naillo

36 ms·

391.
by naillo 4y ago
This makes it seem so underwhelming and overhyped. This just sounds like any ordinary festival.
392.
by naillo 4y ago
I'm just gonna stop paying attention to any fear predictions the hive mind makes at this point. Jobs getting cut, jobs increasing. No idea what's actually happening but I'm not gonna let myself get emotionally involved with w
393.
by naillo 4y ago
Carmack is awesome. Rarely does an interview anymore but when he does he really delivers and gives us friggin 5 hours. Look forward to it.
394.
by naillo 4y ago
I meant for general spammers who goes after tons of sites mostly blind. I agree it would not help for a targeted attack.
395.
by naillo 4y ago
This is an old post but it reminds me of the idea that the blogosphere of the last decade really has moved to arxiv instead. A lot less informal but for technical nuggets of insight that's really the best place with the best culture ri
396.
by naillo 4y ago
I like this solution because spammers are unlikely to try to get around it. A delay eats into their time budget and they can't introduce a human-like waiting time on every site they try to spam, better to just move on to find cheaper t
397.
by naillo 4y ago
This is kind of a weird take to me given that photoshop exists. (Tons of proto-computer vision algorithms in there, like basic convolutional filters.) I suspect you'd still get copyright if you modify it a bit somehow.
398.
by naillo 4y ago
Slightly tangential but if you've ever looked at the data transferred on a youtube video in the network tab they're actually suprisingly tiny (as far as I've seen). Often only like 10 MB for multi minute downloads. Top notch
399.
by naillo 4y ago
I saw a slide somewhere recently that showed that although compute per individual CPU/GPU is reaching an upper bound, the compute per dollar is still growing exponentially. Gave me some hope that we'll still be able to do cool t
400.
by naillo 4y ago
I suspect part of why they closed open APIs down is because it was abused (for spam or similar). Not sure how to convince them to open it back up in light of that.
401.
by naillo 4y ago
It'll be interesting which of these win; scanning real life objects as you describe or just generating 3d assets directly through a DALL-E style model.
402.
by naillo 4y ago
Just waiting for someone to apply CLIP guidence onto this (so I don't have to).
403.
by naillo 4y ago
I agree that that's going way too far. But I also sympathize with people and see the side of people who feel like they've lost some trust in all this after the last few years. There's a spectrum of reactions.
404.
by naillo 4y ago
I wasn't actually gonna use them to solve a diff equation though to be honest :) I was just experimenting with denoising diffusion and instead of a u-net to infer the denoising step, I was thinking about using a multivariate polynomial
405.
by naillo 4y ago
More than one variable :) x/y/z etc
406.
by naillo 4y ago
Anyone here (or the OP) know how you'd use chebyshev in a multivariate scenario? The wikipedia article only defines them for the 1d case.
407.
by naillo 4y ago
It's not clear to me why you wouldn't just turn off your phone if you think you're being targeted by such an extreme attack.
408.
by naillo 4y ago
I wonder if they'll even make back what they spent on training the models before competitors of equal quality and lower cost eats up their margins.
409.
by naillo 4y ago
This news is funny since it doesn't actually change anything. It's still a waitlist that they're pushing out slowly (not an open beta). Nice way to stay in the news though.
410.
by naillo 4y ago
One interesting comment about this is that some models actually benefit from being fed their own output. Alphafold for instance was fed with its own 'high likelihood' outputs (as demis hassabis described in his lex friedman interv
411.
by naillo 4y ago
It's also funny that this likely won't 'unbias' any actual published images coming out of it. If 90% of the images in the world has a male CEO, then for whatever reason that's the image people will pick and choose f
412.
by naillo 4y ago
> And remember, the ML model is never on your side; its primary agenda is the agenda of whoever paid to have it built. AGI ads will be wild. (Hopefully it'll be so good it actually improves my productivity in society and realizes ju
413.
by naillo 4y ago
PCA on high order features of a large language model
414.
by naillo 4y ago
Please be more empathetic to people who beg for money.
415.
by naillo 4y ago
If you film in public places (depending on the country) it's often legal. Still really rude though and should be socially discouraged.
416.
by naillo 4y ago
It's funny how these videos say they're intended to spread love and encourage random acts of kindness, but the actual effect of them is decreasing trust in strangers (that they might be filming us for their tiktok). This makes me
417.
by naillo 4y ago
For all the cynicism in this thread, I can't really deep down dislike sama for investing his money like this. There's plenty of ways to get more money and this isn't one of the most secure ones so money going to physics and s
418.
by naillo 4y ago
I've the same experience roughly. The one recent happy finding I found though is that using the 'ipfs.io' link on library genesis is generally much faster (and as reliable) than the other mirror links. Kinda gave me hope that
419.
by naillo 4y ago
It's funny how many cool things could be created with just a few hours of manual labelling labor.
420.
by naillo 4y ago
Well you will have to do some work to understand them (read papers etc). But distilled (from larger models) versions of these things are fairly capable of being computed even on the web etc. There's definitely cool low hanging fruit he

More ›