Sorry, is this one of the bad endings or one of the good endings?

@blank depends on what happens at the end of the fast takeoff. But lots of optimistic people seem to be optimistic in part because they don't expect a fast takeoff.

@blank not by its nature either

bettting no on this because i its happens it is a win either way

Some of you think this has a >1% chance of occuring??

@diracdeltafunk roughly 4%. There's 80% chance it happens within 10 years.

@jim aight well I'm happy to take the other side for this market :) 80% within 10 years sounds much more reasonable to me than 4% within 8 months.

@diracdeltafunk I think it's unlikely to happen soon because discovering the prerequisite technology requires a lot of compute. Like there's a "no free lunch" thing going on which adds some protection against foom. Got to spend compute to make more compute.

@diracdeltafunk like I think just magically stumbling on significantly better algorithms is very rare. It generally requires a lot of compute to discover them. Even if those algorithms potentially exist that would make everything suddenly go crazy, I think the chance of us suddenly acquiring them without spending huge amounts of compute is very low.

@MalachiteEagle > magically stumbling on significantly better algorithms is very rare

probably not neccesary, we're on track for AGI-soon with the current rate of innovation

@jim yes but the AGI still has to spend compute in order to improve. It’s not a magic wand.

I think this requires a lot of things that seem unlikely at present. It doesn't violate the laws of physics (as far as we know) to suddenly stumble on a new algorithm that far surpasses deep learning, which can immediately be scaled up on existing clusters. But even then, it's not clear that this can be leveraged in such a way that in 12 months the physical environment is radically transformed by technology. It would have to require multiple steps in the physical world, each of which requiring significant physical experimentation