle holds that people are promoted until they prove incompetent in their role, and then they remain there; compet
456.
by krackers 9mo ago
>GPT$$$ (it’s $1B per month) This doesn't follow, open models are not that far behind frontier. And it doesn't take into account geopolitics.
457.
by krackers 9mo ago
In japanese there is ヤッホ (Yahho~) which might be related to english "Yoo-hoo". Apparently this comes from dutch "joehoe". I've also seen etymology sources list "yoo-hoo" coming from sailing jargon "yo
458.
by krackers 9mo ago
All abstractions are leaky abstractions. E.g. C is a leaky abstraction because what you type isn't actually what gets emitted (try the same code in two different compilers and one might vectorize your loop while the other doesn't)
459.
by krackers 9mo ago
The chewed up artwork on the floor makes a better art piece than generic AI art. In fact it too counts as a form of "AI psychosis" so it even fits with the original theme.
460.
by krackers 9mo ago
no in fact that's the proof that adding Bob doesn't help. If Alice & Bob disagree, then since both are correct with same probability it doesn't matter if you pick alice or bob. So WLOG you choose to trust Alice. But now w
461.
by krackers 9mo ago
I'm interested in this too. I was expecting just a chromium reskin, but it does seem to be at least something more than that. https://news.ycombinator.com/item?id=46625189 claims it uses Taffy for CSS layout but the do
462.
by krackers 9mo ago
There is this famous video from 2013 https://www.youtube.com/watch?v=yTeBUpR17Rw
463.
by krackers 9mo ago
Is there any other way to do it though? Even if they implemented some form of auth logic, since it's all client side anyone could spoof it. The only real way to distinguish Claude Code from non-Claude Code is behavioral analysis (e.g.
464.
by krackers 9mo ago
Is this not similar to DeepSeek lighting indexer
465.
by krackers 9mo ago
The killer use case for AI will be bonzi buddy reborn.
466.
by krackers 9mo ago
>Importantly, cognitive impairments are not limited to those with ongoing systemic symptoms; they are also evident in COVID-19 recovered individuals (COVID-RHC) who no longer report active symptoms, compared to healthy controls without S
467.
by krackers 9mo ago
A good test might be to provide it only about a third of the tests, then when it says it's done, run it on the holdout 2/3 of tests and see how well it did. Of course it may have already seen the other tests during training, but t
468.
by krackers 9mo ago
>A JavaScript interpreter written in Python? I'm assuming this refers to the python port of Bellard's MQJS [1]? It's impressive and very useful, but leaving out the "based on mqjs" part is misleading. [1] https:
469.
by krackers 9mo ago
>Why on Earth is this a requirement? Actually I just checked and it's not, technically you can create key equivalents without modifiers as well [1]. For Finder this doesn't work though, because enter seems to be specifically ha
470.
by krackers 9mo ago
You can remap it to any shortcut you want (as long as it has a modifier key in it)
471.
by krackers 9mo ago
The site has a history of trolling people, it used to be the cursor background. Also do check his privacy policy https://tonsky.me/personal-information/
472.
by krackers 9mo ago
The mean minimizes L2 norm, so I guess there's some connection there if you derive OLS by treating X, Y as random variables and trying to estimate Y conditioned on X in a linear form. "L[Y|X] = E[Y] + a*(X - E[X])" If the dat
473.
by krackers 9mo ago
Apparently a new paper from DS shows this is not the case, or rather the information isn't captured with as much fidelity as you'd expect. Intuitively the residual stream apparently doesn't have enough dimension to allow each
474.
by krackers 9mo ago
VLC uses libass for .ass subtitle rendering as far as I can see, shouldn't it be the same?
475.
by krackers 9mo ago
This matches with my observation, VLC tends to be more tolerant of slightly broken files or random issues that you encounter when streaming. Especially for hls streams, vlc often works when ffplay refuses to play it, I believe because vlc u
476.
by krackers 9mo ago
You should do a show HN post, it's definitely interesting. One of the few good cases where blockchain is useful. And maybe find some statistician to define a proper metric for statistical significance here, since my gut feeling is that
477.
by krackers 9mo ago
Maybe slightly related, canon layers provide direct horizontal information flow along residual steams. See this paper, which precisely claims that LLMs struggle with horizontal information flow as "looking back a token" is fairly
478.
by krackers 9mo ago
easyist annoyances filter should take care of a lot of these.
479.
by krackers 9mo ago
Additionally generics can differ in terms of the inactive ingredients or coating or other stuff. See e.g. https://news.osu.edu/all-generic-drugs-are-not-equal-study-f...
480.
by krackers 9mo ago
You should add dilated conv and conv_transpose to the list.

More ›