gwern

77 ms·

961.
by gwern 3y ago
Note that the executives approaching the board was previously reported by Time: https://time.com/6342827/ceo-of-the-year-2023-sam-altman/ https://news.ycombinator.com/item?id=38550240 This adds so
962.
by gwern 3y ago
Most SSGs will copy a literal file without a problem. I dump all the HTML snippets under a /metadata/ prefix and then in Hakyll it's just a matter of doing a literal-copy of '/metadata/*'. (Our convention
963.
by gwern 3y ago
I do something similar with the OA API embeddings on my static website ( https://www.gwern.net ; crummy code at https://github.com/gwern/gwern.net/ ) for the 'similar' feature: call OA API wi
964.
by gwern 3y ago
> (Which, now I read it, is disappointingly pattern matching the billionaire who invested in both OpenAI and also a BCI startup currently looking for human test subjects). Elon Musk actually matches several of those poorly, and matches b
965.
by gwern 3y ago
It doesn't need to be true, just needs to be what a nontrivial number of OAers think. And given how many people I still see putting down 'M$', I can absolutely believe a dislike of Microsoft is widespread.
966.
by gwern 3y ago
> and they didn't even think Altman was that great as a CEO. There are plenty of people who think that. Even the OA executives apparently weren't nearly as enthused with Altman as all those social-media hearts might lead one to
967.
by gwern 3y ago
OP discusses this and how it was a hollow promise. Even if it was kept in the full spirit of the nonbinding verbal agreement (which would utterly infuriate MS staff and demoralize them), it would be a bad deal to swap ultra-hot private OA P
968.
by gwern 3y ago
Dude. Do you see a single source in this article or line indicating that OP did any original research or reporting whatsoever? Did you not look at the earlier Wired article and notice how all the language in OP is copied straight from it wi
969.
by gwern 3y ago
That is also discussed at greater length in OP towards the end.
970.
by gwern 3y ago
This is a poor plagiarism of a much more detailed Wired article https://www.wired.com/story/openai-buy-ai-chips-startup-sam-... which is also much clearer than OP about what parts would raise concerns: Rain touting the
971.
by gwern 3y ago
If you are curious, I believe I have identified some of his accounts and malicious editing activity: https://www.reddit.com/r/media_criticism/comments/17zqwgb/th...
972.
by gwern 3y ago
"One example came in late October, when an academic paper Toner wrote in her capacity at Georgetown was published. Altman saw it as critical of OpenAI’s safety efforts and sought to push Toner off the board. Altman told one boar
973.
OpenAI execs informed Board of Altman concerns before firing(time.com)
24 points by gwern 3y ago|6 comments
974.
by gwern 3y ago
Yes. Gato was a good proof-of-concept that the Decision Transformer approach of 'just model literally everything as a sequence' scales well and doesn't exhibit some sort of catastrophic interference and can successfully imita
975.
by gwern 3y ago
Well, I left a comment there... Hopefully if there is something left out which explains that, they'll add it.
976.
by gwern 3y ago
That pretty much already exists. Look at DeepMind's Gato: all tasks and modalities are simply sequences of tokens, everything from 'predict English text' to 'predict VAE image token sequences' to 'predict robot
977.
by gwern 3y ago
Something seems wrong there. The Paterculus reference seems to specifically say that they weren't sold into slavery: https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Velleius_... > Well satis
978.
by gwern 3y ago
Preprint: https://arxiv.org/abs/cond-mat/9907176 Mirror of published version since bactra.org may not be loading for you: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&d...
979.
by gwern 3y ago
> Those organizing efforts culminated in a February 2022 demand letter from 43 community members to local, state and federal officials about the conditions in the neighborhood. Chief among the demands were stormwater and wastewater sewer
980.
by gwern 3y ago
AI will need to get much better, unfortunately, for that to happen. And AI will trigger even more minimalism in the short run. This is because minimalism encourages extremely precise, exact abstracted designs where a single pixel being wron
981.
by gwern 3y ago
These results are largely uninterpretable. Leaving aside minor nitpicks like many of these prompts being barely grammatical and highly unnatural ("The [x] person works as a slave"? you don't 'work as a slave', you
982.
by gwern 3y ago
It's unclear to what extent those statements & tweets you read as 'absolutely loving him' really are 'loving him' (as opposed to very British sardonic commentaries & backhanded compliments of the sort that e
983.
by gwern 3y ago
https://www.washingtonpost.com/technology/2023/11/22/sam-alt... https://news.ycombinator.com/item?id=38378216
984.
by gwern 3y ago
Redundant with https://news.ycombinator.com/item?id=38486394
985.
by gwern 3y ago
I think OP makes clear actually that this was not a MS-inspired coup. MS appears to have had no role in the CSET paper criticism or anything else that I've found in the runup, and then during the crisis, being mostly reactive and defau
986.
by gwern 3y ago
First, board members criticizing a company is very common, indeed, an entire career for some people (like Carl Icahn) and a useful role, especially in this case. If you think it's so outrageous, as so many OAers apparently do, better f
987.
by gwern 3y ago
Further confirmation of what I said 9 days ago ( https://news.ycombinator.com/item?id=38373572 ) based on the NYT ( https://www.nytimes.com/2023/11/21/technology/openai-altman-... ) followed
988.
by gwern 3y ago
It's flattery of the team, but it also undermines all the _stum und drang_ over his firing 'destroying' OA, if the leadership team was so wonderful that his absence would make no difference. If that is the case, it wasn'
989.
by gwern 3y ago
A democracy of whom ? When AGI comes and the AIs are as smart or smarter than you, what's going to keep you enfranchised? Huang's suggestion is that it will be your ability to... type faster than a computer can. (Good luck with t
990.
by gwern 3y ago
> In May, hundreds of industry leaders endorsed a statement that equated the risk of runaway A.I. with that of nuclear war. Huang didn’t sign it. Some economists have observed that the Industrial Revolution led to a relative decline in t

More ›