317070

34 ms·

451.
by 317070 6y ago
> therefore it isn't really smart and can't actually 'understand' what it writes about. The crux is that you cannot determine a mechanism to distinguish between this "Chinese room" and someone who actually
452.
by 317070 6y ago
> The new result indicates hints of a deviation from one: the statistical significance of the result is 3.1 standard deviations That's not enough, is it? I thought in particle physics, people wanted 5 standard deviations? Does someo
453.
by 317070 6y ago
How to do numerical backpropagation (differentiation) on the result of an integration of a stochastic differential equation. And I am sure I only have scratched the surface of this niche already. Numerical algorithms can be both amazingly i
454.
by 317070 6y ago
My rule of thumb: when there are whole civilizations that exist on a diet, you will be okay. In this case the Jain are a prime example.
455.
by 317070 6y ago
Did anyone see the movie Brazil? This is exactly what that movie is about. The bureaucratic-technological engine being prone to tiny errors, which have cascading effects on the people affected. And despite putting in some effort and telling
456.
by 317070 6y ago
Just xor them together. As long as they are uncorrelated, you will get the best of both worlds.
457.
by 317070 6y ago
Eastern Europe is a word which is a bit contested. In the countries west of the wall, it came to mean all countries east of the wall. Perhaps with the countries around the wall being Central Europe. However, the wall did come down. In many
458.
by 317070 6y ago
> There is a considerable overlap between the intelligence of the smartest bears and the dumbest tourists. Park Ranger at Yosemite National Park (on designing secure garbage cans) https://www.schneier.com/blog/archiv
459.
by 317070 6y ago
> However, AI works best with a database — and those databases are not yet there. They do, but I would tell you that if that is your main motivation to create those databases, don't do it. The AI might spend milliseconds on work tha
460.
by 317070 6y ago
So, I think I explained myself badly. I fully agree on the problems of AI you point out. I agree with you on the reduction of freedom AI will bring. But then again, every relevant piece of technology took away such a piece of freedom. I thi
461.
by 317070 6y ago
> It's a dystopia where AI works because of us trying to conform to it, because otherwise we are out of luck But all technology is like that. Cars work, because we conformed our environment to it. Email works, because we conform to
462.
Why machine learning algorithms are hard to tune and how to fix it(engraved.ghost.io)
4 points by 317070 6y ago|0 comments
463.
by 317070 6y ago
It looks like he changed his mind. I don't know the guy, but I like him even better for it.
464.
by 317070 6y ago
Yes, but you cannot do that when you are short 149% of all stock. There simply is not enough stock available to cover your short position. Which is why GME was targeted in the first place.
465.
by 317070 6y ago
So, from what I understand (I did not do the maths myself), despite Gamestop being the most traded stock right now, not enough has moved in the premarket for Melvin Capital to actually close their position when they claimed they did. It see
466.
by 317070 6y ago
There is 1 company, that company has 1 stock. Alice is the shorter and expects the stock to drop. * How shorting works normally: Alice thinks the price will lower this week. On Monday, Alice borrows a stock from Bob. She sells that stock im
467.
by 317070 6y ago
They are related, but not the same. This recalibrating of logits is basically what you would do afterwards, to correct for the fact that you have been importance sampling your replay buffer.
468.
by 317070 6y ago
Exactly, that is the alternative system which was the default. It has gone through a number of cycles in history and is not always thought of the same way across different cultures (which is why there is a whole book on the topic). But it i
469.
by 317070 6y ago
About rationalization. That the world has not always been as bureaucratic as it is now. And then Jacques Ellul who says that this bureaucracy, which he calls technique, is a self-reinforcing entity on its own outside of the people that make
470.
by 317070 6y ago
I would absolutely love that and use it. Curtis is the reason I got into Max Weber and Jacques Ellul. I cannot remember which of his series led me into that rabbit hole though.
471.
by 317070 6y ago
Also, this digital debt tracker uses numbers, which most of these communal debt systems do not do. Because different people put different value to things, tracking debt in numbers is not how it is done (the book goes in a lot of detail how
472.
by 317070 6y ago
With barter is meant (in the context of Graeber's book) trading tit for tat, without having a currency to go between. E.g. a farmer buying a book and paying with apples. The book says that this barter method is almost never used, and i
473.
by 317070 6y ago
That is basically what the book is telling. You are probably not tit-for-tat bartering, but taking on debts and returning things you owe in some more vague approach. History went through multiple cycles where currency went in disuse in favo
474.
by 317070 6y ago
Something between a sociologist and an anthropologist, but making video art rather than books.
475.
by 317070 6y ago
I disagree. There is often no conspiracy or intent by the people in the individual stories he uses, he is putting them together to show in broad strokes things in society which were not there before or which he thinks are changing. You need
476.
Switch Transformers Scaling to Trillion Parameter Models(arxiv.org)
1 points by 317070 6y ago|0 comments
477.
by 317070 6y ago
> The bank guy said the bank will make no money on this loan. Ha, I believe that the guy told you that, I don't believe he is correct. :) The bank is not a non-profit and is totally making money on that mortgage. A considerable marg
478.
by 317070 6y ago
Well, math is two parts. There is the "prove this conjecture I have" part, which I think will be taken over by machines rather soon (say, within my lifetime). I cannot imagine there is something extra-ordinarily hard that mathemat
479.
by 317070 6y ago
The Schelling point works when there is no communication. But how do you know if your community members are not communicating?
480.
by 317070 6y ago
Check out Hanabi!

More ›