The leap from GPT-4 to GPT-5.5 has been astounding in my opinion. There is no way GPT-4 could run a coding agent harness like Codex at even a fraction of the quality that GPT-5.5 does.
I don’t think that’s exactly indicative of GPT-5.5 being an astoundingly more intelligent model, however. An alternate interpretation is that GPT-5.5 was trained on tool usage/harness patterns and has been optimized for this use case.
I remember that even when GPT-4 was king, the Gorilla paper showed that Llama 7B could be fine-tuned to outperform GPT-4 on tool calling.
On domains that don’t involve agentic tool calling*, I haven’t found the frontier to have advanced that much.
Edit: I should broaden this to domains that naturally lend themselves to RLVR training. Models are drastically better at math now.
None of this matters in the product: it either is capable of agentic loop workflows or it isn’t. A 10% improvement in probability of single task success makes or breaks the use case.
Doesnt seem very testable of a hypothesis. As a company thats trying to get things done, doesnt it make sense to have someone help you prioritize by telling you when some decisions have long term costs?
The context this discussion seems to miss is the push to do more with less humans. Im not dictating the dev workflows. My bosses are. My judgment is as good as the amount of time i have to spend. Even if Im the best engineer, giving me a 2 min budget to review 10k lines wont go well. But it does not matter to my boss's okrs
less is more, I just want the plain old search experience I've had for 25 years so I stated paying for Kagi around a year ago. (Before that I used DuckDuckGo which still seems fine. They seem to also offer AI now, but it's not obnoxious like on Google.)
I pay for ChatGPT and use it as a search engine. It gives exactly what I want and nothing more. Just links. And it has much higher accuracy than traditional search. It doesn’t even show ads. This is what a good search engine is!
Are we supposed to conveniently ignore the mountains of money currently being shoveled into building new data centers as quickly as possible without any regard for negative externalities?
