Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
gwern
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
35 ms
·
511.
▲
by
gwern
2y ago
> What did all this amount to? Neither “Rat City” nor “Dr. Calhoun’s Mousery” seems quite sure. In the preface to the former, Ramsden and Adams explicitly say that they are not going to “evaluate the merit” of Calhoun’s work. In the epil
512.
▲
by
gwern
2y ago
> I'm not saying a commercial motive is a bad thing - hardly... but this quote seems to be odd given the circumstances. It only seems odd to you because you are reading much more into it than he's ever said, like "AI shoul
513.
▲
by
gwern
2y ago
> Cliff Stoll confessed that he doesn't believe in space colonization and interstellar travel. Perhaps you mean Charles Stross?
514.
▲
by
gwern
2y ago
https://www.wsj.com/tech/ai/openais-complex-path-to-becoming...
515.
▲
by
gwern
2y ago
At the time, I read the Apple deal as OA strategically finding an alternative to MS as the sugar-daddy, in terms of free cash flow and datacenter capabilities: it didn't need to be true, it just needed to be a compelling negotiation po
516.
▲
by
gwern
2y ago
Exactly. The problem here is that OA is too valuable, and the non-profit owns too much of it (ie. all of it). The non-profit owns exactly 100% of the OA for-profit right now, and can cancel the "PPUs" any time it pleases, and
517.
▲
by
gwern
2y ago
The original cited paper https://arxiv.org/abs/2210.08277 "Deep Differentiable Logic Gate Networks" struck me then as very clever, ultra-efficient/small, hardly any inductive bias or prior... But you hav
518.
▲
by
gwern
2y ago
https://blog.oimo.io/2022/02/11/clock-core/
519.
▲
by
gwern
2y ago
Reminds me of the falling-blocks one, where IIRC he made the blocks 'spontaneously' create the necessary time by starting with them in the right order and then simply displaying it in reverse, due to the temporal reversibility of
520.
▲
by
gwern
2y ago
Probably not much. As OP points out, this is not even the first time OA accounts have been hacked. It mentions the CTO and chief scientist and a top researcher already got hacked - and in fact, omits that the same researcher got hacked yes
521.
▲
by
gwern
2y ago
A general reply to all of the comments thus far: you are completely missing the point here. OP is not a meaningful forecast, and it's not about nuclear power or any of that. It's about laying the groundwork for the privatization a
522.
▲
by
gwern
2y ago
> If we want to put AI into the hands of as many people as possible, we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips). If we don’t build enough infrastructure, AI will be a very lim
523.
▲
by
gwern
2y ago
The problem with that is that users (1) often click at random or for stupid reasons (I was shocked when one user contacted me to complain about my site acting poorly when they clicked on random empty whitespace - they said they just did tha
524.
▲
by
gwern
2y ago
Yes, but the important thing is the evaluation function of what happens after random actions - his intuitive expectations and things ripping his skin. A web fuzzer can click around randomly like he does, but it can't know things like
525.
▲
by
gwern
2y ago
(You should probably spoil the surprise, because few of the readers of your comment are going to track down & buy a 16-year-old RPG, and spend the necessary 10 or 40 hours to reach that point and learn what the surprise is and how it il
526.
▲
by
gwern
2y ago
Musk has long enjoyed the benefits of treating his empire as a conglomerate, without the costs of things like shared liability, doing things like casually reusing Tesla engineers in Twitter or sending Twitter's GPUs to X.ai, and daring
527.
▲
by
gwern
2y ago
In these cases, the models are so small, and have to run on very limited CPU, that the training time is going to be fairly negligible. The top-listed program, nncp, is just 0.06b parameters! https://bellard.org/nncp/nnc
528.
▲
by
gwern
2y ago
Or Recraft https://www.recraft.ai/ if you want SVG.
529.
▲
by
gwern
2y ago
It was also solved by the even easier approach of aspect ratio conditioning, where you just pass in the dimensions of the crop to the NN like SDXL: https://arxiv.org/pdf/2307.01952#page=3&org=stability
530.
▲
by
gwern
2y ago
OP reads like a lazy ripoff of earlier debates around Lumina, principally https://trevorklee.substack.com/p/please-dont-take-luminas-a... , which I don't see acknowledged or linked at all.
531.
▲
by
gwern
2y ago
Digital Antiquarian has a few posts up recently about Douglas Adams, and it all sounds like it was a great deal more painful than I had ever realized as a kid reading _Hitchhiker's_: https://www.filfre.net/2024/07&
532.
▲
by
gwern
2y ago
> I happened to leave my mouse briefly over one of the drop-caps, and it turns out there's an entire page dedicated to _generating drop-caps_ with AI I'm very pleased to read that! I pushed Achmiz to implement that - "can
533.
▲
by
gwern
2y ago
This is a strange comparison. It is a humorous throwaway line about SF, where the actual 'subject' is not SF at all. (It's nonfiction writing.) That aside, your Kirk/Spock comparison is also wrong: the difference between
534.
▲
by
gwern
2y ago
"Facehugger" is a lot more familiar a term, and yeah, I didn't even realize there was a difference because a facehugger leads to bursting-chests, at least in what I remember of the 3 Alien movies I watched. (I also keep think
535.
▲
by
gwern
2y ago
It is not me.
536.
▲
by
gwern
2y ago
> An OpenAI spokesperson told Fortune and Reuters that the company's non-profit arm is "core to our mission and will continue to exist," as OpenAI stays "focused on building AI that benefits everyone." Not exactl
537.
▲
by
gwern
2y ago
'Total quality management' covers many of the same thoughts. Here's a good write up on exactly this topic, the self-reinforcing nature of inefficient work, I ran into a few days ago: "Nobody Ever Gets Credit for Fixing P
538.
▲
by
gwern
2y ago
Hard to say. They could have fiddled with the rules to make it really unappetizing and kill the market, or slowwalk every last thing (how many vaping products have been approved?). It could have taken many more years and been a lot worse, w
539.
▲
by
gwern
2y ago
Especially here. Expecting good faith in hearing aid regulation - from the FDA ?! Remember, Congress authorized OTC hearing aids back in the Trump administration, in August 2017 ( https://en.wikipedia.org/wiki/Over-th
540.
▲
by
gwern
2y ago
This is a parallel development. It will probably feed into future 'Orion' systems using GPT-5, though, is the current thinking.
More ›