casualscience

12 ms·

61.
by casualscience 3y ago
I don't really know what kind of breakthrough they could achieve. The only other step function improvements I could imagine right now are: 1. A great technique for memory banking: e.g. A model which can have arbitrarily large context w
62.
by casualscience 3y ago
Yeah, I find this kind of dialectic extremely annoying. It's basically just a trivial answer. Yes, everyone wants agency, so in some way you can tie everything we do to 'power'. The actual question above was trying to ask wha
63.
by casualscience 3y ago
Try to find a bathroom in a tourist town in france and then get back to me. In america if you go into most shops and ask, they will let you into the bathroom. In touristy areas sometimes they are dicks about it, but still you can find a way
64.
by casualscience 3y ago
There is no US law which is written statistically. I disagree with your '90%' characterization, but let's run with that. Would you be happy if our murder laws were 90%? Most of the time you can be prosecuted, but 10% of the t
65.
by casualscience 3y ago
I can't tell if you're being serious. Arbitration often means that you can not even discuss the case. If you screw over 1000000 customers, none of them, nor anyone in the public, will ever know. A lot of the reason coorperations d
66.
by casualscience 3y ago
I hate that this has been normalized, we can't seem to pass legislation to get rid of mandatory arbitration. The courts uphold this crap because it's supposed to be agreed upon by your own free will, but withholding service unless
67.
by casualscience 3y ago
Yes, and graphics cards are optimized to do matrix multiplications because most of what they do is projective geometry.
68.
by casualscience 3y ago
I work with LLMs, won't say where, but smaller models stop performing better on benchmarks after a certain point, e.g. they seem to hit their learning capacity (at least with current techniques). Small models struggle to keep context t
69.
by casualscience 3y ago
Most of this is very legit, but this > Still other outsiders, who misunderstand the entire process, are upset that intermediate results are sometimes incorrect. This confuses them, and they’re angry that the process sometimes assigns “p
70.
by casualscience 3y ago
> So maybe the study answers it but that is a lot to read through for one answer, but my big question that you did not quote is still. I don't know whether there is an official answer, but if you look at your link there, it says, e.