skinner_

9 ms·

61.
by skinner_ 4y ago
The point is, I'm annoyed with a company not providing a service for me, am I gonna go the extra mile to make them successful?
62.
by skinner_ 4y ago
...which means that a Tralfamadorian would see an average human as a 6*10^17 meters long snake.
63.
by skinner_ 4y ago
https://aisafety.world/tiles/ lists dozens of institutions. Each hold different positions on AI development, but almost none of them hold the position that you ascribe to their "faction", maybe with the excep
64.
by skinner_ 4y ago
The correctness score is maximized by faithfully imitating humans. Humans do have goals.
65.
by skinner_ 4y ago
> If ChatGPT can write a credible essay about a given subject it means your subject was rather generic to begin with. ChatGPT created a pretty good essay for my daughter's home assignment. She had to write a fictive autobiography fr
66.
by skinner_ 4y ago
Yes, this is amazing. The reason I am a bit less surprised is that I've been playing with textual inversion in the last few days. Just from six photos of me, it can create portraits showing me from any direction. That means it has a ve
67.
by skinner_ 4y ago
If someone played around with textual inversion: what happens if I give it hundreds of images? Does it plateau, does it regress, and if it regresses, what is the issue exactly?
68.
by skinner_ 4y ago
Work is ongoing, this is a whole field of research now, a good keyword to search for is "implicit regularization".
69.
by skinner_ 4y ago
> The central intuition in using T5 is that extremely large language models, by virtue of their sheer size alone, may still learn useful representations despite the fact that they are not explicitly trained with any text/image task
70.
by skinner_ 4y ago
It's not the same internally. Imagen is diffusion model, Parti is autoregressive. Research at Google is quite decentralized, but IMHO it is wise of them to pursue both approaches. At this time it is very unclear what would the successo
71.
by skinner_ 5y ago
It would be awesome to have two interacting communities: AI experts building open source general game playing engines, and gaming fans writing pluggable rule specifications and UIs for popular games. A bit of googling shows that there is a
72.
by skinner_ 5y ago
The original post focuses on the limiting edge case where the only information coming out of the doctor is a single bit: Treatment vs Healthy. I believe the subject can reasonably expect some explanation from the doctor, and then we are in
73.
by skinner_ 5y ago
That makes much sense, thanks! But then why didn't they call it mix() or lerp() or interpolate()? I would have immediately spotted the logic behind it. EDIT: I think I get it, there also a mix() with a different type signature but same
74.
by skinner_ 5y ago
> GLSL supports the ternary ?: operator. WGSL does not support this, but provides the built-in function select(falseValue, trueValue, condition) which does much the same thing (although mind that parameter order!). I'd really like t
75.
by skinner_ 6y ago
Thanks! I have run this command now (not on an M1 but on a 2018 MacBook Pro), and I'm reporting that 31% of my file system accesses were coming from ESET Security (namely esets_daemon, esets_proxy, ERAAgent, esets_gui), and 19% from Ch
76.
by skinner_ 6y ago
I recommend Scratch and editor.p5js.org My kid is 7. He can't really code alone, but he can do nontrivial modifications to Scratch programs based on trial and error, he can create very simple but visually interesting p5js apps like thi
77.
by skinner_ 6y ago
> How cool! I was on the fence about whether or not to put up the source code— I'm _glad_ I did! I'm glad you did, thank you for that! Disclaimer: I'm not claiming that any of my modifications actually help, there's t
78.
by skinner_ 6y ago
This is so great! Frankly, I believe that this kind of low-parameter-count high complexity optimization task is the least suitable kind of task for SGD. Bad local optima everywhere. But I didn't let this opinion of mine spoil the fun:
79.
by skinner_ 6y ago
Google's ambitious BiT (Big Transfer) model is missing from the list: https://ai.googleblog.com/2020/05/open-sourcing-bit-explorin...
80.
by skinner_ 6y ago
Here's an experiment: take two zygotes from each of two animals, and do a DNA swap. That's 2x2 combinations of environment-DNA pairings. Compare the 4 offsprings. If this can be done, it has been done a million times already, I pr
81.
by skinner_ 6y ago
I'm a huge fan of Blokus ( https://en.wikipedia.org/wiki/Blokus ), a deep board game for 2 or 4 players. https://blokee.mattle.online/ and https://pentolla.com/ are places where you
82.
The Google and Verily coronavirus websites are off to a rocky start(vox.com)
1 points by skinner_ 7y ago|0 comments
83.
by skinner_ 7y ago
Absolutely possible, might even be a good idea, but my expectation is that the results won't be robust: the fakes will be uncovered by a slightly differently trained classifier. Maybe even the same classifier with a different random in
84.
by skinner_ 7y ago
One more data point: in Hungarian it's also common, also borrowed from English. It's spelled lúzer.
85.
by skinner_ 7y ago
My attempts below look nicer to me, the all caps probably leads to data sparsity issues. The global coherence of the first one is amazing, and the line "One of the best ways to test if you are a robot is to play against yourself."
86.
by skinner_ 7y ago
It is true that Gerard 't Hooft, the most famous proponent of superdeterminism is an asymptotic quantum supremacy skeptic, but I don't think superdeterminism implies no quantum supremacy. I'm a fan of superdeterminism, but I
87.
by skinner_ 7y ago
Proposition? Does this count as a proposition?: https://www.newscientist.com/article/2114900-concerns-as-fac...
88.
by skinner_ 7y ago
> That paper shows that imagenet classifiers retain their ranking on data drawn from the same distribution as the training set. That isn't even close to the same thing as generalising to real world data. Not the same distribution, i
89.
by skinner_ 7y ago
I'd like to signal boost this, it's a very important line of research. Congrats on a groundbreaking paper! The first time I have seen it, I was completely shocked by that perfect linear fit on "Do ImageNet Classifiers General
90.
by skinner_ 7y ago
Here is a guess: It was https://en.wikipedia.org/wiki/Pomace_brandy , törkölypálinka in Hungarian, created from the leftover fruit pulp after pressing the grapes for wine. That's what Italians call grappa. The oth

More ›