habitue

66 ms·

781.
by habitue 9y ago
A3C is only ever run on one game at a time[0]. This paper gets good performance on all games with the same agent [0] read as: I have only seen papers with 1 agent per game for A3C
782.
by habitue 9y ago
I guess I thought this might actually tell the story of how Apple negotiated not to have the Intel inside sticker. Instead it just states that the sticker wasn't there and goes "we can only guess"
783.
by habitue 9y ago
This strikes me as an interesting case if the credit assignment problem from reinforcement learning[0] (Think of the reward in this case being negative, you get sick) [0] http://www.scholarpedia.org/article/Reinforcemen
784.
by habitue 9y ago
TIL!
785.
by habitue 9y ago
Probably just me but NNabla seems unfortunately close to NAMBLA.
786.
by habitue 9y ago
I think a lot of people are looking at this as "maybe this article is offering advice" instead, the way I'm looking at it is this is giving us a rough quantification of what we all know intuitively: that college is so expensi
787.
by habitue 9y ago
It's probably not time yet, but the winds are blowing that way so hard, I wouldn't be surprised if it becomes completely obvious in a couple years
788.
by habitue 9y ago
Flowtype is seeming more and more like the Mercurial to Typescript's git. (That is, in terms of inertia, it seems like the community has decided on a winner)
789.
by habitue 9y ago
What you're doing is confusing a second-order signal when you have first-order knowledge. We should expect someone who doesn't understand net-neutrality at all to use second-order signals like "Big company X supports net neut
790.
by habitue 9y ago
There is a counterbalancing force, which is hard to see going away, which is that larger companies have a harder time seeing which way the wind is blowing and getting ahead of the curve. They make decisions slowly, and tend to be wary about
791.
by habitue 9y ago
I think the implied subtext of this site is "emacs is sexy, but its logo isn't. Here, download this alternate logo so you can feel ok about using it on a mac"
792.
by habitue 10y ago
Recently we had a meeting about the next phase of the project now that Horizon and RethinkDB are owned by the linux foundation: https://www.youtube.com/watch?v=EG0aucGzJE8&t=279s
793.
by habitue 10y ago
Usually companies have a clause where they have the first right to buy the stock options back before you offer them for sale to anyone else. Employee stock options are sort of intended to be NODROP, and if they want to make shares available
794.
by habitue 10y ago
Luckily, deep learning is built on super complicated mathematical foundations, but the practice isn't. You could quickly get up to speed with a library like Keras and be productive, building things and trying them out without needing y
795.
by habitue 10y ago
So what we're seeing is these fields pre-deep learning were mathematical disciplines making steady progress on well-understood foundations. Then deep learning came in, was exceptionally effective at problems that had been difficult to
796.
by habitue 10y ago
Yeah, this is totally true, but it is also an uninteresting thing to say because basically only Ebert and potentially the author of this piece disagree with you.
797.
by habitue 10y ago
There are a few things I'd point out to summarize the article: 1. The prevailing wisdom now (and at the time the article was written) was that Ebert was dead wrong about games being art. This was a settled issue and not controversial.
798.
by habitue 10y ago
> spanning more than 200 surviving individuals > collective net worth of about $10 billion You are taking his individual contribution as a percentage of the combined net worth of 200 individuals.
799.
by habitue 10y ago
Yeah, exactly. We call it AI because we need a name for this kind of adaptive algorithm, even if it isn't anything close to strong AI
800.
by habitue 10y ago
Benefits of apng vs. video?
801.
by habitue 10y ago
I think I'm naturally a high self-monitor kind of person who has slowly learned to loosen the reins over the years. It feels better when you can be authentic, but you need to be around people you trust and you need to know what the lim
802.
by habitue 10y ago
> Perhaps this article can serve as a warning to those who don’t love programming what they should expect in order to stay relevant in their career. And if what it takes is a turn-off, maybe they should pursue a different line of work. T
803.
by habitue 10y ago
I guess this means we were all using 100% of our brains after all...
804.
by habitue 10y ago
> has no options beyond a) try to seem sincerely humbled and intent on improving, or b) do literally anything else You know, a year ago I would have nodded along to your reply. But after seeing Trump choose B every time and repeatedly
805.
by habitue 10y ago
Just an FYI, local fox affiliates don't have the ideological bent of the cable station. They're often run by entirely different companies
806.
by habitue 10y ago
> The problem with this line of argumentation is atheism does not explain or give reasons for the most basic questions we have as human beings Unfortunately, that has no bearing on whether it is true or not. It would be nice if there wer
807.
by habitue 10y ago
Pathnet is super interesting, but the authors of the paper don't make claims like this blog post does about being general AI, and this blog post doesn't back up those claims well either.
808.
by habitue 10y ago
> making you not consider other techniques that might work better Currently, DL is the most powerful technique for many problem types. I think for a beginner, learning that is a safe bet, the "next thing" is likely to be an ela
809.
by habitue 10y ago
> I came to the conclusion that we need to create a single app that does 80% of things This is called a web browser
810.
by habitue 10y ago
It depends on what conclusions you're trying to draw from that information. What conference a paper was accepted to is a second-order signal of the noteworthiness. It's probably easier for someone versed in the field to just read

More ›