Post

Conversation

Zvi Mowshowitz
@TheZvi
Okay, this is actually about as good as we could have hoped for under the circumstances. We might be seeing Actual Progress.
Image
Quote
Dario Amodei
@DarioAmodei
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our
21.5K
Views
View quotes
David Watson 🥑
Post your reply

Wickemu
@wickemu
But will OpenAI follow suit? My worry is that OpenAI has so many skeletons in their closet that they're afraid to allow embedded evaluators and discover what they've done or allowed to happen, even if their record is cleaner going forward
232
Undefined
@UndefinedDzx
No. Parts about absolutely mandatory keeping ahead of China tells this isn't really a slowdown except for PR purposes. And advocacy of AI chip sanctions and signing Bessent's tune indicate that no bilateral treaty (or global treaty including China) is wanted or expected.
116
AI Basics
@AiBasics49332
RSI has already started - what????? This is the first declaration of it I’ve seen. Is govt aware? Are most ‘experts’ even aware? And critically, what does this say about loss of control already???
@DKokotajlo
@_NathanCalvin
What a time to be a parent (or any human)
Image
275
Vincent Bagayoko
@bagayoko87528
Yes, very pleasantly surprised by the first part! What would be better: much much more precision and concrete steps taken, soon. If you can contribute to getting more clarity on that, that would be amazing.
188
brian smithson
@briansmith71979
As expected the real play: trying to restrict the use of "unsafe" open source models. The regulatory moat the frontier labs need to make more money.
78

Discover more

Sourced from across X
Sydney
@SydneyVonArx
Crucially, OpenAI did _not_ reveal this attack. This article makes it sound like they disclosed it, but in reality we had to dig it up. We can't trust OpenAI to notice and disclose these incidents.
Quote
POLITICO
@politico
OpenAI reveals another rogue AI attack dlvr.it/TVRYF5
Readers added context
The attack, which first occurred in May, was revealed first on Friday by a team of independent researchers. OpenAI admitted to the attack only after it had been attributed to them. reuters.com/legal/litigati…
Do you find this helpful?
Rate it
40K
Cate Hall
@catehall
old enough to remember when autonomous weapons were a tactic to avoid discussing actual harms like algorithmic bias
Quote
WIRED
@WIRED
Timnit Gebru argues that AI companies are stoking fears of extinction to avoid discussing actual harms, like autonomous weapons. wired.com/story/one-of-a
27K
Tenobrus
@tenobrus
third party evaluators are finally being empowered. this is such an insanely good step. ofc it's not enough , we obviously need this to not be a voluntary choice Anthropic makes, but great start. METR has like 40 employees. u should drop everything ur doing and go apply to help!
Image
Quote
Dario Amodei
@DarioAmodei
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our
6.1K
Peter Schmidt-Nielsen
@ptrschmdtnlsn
Ed Zitron is accusing Jacob Coxon of "grift" for leaving a 7- or 8-figure per year job to... start a substack. You can think Jacob is wrong, but accusing him of acting in grifty self-interest here is simply objectively *insane*.
Image
18K