Vecr

42 ms·

751.
by Vecr 3y ago
Dark and/or 1983 reference?
752.
by Vecr 3y ago
They should try to implement some kind of resolute agent in that case. Might be hard to do if it needs to be "not technically search" though.
753.
by Vecr 3y ago
LPDDR5 has lower total latency (or can, and I assume it very much does in this case). Essentially the "standard motherboard layout" normal DDR5 is designed for is too big to target a lower latency.
754.
by Vecr 3y ago
What does a "Black Sun night club" t-shirt look like? I've had people (claiming to be US police) joke online about shooting anyone wearing a plate carrier with a black sun patch on it on sight.
755.
by Vecr 3y ago
Is Jane Fonda still alive?
756.
by Vecr 3y ago
It's possible he does apply the reasoning, whatever you mean by that, but just does not write about it publicly.
757.
by Vecr 3y ago
Yeah but people don't like typing in their password twice, if they even have one and don't use a fingerprint/face scan or something. I use an entirely separate and very long password for unlocking my device and for nothing el
758.
by Vecr 3y ago
Do be clear, people have been pwn'd due to this. It's not a theoretical problem.
759.
by Vecr 3y ago
Do you also have a full printed out copy of Harry Potter and the Methods of Rationality on that bookshelf? I'm not sure you understood your parent comment's point.
760.
by Vecr 3y ago
I've only read a few pages around a couple of cited parts, just to make sure the citing author is not just making things up. It's too long otherwise.
761.
by Vecr 3y ago
It's interesting to speculate about what C. S. Lewis would think of current AI systems and their future. Might not be too different from his demons.
762.
by Vecr 3y ago
> why aren't they demanding nvidia start building DRM enabled thermite charges into A100s? Tweet that at Yudkowsky, he'll probably endorse it.
763.
by Vecr 3y ago
> Asking if humanity can invent a machine to protect itself from the existential threats created by the other machines it has invented to me does not sound that intelligent. > Humanity can barely manage the existential risks for which
764.
by Vecr 3y ago
https://news.ycombinator.com/item?id=39173034 If this paper replicates, LLM-like large transformer models can approximate Solomonoff induction pretty well, with other neural net types showing clear evidence of generalizing
765.
by Vecr 3y ago
You don't want RNA anyway, you want to turn the virus's RNA sequence into a DNA sequence (reverse transcription), then do whatever editing you want once that's in your computer, then synthesize the DNA, and grow up the virus
766.
by Vecr 3y ago
> Cruise incident response playbook outlines roles of Incident Commander, SLT, CMT as well as how to respond in the weeks after incident. Playbook was not followed, said to be "aborted" because "too manually intensive&quot
767.
by Vecr 3y ago
Properly trained, well rested, and sober humans are very, very good. If you can create an AV with a single open-source software version, entirely fixed with zero changes during the whole duration, and get it to drive 200 million miles, I&#x
768.
by Vecr 3y ago
In the US people on the Internet will have ATC recordings almost immediately. The government's accountable to the people, not the other way around.
769.
by Vecr 3y ago
I actually don't know. They say in the paper they can't figure out a better way to approximate Solomonoff induction in the UTM (universal Turing machine) section, so they just guess some reasonable value (based on knowledge of the
770.
by Vecr 3y ago
> Astronomers are making the first exact maps of the Milky Way: locations of stars in three dimensions Maybe science fiction authors can stop making up locations entirely now.
771.
by Vecr 3y ago
I'm not sure what you mean, if I need something shipped directly from Colorado to my house I want it shipped directly to my house, I don't want to require the the delivery to happen somewhere else and then be put on some short tru
772.
by Vecr 3y ago
It's literally impossible to look at more than a few pages (10 maybe? that's optimistic at this point) of those results, if they even do exist.
773.
by Vecr 3y ago
Yes that's why it would be something like a Gödel machine[0]. Ignore everything it can't prove is not malicious (won't halt, runs too long, takes too many resources, etc.). [0]: https://en.wikipedia.org/wiki&#
774.
by Vecr 3y ago
Marcus Hutter invented AIXI (Hutter 2000), the theoretically optimal AI system. It's not computable though, and the computable version is hilariously inefficient. Yes it uses Solomonoff induction.
775.
by Vecr 3y ago
Wear a helmet and hip pads, but failing that split the password in half (as in literally the first half and the second half, don't try to get fancy with crypto) and give the halves to two people unlikely to collude. Your computer will
776.
by Vecr 3y ago
Maybe it's usually stored in an unheated shed or similar.
777.
by Vecr 3y ago
Yeah, probably, and the counterfactual where airlines were worse at preventing crashes would probably have worse crash survival rates as well. Maybe planes bumping into each other at very low speeds on the ground would skew the stats in tha
778.
by Vecr 3y ago
Define the enum to be represented as a two's complement integer made of hardbools.
779.
by Vecr 3y ago
If you can imagine an AI system that records some approximate representation of its own thinking, and can play that back, and can record its thinking during the playback, and play that back, and then 2-ish times more at decreasing fidelit
780.
by Vecr 3y ago
Neither have you met someone who can move 50% of the probability mass in their world model probability distribution with every bit of information they receive, but that's theoretical perfection. There are things that provably can'

More ›