← Back to context

Comment by gptfiveslow

17 hours ago

GPT5 is HELLISHLY slow. That's all there is to it.

It loves doing a whole bunch of reasoning steps and prolaim how mucf of a very good job it did clearing up its own todo steps and all that mumbo jumbo, but at the end of the day, I only asked it a small piece of information about nginx try_files that even GPT3 could answer instantly.

Maybe before you make reasoning models that go on funny little sidequests wher they multiply numbers by 0 a couple of times, make it so its good at identfying the length of a task. ntil then, I'll ask little bro and advance only if necessity arrives. And if it ends up gathering dust, well... yeah.

This. Speed determines whether I (like to) use a piece of software.

Imagine waiting for a minute until Google spits out the first 10 results.

My prediction: All AI models of the future will give an immediate result, with more and more innovation in mechanisms and UX to drill down further on request.

Edit: After reading my reply I realize that this is also true for interactions with other people. I like interacting with people who give me a 1 sentence response to my question, and only start elaborating and going on tangents and down rabbit holes upon request.

  • > All AI models of the future will give an immediate result, with more and more innovation in mechanisms and UX to drill down further on request.

    I doubt it. In fact I would predict the speed/detail trade-off continues to diverge.

  • > Imagine waiting for a minute until Google spits out the first 10 results.

    what if the instantaneous responses make you waste 10 min realizing they were not what you searched for?

    • I understand your point, but I still prefer instantaneous responses.

      Only when the immediate answers become completely useless will I want to look into slower alternatives.

      But first "show me what you've got so far", and let me decide whether it's good enough or not.

      2 replies →

> It loves doing a whole bunch of reasoning steps

If you are talking about local models, you can switch that off. The reasoning is a common technique now to improve the accuracy of the output where the question is more complex.

The article(§) talks about going from Sonnet 4.5 back to Sonnet 4.0.

(§) You know that it's a hyperlink, do you? /s