Obviously we're all in the same epistemic boat here so there's not much to do from outside Russia but wait, but I do want to comment on the shared assumption here that this might be a strain strengthened by post-2022 DL techniques. That's certainly possible, but by no means guaranteed and IMHO unlikely -- many established labs have yet to integrate models deeply, and Russian labs would be working with an abliterated GLM-5.3 at best.
So... kinda a reason to have some hope, here. AI could even help us mitigate a potential global catastrophe with fast iterati... (read more)
Obviously we're all in the same epistemic boat here so there's not much to do from outside Russia but wait, but I do want to comment on the shared assumption here that this might be a strain strengthened by post-2022 DL techniques. That's certainly possible, but by no means guaranteed and IMHO unlikely -- many established labs have yet to integrate models deeply, and Russian labs would be working with an abliterated GLM-5.3 at best.
So... kinda a reason to have some hope, here. AI could even help us mitigate a potential global catastrophe with fast iterati... (read more)