@ retweeted
I'm pretty confused by the reaction to Dean's essay. It seems like Anthropic are aiming to build AGI and then hand off all decision-making, and getting criticised for it, and OpenAI are saying that handoff isn't the plan, and getting praised for saying this. But OpenAI's actions aren't consistent with that at all! They're still trying to develop AI which outperforms humans in all economically relevant domains.
It also seems underdiscussed that there's nothing in Dean's essay that would actually prevent the kind of human disempowerment he's worried about. He talks about maintaining human oversight and holding humans accountable for their agents' actions, but I don't see how that would be anything other than rubber-stamping, if AIs do indeed become faster and more capable than humans at all the tasks they're doing. It'll be like "overseeing" hundreds of workers running hundreds of times faster than you at a task they can do better than you; of course, you can check their work carefully, but this will just slow you down and disadvantage you, even if you can understand it all! (Dean discusses this dynamic in the essay. I'm not sure how he thinks his ideas would escape it.)
Figuring out how you could prevent human disempowerment is an extremely hard problem. I'm not sure it's compatible with building AGI. So I'm not criticising Dean for not having fully solved the issue: I'm glad OpenAI are thinking about this and that Dean wrote this essay, and it's important to have more public discussion! But the positive response I've seen to this essay seems overstated, given that OpenAI are continuing to build AGI and have no good ideas for what to do to prevent disempowerment.
@ retweeted
If I have a subscription to your publication or service, the AI model I'm working with should have the same level of access that I do and not be paywalled or blocked. I think this is part of the intended purpose of computer use, but we should be able to designate a model as well.
@ retweeted
The purpose of the economy should be to make one guy really rich. And I think we can all agree that guy should be a computer
@ retweeted
"I personally don't want to ever die, therefore accelerate AI to the max" was already a motivation for a significant fraction of AI prominent researchers in the 1970s. (I know because I was there then.) xitter.gq/gwern/status/209781173…
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
Since the vulnerability has leaked: Calif discovered and reported Java4Shell (CVE-2026-114938).
We're publishing technical details by 1 PM Pacific today, after Oracle releases a patch.
Stay tuned!
@ retweeted
It’s hard to believe how hyper-literary america was 20 years ago in the age of David Foster Wallace, maximalist “hysterical realism,” Joanna Newsom’s Ys, Sufjan Stevens’ America project, the heyday of McSweeney’s etc. I was psyopped as a child into thinking it would go on forever.
@ retweeted
Google Maps users rate the Getty Center 4.8/5 stars (26,199 ratings) and the Art Gallery of New South Wales 4.7/5 stars (9,838 ratings), but NSW is full of mediocre art while the Getty Center is a continuously phenomenal experience of some of the best art and outdoor views.
@ retweeted
I regret to inform you all that we have found our first long false positive from 2006
it's a poem where the same 6 words are repeated four thousand times
from 2006, earliest archive is 2013
somehow i don't think you're going to count this one though
web.archive.org/web/20130214…
@ retweeted
The incredible thing about COVID is that people want to avoid thinking about it so badly that even when people do make movies about it, they get memoryholed to the point that people don't even know they exist. It's a genuine antimeme.
It’s really fascinating how “train a model to directly output art” is being progressively displaced by “train a model to be an artist.” (Image + video generation being the other recent examples)
check out pr #24 i opened to convert musicxml notation into notes so ppl can do fun arrangements and performances from musicxml files




























