Yet another Hacker News

Anthropic CEO Says It's Time to Slow AI Model Advances

20 points sbulaev 54 minutes ago 16 comments

elmer2 43 minutes ago | parent

If they do this, it will just give China time to catch up.

achow 19 minutes ago | parent

China will also slow down. They are catching up with SOTA models due to distillation https://www.cisa.gov/news-events/cybersecurity-advisories/aa...

atmosx 14 minutes ago | parent

lol. Every time one of these ppl says something like that, it's because he just came out of a meeting where he learned that his direct US rival is beating him... So he is suddenly interested in regulation. We've been there with Altman, Musk, etc already. But at the time Anthropic was ahead :-)

exabrial 27 minutes ago | parent

Meaning: slow his competitors; mainly because anthropic stopped focusing on product and quality. Instead their big push right now is preventing others from catching up.

throw03172019 19 minutes ago | parent

Is this to strengthen financials? Less burn on training?

aczerepinski 19 minutes ago | parent

I was prompted to accept binding arbitration in order to read this, so I didn't. Feel free to summarize.

winstonwinston 17 minutes ago | parent

Why makes no sense at all unless you are very naive trusting they would not jump straight to the endgame if it was possible with the technology available.

misnome 17 minutes ago | parent

Sounds like they hit a wall and are looking to score points for it.

conqrr 16 minutes ago | parent

Meaning they realized they hit a wall. The next new model isn't that much better than the last in quality.

jordanb 6 minutes ago | parent

They can get a "government mandated freeze" on the grounds that "ASI is at hand" then they IPO promising to deliver right after the freeze ends and you, lucky investor, will be able to take advantage of the pause to get in right before this thing roars past Pluto!

s1artibartfast 15 minutes ago | parent

Good.

It will take decades of economic and social innovation to catch up with current model capabilities.

No upside is worth the potential downsides of getting this wrong, even if you set aside all the x-risk stuff.

5555watch 15 minutes ago | parent

Recalling their recent misuse report [0], I kind of agree with them. And performance/intelligence wise I think we are already at quite a nice local optima.

[0]: https://www.anthropic.com/threat-intelligence-report-septemb...

karim79 12 minutes ago | parent

Yes. We must always trust these ass hats because they have our well-being at heart.

encyclopedism 5 minutes ago | parent

Everyone, collectively, stop thinking about neural networks too 'hard'. Whilst you're at it stop doing maths too!

LogicFailsMe 3 minutes ago | parent

Math is tough! Let's go use open weight models!

LogicFailsMe 3 minutes ago | parent

Those regulated models aren't just going to capture themselves!