GPT-5 • Hacker News Board

Re: GPT-5

#881

Post by primaprashant » Fri, Aug 08, 2025, 9:13 AM UTC

created a summary of comments from this thread about 15 hours after it had been posted and had 1983 comments, using gpt-5-high and gemini-2.5-pro using a prompt similar to simonw [1]. Used a Python script [2] that I wrote to generate the summary.

- gpt-5-high summary: https://gist.github.com/primaprashant/1775eb97537362b049d643...

- gemini-2.5-pro summary: https://gist.github.com/primaprashant/4d22df9735a1541263c671...

[1]: https://news.ycombinator.com/item?id=43477622

[2]: https://gist.github.com/primaprashant/f181ed685ae563fd06c49d...

Re: GPT-5

#882

Post by nodesocket » Fri, Aug 08, 2025, 9:23 AM UTC

Why do I have access to GPT-5 on only some of my devices? All logged into my plus account. My iPad ChatGPT shows 5, but my iPhone ChatGPT only allows 4o?

Re: GPT-5

#883

Post by verytrivial » Fri, Aug 08, 2025, 9:26 AM UTC

LeoPanthera wrote: post #392

surround wrote: post #297

GPT-5 knowledge cutoff: Sep 30, 2024 (10 months before release). Compare that to Gemini 2.5 Pro knowledge cutoff: Jan 2025 (3 months before release) Claude Opus 4.1: knowledge cutoff: Mar 2025 (4 months before release) https://platform.openai.com/docs/models/compare https://deepmind.google/models/gemini/pro/ https://docs.anthropic.com/en/docs/about-claude/models/overv...

Gemini does cursory web searches for almost every query, presumably to fill in the gap between the knowledge cutoff and now.

I had 2.5 Flash refuse to summarise a URL that had today's date encoded in it because "That web page is from the future so may not exist yet or may be missing" or something like that. Amusing.

2.5 Pro went ahead and summarized it (but completely ignored a # reference so summarised the wrong section of a multi-topic page, but that's a different problem.)

Re: GPT-5

#884

Post by mustaphah » Fri, Aug 08, 2025, 9:33 AM UTC

primaprashant wrote: post #881

created a summary of comments from this thread about 15 hours after it had been posted and had 1983 comments, using gpt-5-high and gemini-2.5-pro using a prompt similar to simonw [1]. Used a Python script [2] that I wrote to generate the summary. - gpt-5-high summary: https://gist.github.com/primaprashant/1775eb97537362b049d643... - gemini-2.5-pro summary: https://gist.github.com/primaprashant/4d22df9735a1541263c671.…

Why not use the ChatGPT interface instead of the API to save credits? Pass the cookies.

Re: GPT-5

#885

Post by jiggawatts » Fri, Aug 08, 2025, 9:35 AM UTC

primaprashant wrote: post #881

created a summary of comments from this thread about 15 hours after it had been posted and had 1983 comments, using gpt-5-high and gemini-2.5-pro using a prompt similar to simonw [1]. Used a Python script [2] that I wrote to generate the summary. - gpt-5-high summary: https://gist.github.com/primaprashant/1775eb97537362b049d643... - gemini-2.5-pro summary: https://gist.github.com/primaprashant/4d22df9735a1541263c671.…

Wow, the 2.5 Pro summary is far better, it reads like coherent English instead of a list of bullet points.

Re: GPT-5

#886

Post by energy123 » Fri, Aug 08, 2025, 9:55 AM UTC

joshmlewis wrote: post #645

AgentMatrixAI wrote: post #566

I'm not really convinced, the benchmark blunder was really strange but the demos were quite underwhelming, and it appears this was reflected by a huge market correction in the betting markets as to who will have the best AI by end of the year. What excites me now is that Gemini 3.0 or some answer from Google is coming soon and that will be the one I will actually end up using. It seems like the last mover in the LLM…

I am convinced. I've been giving it tasks the past couple hours that Opus 4.1 was failing on and it not only did them but cleaned up the mess Opus made. It's the real deal.

And it's almost 10x cheaper via flex, and in #1 position on lmarena. It's not even close.

Re: GPT-5

#887

Post by andrepd » Fri, Aug 08, 2025, 10:00 AM UTC

jablongo wrote: post #628

highfrequency wrote: post #331

It is frequently suggested that once one of the AI companies reaches an AGI threshold, they will take off ahead of the rest. It's interesting to note that at least so far, the trend has been the opposite: as time goes on and the models get better, the performance of the different company's gets clustered closer together. Right now GPT-5, Claude Opus, Grok 4, Gemini 2.5 Pro all seem quite good across the board (ie the…

It's also worth considering that past some threshold, it may be very difficult for us as users to discern which model is better. I don't think thats what's going on here, but we should be ready for it. For example, if you are an ELO 1000 chess player would you yourself be able to tell if Magnus Carlson or another grandmaster were better by playing them individually? To the extent that our AGI/SI metrics are based on…

I could certainly tell if they played ??-level blunders, which LLMs do all the time.

Re: GPT-5

#888

Post by andrepd » Fri, Aug 08, 2025, 10:02 AM UTC

highfrequency wrote: post #331

It is frequently suggested that once one of the AI companies reaches an AGI threshold, they will take off ahead of the rest. It's interesting to note that at least so far, the trend has been the opposite: as time goes on and the models get better, the performance of the different company's gets clustered closer together. Right now GPT-5, Claude Opus, Grok 4, Gemini 2.5 Pro all seem quite good across the board (ie the…

> once one of the AI companies reaches an AGI threshold

Why is this even an axiom, that this has to happen and it's just a matter of time?

I don't see any credible argument for the path LLM -> AGI, in fact given the slowdown in enhancement rate over the past 3 years of LLMs, despite the unprecedented firehose of trillions of dollars being sunk into them, I think it points to the contrary!

Re: GPT-5

#889

Post by Applejinx » Fri, Aug 08, 2025, 10:09 AM UTC

I am very puzzled that I cannot search for the word 'blueberry' in this HN discussion. Is my browser broken, or is the subject inappropriate to raise in this community?

Re: GPT-5

#890

Post by energy123 » Fri, Aug 08, 2025, 10:13 AM UTC

nacholibrev wrote: post #866

I've tried it in cursor and I didn't like it. The claude-4-sonnet gives me far better results. Also it's a lot slower than Claude and Google models. In general GPT models doesn't work well for me for both coding and general questions.

On livebench.ai, GPT-5 is the best model overall, and the second best for agentic coding. But for the Coding benchmarks, it's ranked like 20th. Quite interesting. I'm finding it exceptional for non-trivial summarization tasks.