eru

24 ms·

691.
by eru 3mo ago
Sure. Assume the one that's attached to your payment is public.
692.
by eru 3mo ago
It does make a difference: you only mark enough to be kept to make your phone looks like it's in use, as opposed to obviously wiped. By default, new material will be deleted. So you don't have to prepare again and again. > you
693.
by eru 3mo ago
> The LLM does not prove anything (it cannot reason). What is that supposed to mean? > It generates Lean code, and conveniently, in Lean, the code is also the proof. It’s not merely a model of the stated system, it is the system. Yes,
694.
by eru 3mo ago
The article says that their zstd decoder is 10x slower than the 'normal' one. I would be interested in seeing if that can be driven down to (almost) parity, with a bit more LLM time and perhaps human effort.
695.
by eru 3mo ago
The US have things like exit taxes etc which make it a pain to leave.
696.
by eru 3mo ago
Obviously, you'd go the other way round and marks things that you don't want deleted.
697.
by eru 3mo ago
That's why you do it the other way round: you mark things that you don't want deleted.
698.
by eru 3mo ago
Well, assume your phone number is public knowledge.
699.
by eru 3mo ago
That's interesting. I don't even know most of my passwords (thanks to password managers).
700.
by eru 3mo ago
Why are you staying in this place? (Is this in South Africa?)
701.
by eru 3mo ago
Has the EU ever ran an honest (or otherwise) cost benefit analysis on the cookie banner rules?
702.
by eru 3mo ago
So the question is: why are they designing AI cards so unbalanced?
703.
by eru 3mo ago
> Because markets succeed or fail based on perceived fairness. Commodities markets in the US only got anti-insider-trading rules very recently. I don't think fairness perceptions have changed much?
704.
by eru 3mo ago
You are right about RICO. Sorry, I was generalising a bit without making that explicit: > However, the [authorities] will not [use some laws] against the elite class, as ex-prosecutors end up working for the same elite through law firms.
705.
by eru 3mo ago
Or just show us your short position. :)
706.
by eru 3mo ago
Yes, but nothing 'current' about that.
707.
by eru 3mo ago
I find that modern harnesses mostly get rid of hallucinations: ask your agent to provide quotes and citations and tests for everything, and have another agent cross check that. Same way we used to do that as humans: I vaguely remember a cas
708.
by eru 3mo ago
When the author leads with a part I know a few things about (the 2000s mortgage market) and gets it so wrong, that kind of makes it hard to put much credence in the point they actually want to make later. Well, as an escape hatch you could
709.
by eru 3mo ago
> Yeah it's extremely cost-inefficient, but it works. And competition is fixing that as we speak.
710.
by eru 3mo ago
Keep your head down, while you look for alternatives. No need to pre-emptively quit, if you can coast and collect a paycheck.
711.
by eru 3mo ago
AI optimised cards is the biggest cashcow for nvidia. They are definitely doing custom silicon. And Google et al have cards that don't even pretend to be able to do graphics.
712.
by eru 3mo ago
Why should that be necessary? You can unilaterally stop trading 'before a market closes with some time buffer to prevent this.' No need for centralised action.
713.
by eru 3mo ago
That's a bit silly. Tough prosecutors and regulators get top jobs all the time. Being tough is actually a good way to raise your profile.
714.
by eru 3mo ago
Yes. As a small wrinkle: the actual observed number of producers can be small, but the market still be competitive. See https://en.wikipedia.org/wiki/Contestable_market for one case: when potential suppliers are waiti
715.
by eru 3mo ago
Agreed. But also formalising the statement of a theorem, or rather understanding the formalisation that the LLM suggested to you, is often a lot easier than understand the whole proof, especially if it's a formal proof.
716.
by eru 3mo ago
You read the stated theorem?
717.
by eru 3mo ago
No worries. I just thought I was misunderstanding something.
718.
by eru 3mo ago
Why not? Seems like they would be poorly optimised?
719.
by eru 3mo ago
Binning can take in a supposedly uniform stream of chips and produce different tiers on the output.
720.
by eru 3mo ago
How is a bathtub curve memoryless?

More ›