I think open source is <12 months behind the frontier, and the same security reaction you saw with Mythos/Fable will likely happen with open source sometime in the next year.
To my mind, the escape routes are
(1) Frontier models desensitize the security apparatus
yep - there is no way to hold a consistent belief set where you’re agi pilled and pro open source and this has been obvious since ilya wrote this 2015 or whatever. enormous cope ensues
substack's adding pangram integration, and i really like both the initial intervention - click-to-reveal - and the described (uncertain!) path to future interventions, like filtering out anything that scores highly. seems thoughtful & measured.
my feed's j-space reactions are
* 50% 'the authors claim too little about AI consciousness'
* 50% 'the authors claim too much about AI consciousness'
which is probably more a statement about my feed's composition than anything but still feels like a W
i suspect humans overrate the importance of continual learning because humans don't come with a make-me-generally-smarter lever. and so humans have to fall back on other, less-general levers
