Date: November 18th, 2026 9:28 AM Author: zarathustra
a kind of diametric opposite of the typical positive astroturf campaign that usually follows a new model release, same methods opposite sentiment. suspect that these flame wars are a new part of token merchant warfare going forward.
Date: November 18th, 2026 9:44 AM Author: The Metaphysician
Yeah I have noticed this too. Funny thing is none of it has to do with actual intelligence metrics. It is all bullshit subjective stuff like "it feels lazy", and "it isn't enough of an advance". Like bro, chill out, 6.2 sol is literally same intelligence index as fucking Astra 6 was for a fraction of the price, that's big enough. 6.2 Astra isn't insanely better at coding, but it is significantly more intelligent on other things, and "coding" is pretty much saturated at this point anyway. There is no way you can need something "better" at coding unless you want it to run autonomously 24/7 and not understand any of the underlying frameworks and systems you are using at all.
Another thing, nothing they are saying to shit on it can override for me the fact that Claude has an insufferable personality. IMO opus 5.5 is overrated even based on objective ability, it is all hype. Biggest thing it is better at is graphic design stuff. It is worse at coding and problem solving than sol 6.2 imo. New Fable model barely scores higher than opus 5.5 on intelligence metrics too. The companies are pretty much neck and neck right now on objective metrics for models. The people saying anthropic has a huge lead are lying, unless you are talking revenue and compute access and even that is exaggerated. My reason for sticking with open ai is that anthropic and claude are more annoying.
Date: November 18th, 2026 10:12 AM Author: zarathustra
model choice no longer really matters for normies, prompting and model management is where you get your roi. real takeaway here is that twitter is pr theater that is no longer worth participating in, if it wasn't clear already.