I know I sound like a broken record at this point, but GPT 5 really doesn't seem like a paradigm shift at all. Yes there's progress being made, but it feels to me it's becoming marginal. But surely, in 6 - 12 months, everything will be vastly different right? Or is that just what these labs have to parrot to maintain their funding? It's worth mentioning that even without improvements to the LLMs themselves, there's still a lot of work being done on tooling and UX. That seems to be where most of the progress is happening these days.
@neetcode1 As far as I can see, at least on my timeline, no is claiming it’s a paradigm shift. Personally, it’s a decent bit better than sonnet 4 on coding tasks, specifically monorepo full stack work (purely just vibes based from using it today).
@neetcode1 no you are not a a broken record, it hits right when u post your take
@neetcode1 it's not scaling properly, the right approach is still open source
@neetcode1 Dude, just enjoy it while you can, ’cause in 6–12 months we’ll be deciding which one to use: - GPT 5o - GPT 5o mini - GPT o5 mini-High - GPT 5o nano - GPT 5.1 - GPT 5.1 mini - GPT 5.5 - GPT 5.5 mini - GPT 5++ - GPT 5# - GPT 5 Air - GPT 5 Pro - GPT 5 Pro Max - GPT 5 Ultra
@neetcode1 Everything you said is correct. House of Stark be like, "Winter is coming."
@neetcode1 how do u benchmark whether there is an improvement or not?
@neetcode1 i'm sorry but why every update has to shift paradigm.
@neetcode1 AGI has become a hype term. Everybody would be better off if we were working on getting AI for niches.
@neetcode1 I tried to add some js on my easy admin custom html forms, nothing too complex need a couple of events where changing value js should fetch some data and populate next field. The result was very off and at the moment claude4 still does it better
@neetcode1 I think, I agree, but from the perspective of a university student, even if the benchmark improvements are marginal, the cost to intelligence ratio is quite something to look at, especially since I feel like this will cause another layoff season, and lower job positions for me :(
@neetcode1 i think there is a lot of hype for this, both people and companies are doing it for AGI, ASI, etc... I think all of this is only for funding :) But to get to that level, it would really take more timeke maybe 5 or 6 years ig Wdyt? Can we see tht drastic change in just one year?
@neetcode1 diminishing returns. seems almost a law of nature at this point
