gs17

11 ms·

241.
by gs17 10mo ago
Sounds like a good opportunity for some kind of batching feature.
242.
by gs17 10mo ago
> I can safely assume that this dictionary of bad words contains no people’s names in it. This is a big one for this kind of project, and I've never been sure how usernames for people named Kike should be handled.
243.
by gs17 10mo ago
I could see social-media-ish websites not wanting those names to prevent impersonation. They'd be deciding if they want to risk friction when a big name joins the platform (@cocacola needs Coca-Cola to verify) or risk threats from that
244.
by gs17 10mo ago
It's odd that they focused so much on "it's better than regexes" when it doesn't handle these cases where a regex would do well.
245.
by gs17 10mo ago
I'm not understanding your categories. Every dictionary word is flagged? It seems any first or last name is a "public_figure" ("apple" is a "public_figure" and also a "brand", I guess that means
246.
by gs17 10mo ago
I think there's some value in providing a huge dictionary of things to test against, with tagging for what things are to help filter. This doesn't do a great job at it, and it would make 100x more sense as a library, but it's
247.
by gs17 10mo ago
GeForce NOW is supposedly decent for a lot of games (depending on connection and distance to server), although if Nvidia totally left gaming they'd probably drop the service too.
248.
by gs17 10mo ago
> ESPECIALLY if you're in seattle Especially especially if you're at Pike Place. It's not even the real original Starbucks location!
249.
by gs17 10mo ago
It's not quite as bad as I expected (still bad), since it does at least clearly tell you how to not buy it, in normal size type even. Except then they decided to make it out of place alphabetically (and it seems to have moved at least
250.
by gs17 10mo ago
It's fiction, there's breadcrumbs at the top that list it as in the "Fiction" category. qntm is good at plausible sci-fi, e.g. https://qntm.org/mmacevedo
251.
by gs17 10mo ago
> 1.89gb executable file Add me to the list of people curious about this. It feels more like some sort of bug than a real attack, it would be odd to use such a huge file for every torrent.
252.
by gs17 10mo ago
> someone will have to come in to make the actual product. My experience has been more that they expect you to fix the broken mess, not rebuild it properly.
253.
by gs17 10mo ago
> However he disagreed that the ‘don’t insure me’ option was hidden, and said that 98% of Ryanair’s passengers could “find a way to decline insurance”. I'm not surprised, but still a bit impressed by the ability to lie like this. So
254.
by gs17 10mo ago
The future is Google People: https://qntm.org/perso
255.
by gs17 10mo ago
Or if they were simply being harassed because their ex was a cop who decided to use the cameras to stalk them, where there's not even an excuse.
256.
by gs17 10mo ago
They should also require a warrant at least, especially for any data sharing. With "they can only be accessed by law enforcement", we've already had plenty of police harassing their exes. If they couldn't convince a judg
257.
by gs17 10mo ago
> Show them the problem on a shared screen I'm pretty sure the intent is to weed people out well before they get to a point where you could share a screen with them. He mentioned a few people "resubmitted the application",
258.
by gs17 10mo ago
Is it? The page implies it's allowed, but they want people who think running it is "more of a hassle".
259.
by gs17 10mo ago
> I went in with doubt I had done it right. And if in your doubt you decided to run it through the interpreter to get the "real" answer, whoops, you're rejected.
260.
by gs17 10mo ago
It depends why they didn't get it "correct" (asked ChatGPT bad, used Python REPL not so bad, used screen reader very not bad) and what "correct" even means for this problem. There's a bizarro version of this gu
261.
by gs17 10mo ago
> Let some % of candidates who get this wrong through Really, the better test would be to not discriminate on it before you know it's useful, but store their answer to compare later.
262.
by gs17 10mo ago
> into an AI solver Not just an "AI solver", a Python interpreter will also do it "wrong". The idea is that it's so simple that anyone qualified should be able to do it in their head, so they get the answer witho
263.
by gs17 10mo ago
The trick is that the = has CSS styling with "opacity: 0; font-size: 1px;".
264.
by gs17 10mo ago
"Original" as in the original of the one they used in their tweet.
265.
by gs17 10mo ago
The Hebrew version doesn't seem to have that error, I wonder what they used to translate it, it feels like a rare hallucination for recent LLMs.
266.
by gs17 10mo ago
> however im not sure if these are true uv maps I can tell you with 100% certainty they are not. For example, Crash doesn't have a backside for his torso. You could definitely make a model that uses these as textures, but you'd
267.
by gs17 10mo ago
> Still some scientific inaccuracies, but ~70% correct That's still dangerously bad for the use-case they're proposing. We don't need better looking but completely wrong infographics.
268.
by gs17 10mo ago
This discussion isn't really about red light or speed cameras, although they suck in different ways. They are technically "license plate cameras", but they only capture for a specific purpose. ALPR cameras are about a surveil
269.
by gs17 10mo ago
> and try to advocate for and enact anti-car, pro-public-transit policies If you're lucky. Sometimes you just get anti-car. I'd love to not need a car at all, but where I am now it would mean Ubering instead because they'v
270.
by gs17 10mo ago
The gaussian splat output can be generated with CPU (this was honestly one of the easiest AI repos to get running).

More ›