pfortuny
pfortuny on Hacker News
That is the joke, indeed.
The Jacobian Conjecture settles the "solve a math problem" and is controversy-free.
The basic idea of EA is the same many people have (especially those not used to the idea of
nonlinear orders): quantifying everything, even "goodness" or "utility".
Not everything is quantifiable, and even more, not any family of things (investments, in this case) admits a linear ordering (they may even not admit a total order).
There can be incomparable outcomes from actions, and they may be maximal in their suborder.
This is difficult until you stop measuring everything and dedicate a minute to think of ordered trees.
Right now, "The Lord", by Romano Guardini, a large collection of homilies covering the whole life of Christ. Obviously, this is a Catholic text.
Also, "Heaven on Earh", by Emma Wells, about nine or ten Gothic Cathedrals (mostly the history of their building and preservation). A bit too plain for my taste but includes interesting facts and anecdotes.
Finished recently:
-Flowers for Algernon, which I liked a lot.
-Piranesi, by Susanna Clarke, so beautiful and surprising.
-The Great Gatsby, which I had not read yet and which I have loved reading.
Are you from a main Spanish city? Because the chargers are being set up but not so fast, and parking is not exactly an easy thing to find for most people.
The father-mother-child-donkey conundrum. Its only solution is: to hell with anyone's thoughts, we'll act as we prefer.
Because egress firewalls have existed since way before "ai" and are very easy to set up.
And memorizing. Rote learning makes the brain.
[un]believable. Thank you.
Sorry what? Honest question.
That is exactly what "training AI" needs: labeled data.
It might be that the velocity at the spiraling axis is not infinity (i.e. there is no blow-up in finite time). Viscosity is a bitch.
That has another name but I forget. Essentially contour plotting in high res.
Yep: the problem with IP is exactly the "Property". The texts are economically (even if only for leisure) valuable, so the producer deserves a fair comoensation unless he has explicitly dismissed it.
So paying a price for an intelectual product is reasonable. Nothing to do (per se) with property.
Wow, this is really interesting. Thanks.
Apart from the well-known dubious position of OpenAI wrt truth, the prompts/inputs do mot include the outputs.
You can train on a sequence of outputs. In the end, OpenAI outputs are OpenAI's property.
You can learn a lot from a single side of a conversation.
Courts exist, among other things, to specify the rules in ambiguous cases. Rules will be ambiguous always.
Colima is in Mexico.
Kolyma is in Russia.
As long as you can get an exogenous reaction from it, which is not guaranteed at all.
The self-replication assumes also that there is enough energy stored in each planet (or coming from a Sun) to do the work... That is pretty much unlikely.
Unbelievable: I read the title and the only thought I got was that households get 3h free energy in Australia.
Also "neuron" should be at least italized.
Well, that is what markets do: behave. Rationally or not.
The 1929 panic was also irrational but it happened. There is no way to solve that.
Panics happen and bubbles too, and in the meantime, very long term investments tend to grow (in average), as long as the economy works as expected (improving in average, long-term).
Of course, catastrophes are not impossible (the Black Death, the French Revolution, World Wars...).
That summurizes it pretty well. The Commission is essentially omnipotent, and there is no relevat judiciary to stop it.
This is farcical.
Yes, from a PURELY MONETARY viewpoint, Brexit was a disaster. But there are more important things than money.
I'm from Spain btw.
The first thought when moving people must be "where/how will they shit?" Only once this is solved can you ask "what will they eat?".
This is not even dumb. Unbelievable coming from act....
I guess they do not understand the relevance of the median in this. They just take the average and think it is "good enough"...
I would call it "Engineered (as machine-made trustworthy) Mathematics": you have results but nobody undrstands them but they were not produced by humans.
Despite that, people use them. In that sense, similar to the Finite Elements method. But the tools (statements) are machined like any other tool (screwdriver).
Similar to microprocessors in your example. But about statements.
Well, by "understanding" I mean "understanding by humans", indeed.