Tweet fixado
Moon
1.7K posts
English
So @Anthropic about to learn the @SpaceX ITAR/EAR lessons
Will be very hard for non-nationals to work there and @OpenAI on frontier models.
Suppose AGI is the ultimate dual purpose technology
English
Two possibilities, both extremely bad:
1. The admin is deliberating fucking with Anthropic for Schmittian reasons
2. This is now precedent for any model with Mythos-level capabilities and beyond
When OpenAI releases a comparable model will they face the same controls?
Andrew Curran@AndrewCurran_
From the statement: 'The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, 𝘪𝘯𝘤𝘭𝘶𝘥𝘪𝘯𝘨 𝘧𝘰𝘳𝘦𝘪𝘨𝘯 𝘯𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘈𝘯𝘵𝘩𝘳𝘰𝘱𝘪𝘤 𝘦𝘮𝘱𝘭𝘰𝘺𝘦𝘦𝘴. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.'
English
Anthropic's "core views on AI safety" used to include that they "do not wish to advance the rate of AI capabilities progress", even to the point that they thought they needed to be careful in even demonstrating frontier capabilities to anyone (in publications or otherwise).
anthropic.com/news/core-view…
English
I don’t remember if this was direct from Dario / early team / EA community or this was a public position, but I distinctly remember Anthropic’s stance as a safety org was to not encourage a capabilities race by not releasing the strongest model at any given time. So… why did they launch Mythos / Fable??
Phil Chen@philhchen
When I was considering joining Anthropic in 2023, their main pitch was “we have prioritized safety over commercialization and actually made decisions that hurt our business because they were the right moral choice.” Everyone I met at the company was incredibly sincere about their mission alignment. I’m cynical and don’t believe that some people joining Anthropic today would hold this position, but I do think the early team really believed in this mission.
English
@TomDavidsonX Hope there is alternative to standoff betw. middle countries + US + cooperation with 🇨🇳 instead. 🇨🇳 is strong on AI but as long as the government is authoritarian we might rather demand NVIDIA chip allotment and frontier access from US for providing key tec (ASML, Zeiss, Micron).
English
Moon retweetou
@flowersslop you're aware that they can neither catch up ever on aura nor revenue?
English
@scaling01 Europe will fumble this but probably invest more than 2B.
China will not get substantially more AGI pilled from this, they never had access to OAI or Ant models before.
English
Prediction time:
- Europe will absolutely learn nothing from this incident. At best they are going to invest an additional 2B Euros in some robotics or image gen startup, because they think that's somehow relevant
- China will get even more AGI pilled and restructure their entire economy for one purpose
English
Can someone explain to me like i am dumb, that AI American private companies not publicly traded, are worth almost 1T, without revenue or assets to back it up?
Are Europe's financial investment markets that fragmented, that we can't raise 100billion for Mistral, or are the Americans just king at exaggerating their valuations?
Emmanuel Pernot-Leplay@PernotLeplay
🇪🇺 @MistralAI is about to raise €3B at a €20B valuation 🚀 It could even be more, investors are not yet disclosed. Last year it raised €1.7B at a €12B valuation, when ASML became its biggest investor. OpenAI is valued at €736B and Anthropic even more.
English
Moon retweetou
So everyone was showing off a bunch of sims made with Claude Fable. I thought I'd try too. A year ago I was nerdsniped figuring out how to find a projection function from a point to the screen around a black hole. Couldnt do it before, but Fable just oneshotted it😅
English
Moon retweetou
My first (and only, it seems) test of Fable: use the opengl-cuda interop to make a demo renderer that bridges between moderngl and my RT cores. Opus took that demo and adapted it for Fluoddity.
Most videos I post take a half hour plus to render. These are all realtime on my 5060
English
@MLStreetTalk i don't think you understand the overpowering dominance of aura
English
This will become a textbook example of how a company snatched defeat from the claws of victory, their bullshit game spectacularly backfired.
You reap what you sow.
Jeremy Howard@jeremyphoward
I disagree with this decision and I don't like it. But also... HOW DID ANTHROPIC NOT SEE THIS COMING‽ It is *the* obvious response to "this is too dangerous for anyone except us to use", since that relies on a premise ("we are uniquely good") that almost no-one agrees with.
English
@bpodgursky If she has a green card, she's a US person for these purposes
English
Amanda Askell isn't a US citizen. Anthropic chief philosopher can't talk to Mythos until this is sorted.
English
What Europe should do right now:
1. Call all the European researchers working on AI and return them back with same salary (or they can stay but switch career).
2. Fill EU places having GPUs with money, and put those people there.
3. AI partnerships with China + India.
English
i am genuinely wracking my brain and cannot decide if SpaceX is under or over-priced
Brew Markets@brewmarkets
SpaceX is now the 6th largest public company in the U.S.
English
@lacker @Afinetheorem more than average could be enough for alignment purposes re avoiding catastrophe
English
@Moonicker @Afinetheorem I would expect them to be more aligned than the average person. The same way that the average opinion of a group of 100 people is closer to the opinion of a random person, than the opinions of two random people are to each other.
English
The fundamental paradox with AI safety: intelligent *humans* with tools made the nuclear bomb, and the cure for polio, and the mindless YouTube videos. We have not solved the *human* alignment problem. (But we are still better off on net with this bundle, to be clear!) 1/2
English