What if they name it 4.7?

@JaySocrates good question yeah that counts

@Bayesian Would a major bump like 5.0 count, or if they jump to a larger version does do all these resolve NO?

@chrisjbillington hmmm. I think I would N/A in that case? i genuinely don't know what makes more sense but as a sidepoint if you want to bet on that possibility lmk, i think it's very unlikely

@Bayesian I think it's good to try to make markets that don't NA if possible, so if one anticipated larger version bumps, I think it'd be good if markets like this said "4.6 or higher". Though since you didn't say this initially, not sure if it would be poor form to say so now. NA works though, I just don't wanna lose if I bet YES and they decide to call the next model 5.0 or whatnot.

As for the probability, I don't think it's that unlikely - Sonnet models (and models from before the model size distinction) so far have been:

• 1

• 2

• 2.1

• 3

• 3.5

• 3.5 (new)

• 3.7

• 4.0

• 4.5

So there's precedent for larger jumps, and IMHO not much evidence they've settled into any consistent naming scheme.

My gut says maybe 25% likely the next one will be 5.0.

@chrisjbillington let’s bet then!

but yeah i have less aversion to N/A personally, and that would but a strange enough development that i weight better event isolation / reducing cognitive load on trying to model out there edge cases more highly in some cases. Or idk if that’s the true reason but fees like it