Imnimo

38 ms·

481.
by Imnimo 4y ago
I go even further - I OWN a car I don't drive every day. I don't even split it with my neighbor.
482.
by Imnimo 4y ago
Totally agree. Data is a problem when you need millions of labels, but when your training corpus is just "all the text I can find on the internet" and the training objective is self-supervised, it's not what's stopping
483.
by Imnimo 4y ago
>There are only a handful of companies with enough data to be able to train artificial intelligence algorithms. https://commoncrawl.org/
484.
by Imnimo 4y ago
So is the idea here that the submitters figure they have a miniscule chance of sneaking through, but it costs them basically nothing to generate, and if by some miracle they win, they get a payout that's a significant amount of money i
485.
by Imnimo 4y ago
Do the kids participating in, say, the math Olympiad not receive an amount of tutoring and exposure to satisfy these criteria?
486.
by Imnimo 4y ago
I must not understand how Facebook works anymore. I didn't realize you can follow people, I thought you just sent mutual friend requests.
487.
by Imnimo 4y ago
Maybe I'm behind the times, but isn't the whole deal of Facebook that I've chosen to be friends with these people who I know in real life? Why do I need them to be verified?
488.
by Imnimo 4y ago
While it's possible that it could repeat (in a loose sense - the contents of the context window has shifted, so the exact calculations in the next forward pass will necessarily be different strictly speaking) the calculations that arri
489.
by Imnimo 4y ago
What is your proposed mechanism by which using previous tokens allows it to "build [a character] for itself"?
490.
by Imnimo 4y ago
>The character it has built for itself is extremely suspicious when you examine how it behaves closely. And I don't think Microsoft has created this character on purpose. The thing doesn't even have a persistent thought from on
491.
by Imnimo 4y ago
MUDs (specifically Achaea) got me from hunt-and-peck to competent typist. Not even Mario Teaches Typing could manage that.
492.
by Imnimo 4y ago
Is it easier to have a theory of mind when you don't have a mind of your own? Like the part that makes the ToM test hard is that you know what's in the bag, and you have to set that knowledge aside to understand what the other
493.
by Imnimo 4y ago
How did this get to the front page of HN? This is so transparently asinine. In what universe does finding a nearest neighbor constitute attribution?
494.
Ways to Spaceflight (1929) [pdf](ia800304.us.archive.org)
2 points by Imnimo 4y ago|0 comments
495.
by Imnimo 4y ago
I will certainly agree that if police training were completely different, then it might be helpful to have more of it.
496.
by Imnimo 4y ago
That would be great, but it's not the option we're presented with. Police training emphasizes escalation of force, it frames interactions with the public as a battle, and spends an inordinate amount of time on firearms and combat
497.
by Imnimo 4y ago
>the significant bias in OpenAI's treatment of the same adjectives when applied to different groups is certainly also a story. What is the story? If the bias shows up on sentences that no one will ever say, why is that interesting?
498.
by Imnimo 4y ago
>Why would OpenAI be flagging one kind of meaningless sentence more often than another? Well, let's consider a specific example. His first chart is women/men. Let's take five of his top-line adjectives (the ones used in
499.
by Imnimo 4y ago
Just eyeballing the list of adjectives used, the real story here is that OpenAI flags a lot of sentences that are essentially meaningless as hateful. I notice that he uses words like "evil" and "idiotic" in his diagram,
500.
by Imnimo 4y ago
Yes, clearly the officers were just confused and trying their best when they beat him to death. If only someone had told them not to do that.
501.
by Imnimo 4y ago
>Why would you think that? The purpose of police training is to indoctrinate recruits into police culture. https://twitter.com/KatieSponsler/status/1619341593827487744 > I don't think it's controve
502.
by Imnimo 4y ago
Probably for the best. I'm very skeptical that police training decreases misconduct. If anything, it probably increases it.
503.
by Imnimo 4y ago
Only 109 retrievable images out of the 350,000 most-duplicated is fewer than I expected. Maybe it's just the stringent definition of retrieval, but I would have expected many famous works of art like the Mona Lisa and Girl with a Pearl
504.
by Imnimo 4y ago
We should just operate on pixels of images of text: https://arxiv.org/abs/2212.08045
505.
by Imnimo 4y ago
Is that not his job description?
506.
by Imnimo 4y ago
One piece of Magic history that is still well-preserved is the ancient usenet groups: https://groups.google.com/g/rec.games.deckmaster https://groups.google.com/g/rec.games.trading-cards.magic.mi..
507.
by Imnimo 4y ago
What percentage of successful impersonations are really attributable to 0/O confusion?
508.
by Imnimo 4y ago
I dislike the focus on district shapes in gerrymandering discussions. The thing that matters is how severely the districting process subverts the will of the electorate, and that can be done with squares instead of squiggles.
509.
by Imnimo 4y ago
Wasn't the gun supposed to be loaded, just with blanks?
510.
by Imnimo 4y ago
At least he didn't talk about politics at work. Can't put a price tag on that!

More ›