317070

51 ms·

541.
by 317070 6y ago
I have seen the figure 0.4 kWh for 100 pages of text come about in the paper for generating text with the model. So that goes up to $0.002/page of text when you take into account hardware cost too. Source: https://arxiv.org&
542.
by 317070 6y ago
Black holes do not have more gravity per mass than any other object. Put the other way, if the sun would suddenly change into a black hole without losing a lot of mass, the earth would still follow the exact same orbit.
543.
by 317070 6y ago
I'm curious though, what DO you talk about as general topics?
544.
by 317070 6y ago
In order to maximize impact. Press doesn't want to talk about this unless there is an embargo, otherwise they might be 'late' in which case they would rather not mention it at all (and not have to spend costly journalist time
545.
by 317070 6y ago
Yes, Brazil is practically halfway between Portugal and Angola. It's half the GDP per capita of Portugal, but twice that of Angola: https://www.google.com/publicdata/explore?ds=d5bncppjof8f9_&...
546.
by 317070 6y ago
But Cetaceans do have language? And we do have communication to some level. > Experiments have shown that they can learn human sign language and can use whistles for 2-way human–animal communication. Phoenix and Akeakamai, bottlenose dol
547.
by 317070 6y ago
The whole atari suite has been solved recently by a single algorithm: https://deepmind.com/blog/article/Agent57-Outperforming-the-...
548.
by 317070 6y ago
>and there are ways to obtain justice that are not available in most other countries (mass media, social media, courts). How many countries in the world do you imagine don't have courts, social media or mass media? Or even lack any
549.
by 317070 6y ago
I would say it is exactly the opposite. :) It is showing how you can get drastically better at deep meta-learning by being better at deep learning. But it does not really show how you can be better at deep meta-learning outside of the impro
550.
by 317070 6y ago
There have not really been a lot of major breakthroughs in meta-learning in the last year, as far as I am aware. The paper is basically saying there was not a lot of progress in the 3 years before that either. All in all, nobody really has
551.
by 317070 6y ago
The field of meta-learning is still very immature though. I can see why you would already want to start a scholarly discourse on the topic, but I am not sure how useful these techniques are for the students involved. They are still very ad
552.
by 317070 6y ago
The common example for the "would this even work as a parliament" is the EU parliament. It has 705 seats, and is filled with people coming from ±150 different parties: https://en.wikipedia.org/wiki/Table_of_po
553.
by 317070 6y ago
Germany has more than twice the GDP of Russia. They could reasonably handle themselves. US and China are considerably larger however.
554.
by 317070 6y ago
That is very cool. Though, part of the pedal are of course the knobs. You'd need to condition the wavenet on the knobs. Did that work well (I assume that you tried that already)? Also, what is the inference latency on your model? A nic
555.
by 317070 6y ago
Yes, for me it was also an epiphany that people weren't professional, they _acted_ professional. Suddenly a lot made sense! Careerwise, it can help to ask what a professional would do, as it is a solution that while it might not work,
556.
by 317070 6y ago
"Life is unfair." Simple, but for me it was an important insight. I used to think I deserved things, and now I know that while that might be true, you still need to be assertive to get them. There is no karma or divine interventio
557.
by 317070 6y ago
> for every λ there exists an M and for every M there exists a λ such that the resulting problems are equivalent Sorry, this might be a tad off-topic. I have over the last month being thinking extensively on the topic. As far as I can se
558.
by 317070 6y ago
But now zoom-sales is competing with zoom-sales, no-one is traveling. I doubt zoom-sales can compete with travel-sales.
559.
by 317070 6y ago
You are correct, I was unaware of the German case! In fact, it seems that many European countries do have a credit score system. Some do have other systems: https://en.wikipedia.org/wiki/Credit_score#Germany Here are s
560.
by 317070 6y ago
EDITED: the below is wrong (as stated by a child comment). There is a diversity of systems in Europe. OLD COMMENT: Here's the thing, as far as I know, Europe does not have those companies. So it definitely should be possible to do with
561.
by 317070 6y ago
It depends on more than just ZFC, also on the definitions of the real/complex numbers. The crux of the proof is that 0.99999... is being constructed within the real/complex numbers, and in that system it is equal to 1. And at the
562.
by 317070 6y ago
Why not have wheels? What is it that makes spot better at telepresence, telemedicine, remote vital inspection, internal delivery or disinfecting hospital rooms than, say, a roomba platform?
563.
by 317070 6y ago
That is one side of the scale. The other side of the scale has the meat industry in it (which is absolutely not like 100 years ago). As someone who used to work in a slaughter house, I am not convinced that today is better. I can see our ch
564.
by 317070 6y ago
Except that it wouldn't speed things up at all. Academia writes run-once code, which changes spec fourteen times in one week. Their use case is orthogonal to industry. Have you considered that maybe the academics actually know what the
565.
by 317070 7y ago
Possible actions: - If you had a JP Morgan account. Cancel it. - Evaluate if they did something illegal, organise a group and go to court over it. - If you live close to a JP Morgan office, go there and do something. Try to not attract poli
566.
by 317070 7y ago
> As a side note: in my experience, people who say things like “voting is not acting, you need to do more!” are not reliable voters, and not reliable activists either. Well, sorry if I take the above personally, but here is another datap
567.
by 317070 7y ago
I mean, if there ever was a point where you take a step back and ask yourself "will voting actually solve this" this surely must be it? How on earth would voting actually solve something like this? I'm not advocating for &quo
568.
by 317070 7y ago
However, the further it goes, the more clear the second one gets. I.e. "am I already at the point of no return, I guess not, it's unclear. Oh wait, now I am, let's trigger." So, Buridan’s Principle is correct for time-in
569.
by 317070 7y ago
Well, it will be still here in the same way Fortran is still here. I.e. while now it does not seem to bad, there will come a time where people will wonder why anyone put up with coding in C, when it was so prone to making bugs with things t
570.
by 317070 7y ago
Correct me if I'm wrong, but doesn't PAC on real data usually just tells you that "it doesn't know the answer"? Like, it says the equivalent to: "I think this is the answer, but it might be miles off and I give

More ›