>"We can say categorically that it is impossible for Dr Buckmaster's Codex prompts over the last two months to have influenced the system in any way, including training," they told the ABC.
When did it become "impossible" for it to happen? Wasn't it left ambiguous when Buckmaster asked the team [0], and kind of likely given the timing and both of the proofs being based on Diego C´ordoba and Luis Mart´ınez-Zoroa's work?
Anyways, I'd say this is more about AI Companies not being
nearly as interested in scientific progress as they are in marketing (who would've thought?) than anything else. This also disregards Diego C´ordoba and Luis Mart´ınez-Zoroa's multiple years of work on the subject.
>"We can say categorically that it is impossible for Dr Buckmaster's Codex prompts over the last two months to have influenced the system in any way, including training," they told the ABC.
When did it become "impossible" for it to happen? Wasn't it left ambiguous when Buckmaster asked the team [0], and kind of likely given the timing and both of the proofs being based on Diego C´ordoba and Luis Mart´ınez-Zoroa's work?
Anyways, I'd say this is more about AI Companies not being nearly as interested in scientific progress as they are in marketing (who would've thought?) than anything else. This also disregards Diego C´ordoba and Luis Mart´ınez-Zoroa's multiple years of work on the subject.
[0] https://cims.nyu.edu/~tristanb/statement.pdf (page 3)