naillo

32 ms·

361.
by naillo 4y ago
I feel like they're just dying to figure out a way to get rid of extensions on desktop as well. Soon enough they'll figure out a way to rationalize them away.
362.
by naillo 4y ago
Also funny that their first sentences in the second article is > "You may be wondering if such a bold headline is true. The answer is yes." Feel pretty save to tag 'The Algorithmic Bridge' as not something to pay atte
363.
by naillo 4y ago
This is such a silly argument. People have been drawing and photoshopping non real depictions of lion elephant hybrids or whatever else for ages. It's already "clogged".
364.
by naillo 4y ago
You should check out taichi: https://github.com/taichi-dev/taichi They have a ton of great demos for doing physics but check out this example in particular for something related to your project: https://gith
365.
by naillo 4y ago
Good survey on many other extensions too: http://mmacklin.com/2017-EG-CourseNotes.pdf
366.
by naillo 4y ago
Position based dynamics :) With normal `next_pos = current_pos + velocity * dt` schemes you can get instabilities and it can be non obvious how to implement certain physics phenomena. With PBD you basically define a 'constraint' t
367.
by naillo 4y ago
For a more modern approach (that's one of those rare cases where things have gotten more powerful while getting simpler over the years): https://www.youtube.com/c/TenMinutePhysics/videos (by basically the inv
368.
by naillo 4y ago
They can definitely be that bad quite frequently. I've actually been a lot happier with stable diffusion outputs lately (doesn't hurt that they're free too).
369.
by naillo 4y ago
Well it comes with an NSFW filter that's activated by default so that's a non-issue (but importantly you can turn it off if you're ok with accidentally seeing stuff like that without explicitly asking for it).
370.
by naillo 4y ago
On the other hand I doubt many more people noticed that it went away than ppl on HN who saw the previous thread. So now they've all been re-informed and know they can rely on it again.
371.
by naillo 4y ago
This should be possible if someone just exported them to tflite or onnxruntime etc (quantization could help a ton too). Not sure why ppl haven't yet. Sure it'll come in the next few days (I might do it).
372.
by naillo 4y ago
Img2img and inpainting is usually a lot easier to get cool results (with a lot of control on your end) with in my view.
373.
by naillo 4y ago
If it is made illegal it'll probably be applied to the sites that distribute stuff like that, not this model itself.
374.
by naillo 4y ago
I much prefer a company that asks me to not abuse it but lets me, then a company that treats me like child and force filters it out for me.
375.
by naillo 4y ago
Fucking awesome
376.
by naillo 4y ago
We really gotta come up with another thing to repeat as the go-to diss than this. I'm all for putting pressure on them to release weights etc but they are a lot more open than e.g. google or meta, and the whole "ClosedAI" is
377.
by naillo 4y ago
This person has an awesome set of blog posts. One of the few rss feeds I keep track of.
378.
by naillo 4y ago
I'm just building ray tracers from scratch mostly and having raw BRDF data would make that easier.
379.
by naillo 4y ago
Just color? https://physicallybased.info/ I would have hoped for BRDF info as well.
380.
by naillo 4y ago
Makes me think of this https://float.exposed/
381.
by naillo 4y ago
Karpathy is awesome and I love that he's doing this and going down this route. (Hope he doesn't get that "offer he can't refuse" like carmack predicted and keeps doing educational stuff for us all :).)
382.
by naillo 4y ago
It's weird that it still feels like this is almost something I'm 'supposed' to know and at some point work my way through. (Maybe because it was on the brink of still being relevant when I got interested in programming i
383.
by naillo 4y ago
This is kind of the most important link in any case: https://github.com/CompVis/stable-diffusion
384.
by naillo 4y ago
The fact that they're releasing the code and weights is so insanely cool. Love enriching the public goods.
385.
by naillo 4y ago
That's a good point. In general there's probably a ton of easy ways to adverserially rewrite the binary against a deanonymizer like this. You could probably even make a program that rewrites your own code into someone elses style
386.
by naillo 4y ago
This uses 600 'candidate' programmers. But I wonder how much harder it becomes when on e.g. an arbitrary piece of github gist code. As the number of candidates increases (with many writing the same styles) I'd imagine the pro
387.
by naillo 4y ago
I'm much more scared of people who don't share Theils worldview.
388.
by naillo 4y ago
There are actually a huge chunk of the population that just isn't very online, and genuinely don't know these terms. (Even many young people.) It was a real eye opener to meet some 'terminally offline' people and realize
389.
by naillo 4y ago
I suspect it could be done for sort of periodic noise (or possibly with a predictive net). I've always wanted to implement this on the web but never got around to it.
390.
by naillo 4y ago
What's the story of how gizmodo is still around? I thought Peter Thiel killed them off.

More ›