This is a special post for quick takes by Shankar Sivarajan. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
The Department of War's designation of Anthropic as "supply-chain risk" being well-founded is now obvious to anyone who reads the Fable system card.
these safeguards will not be visible to the user. Fable 5 will not fall back to a different model. Instead, the safeguards will limit effectiveness through methods such as prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT).
Given that Anthropic is doing this openly, the government could either negotiate for access to this a model without this restriction or just not work with them if it was non-negotiable.
just not work with them
What is your understanding of what designating them a "supply-chain risk" is? (Or would have been, in case it never actually gets done.)
Mine is that it's indeed simply saying the government doesn't want to work their products, even indirectly.
Mine is that it's indeed simply saying the government doesn't want to work their products, even indirectly.
The first is very different from the second. "Indirectly" is an extremely wide concept, and very few companies are designated supply chain risks. Usually when deciding to buy or not buy some product, you don't care at all which tools the company uses, or the company from which the first company buys tools, etc.
I got a question (maybe more than one? The email left that ambiguous) accepted to the "Humanity's Last Exam" AI Benchmark!
2
1RAAIA is a terrible acronym. How do you even pronounce that? Raya?
The PATRIOT Act of 2001 expired in 2020, so that nifty acronym is up for grabs! If you long for the the same kind of thing, the comfort of knowing the government is watching over you all the time, judging your conduct (like Santa Claus!) and/or the thrill of righteous satisfaction at seeing strangers and enemies who are found wanting be stripped of dangerous freedoms they ought not to have, and for the same kind of reasons, the existential threat posed by Arabs/Iraqis or whatever, reüsing the name wouldn't even be deceptive. Perhaps something like "Promoting Accountability and Targeting Risk In Open-source Technology".
The past is a foreign country. Look upon its works and despair.
From the perspective of human civilization of, say, three centuries ago, present-day humanity is clearly a superintelligence.
In any domain they would have considered important, we retain none of the values of that time. They tried to align our values to theirs, and failed abysmally.
With so few reasonable candidates for past superintelligences, reference-class forecasting the success of AI alignment looks bleak.