Pinned Tweet
What's maddening — but also fascinating — about both "You Can See Everything" and the discourse surrounding it is how simultaneously obsessed with self-perceptiveness it manages to be without turning the camera back on Fielder as *director and producer* the full 180 degrees. 🧵
2
0
17
9,668
Great evening Black Horse and a Cherry Treemaxxing with @Abi0lvera @earth_bagel et al, putting Song Abundance into practice
1
0
1
189
It is extremely cool that this is happening and awesome people should apply to this.
I'm especially interested in exceptional nonstandard applicants: People who are new to AI safety work (although AI familiarity helps), and with competing viewpoints.
If you think RSI soon is unlikely, but evidence for and against should be shared openly, I'd love to talk.
0
0
2
291
Sigal on the bell-curve-meme effect of why just doing the thing you believe is good is good, actually.
What if Anthropic just...stopped? Unilaterally? Right now?
I talked to @DKokotajlo @_NathanCalvin @geoffreyirving about why it's time to stop framing AI as a prisoner's dilemma and just do the right thing.
I hope a lot of people inside and outside Anthropic read this
🧵 1/n
0
0
1
258
Tiny violin playing, tinily
Paper straws worse for the environment than plastic ones, says study trib.al/TEphuYk
1
0
4
320
Indeed, partially because policy choices surrounding how we structure its development and deployment (as distinct from the current “YOLO 🫰” approach) have such outsize implications for, and on, the economy!
I think there's an excellent chance AI skyrockets into being a defining issue for the 2028 election, probably alongside (as always), the economy.
0
0
3
229
Something that is important for time-buying measures like RSI research verification (“the dial”) is the capacity to know what is going on. This is a solid, constructive step in that direction.
NEW from ARI: As AI increasingly helps build the next generation of AI, Washington needs an early warning system.
Our new proposal would create an office at Commerce to track whether AI development is outpacing human oversight and safety testing.
ari.us/policy-bytes/k…
0
0
2
216
A Modest Proposal but it’s just suggesting that Kyshawn George not play the point guard position (@nathanjandrews)
1
0
0
282
The whole framing in this thread is absurd but I just have to point out that the exact same arguments in the exact same way at basically at the exact same level would also apply to Mira Nair, yet are strangely absent from the discourse for SoMe UnSpEciFIed ReAsOn
My twitter beefs are all a long con to get more people to watch Fearless (1993).
0
0
1
210
You Can Just Do Things (possess textured multidimensional opinions involving ground-truth reality and assumptions)
I often discuss sexism in AI safety & strongly advocate that the field needs to improve in how it handles harassment, but I think AI risk is serious & backed by strong arguments & now empirical evidence. There are many reputable nonprofits & professionals working on AI safety.
0
0
4
275
If you read one thing today, read this.
"The Defector" - a 5,500-word espionage tale in today's WSJ that I can guarantee is worth your time.
I spent months speaking to one of the deadliest saboteurs in recent military history, a Russian officer who helped Kyiv kill hundreds of his fellow troops wsj.com/world/russian-…
0
0
1
265
One reason I don’t conflate “we need to slow RSI progress” with being a ‘doomer’ is that the existing baseline has to be taken into account, the same way e.g fiscal austerity can be a bad/good idea depending on the environment.
Good post here from Nathan. Sure, Anthropic has taken actions that may have cost it money. But in Dario’s post, there’s a conspicuous absence of any actual pacing or slowing progress towards RSI, despite a widespread view that “alignment” is nowhere close to being solved.
0
0
0
245
.@JimVandeHei has absolutely been out ahead of AI trends for years now, but I just want to highlight how much "stuff is about to get cray" is not longer a weirdo AI-safety-people Berkeley, CA-coded fixation.
0
1
2
194
Zac Hill retweeted
@zdch Me: Democrats nominating candidates like the pastor at MLK's church against a domestic abusing former football player with a fake police badge (who still only lost by 1.5%) is why they're competitive in the Senate
Republicans: James Talarico is a pedophile
Republicans: James Talarico is a pedophile
1
1
60
1,061
Fielder employs [redacted] as another vector of processing his conflictedness, but again parameterizes what is and isn’t ’open for uncertainty’ in a revealingly narrow way.
What's maddening — but also fascinating — about both "You Can See Everything" and the discourse surrounding it is how simultaneously obsessed with self-perceptiveness it manages to be without turning the camera back on Fielder as *director and producer* the full 180 degrees. 🧵
0
0
0
277
This feels like an important forcing function as the space aims to insinuate itself into greater positions of consequence.
What I intend to do: I am looking to write a detailed report which investigates issues of sexual harassment & sexism in AI safety. I want to be able to cite as many concrete examples as possible. I plan to use this report in discussions with org leaders & may write a public post
1
1
4
1,066
Paging @stripepress or eg? Maybe @aishdoingthings??
Also, I have drafted a book-length treatment of Maximum Republicanism. If you want to partner to publish it, or know someone who would, DM.
It covers: the history of the party, the contemporary landscape, the maximum philosophy, and why New York is an excellent place to grow the party.
1
0
5
373