Fuwn

When GPT-5.4 was initially released, I was able to run 10+ hour-long autonomous sessions, but that hasn't been the case recently. I saw someone post about how GPT-5.3-Codex is a code addict that cares about nothing else. They were right. GPT-5.3-Codex will run til heat death.

I thought I was tight with Kimi K2.5, but it is definitely not following my AGENTS.md guidance today.

Has anyone else noticed that ChatGPT no longer drops the clickbait “If you want, I can also show you …” snippets in its responses? Can’t say I miss it.

Genuinely, what is it saying ...

My subagents are subagenting.

GPT-5.4 is incredibly good at SwiftUI, which shouldn't surprise me. Fundamentally, there are many reasons why it should perform well with it, but it's still a delight to see it resolve any bug I throw at it in a large SwiftUI application in almost no time.

Using GNU Make feels gross unless I want to feel extra C.

WTF is Codex on about? All I did was try out Garry Tan's gstack. It's done this multiple times now. I know gstack is meant for Claude Code, but this is bizarre.

@btibor91 Finally, a hard update. I’ve given ChatGPT a custom instruction to stop the clickbait follow-ups, but it has repeatedly ignored it.

GPT-5.3 Instant in ChatGPT is getting an update that improves follow-up tone and reduces teaser-style phrasing in responses

GLM-5 makes so many character set and punctuation mistakes in its responses that I’m seriously unsure how anyone would use it for serious software. It seems fine for personal use, though.

I'm still running low on GPT-5.4 usage in Codex, so I've been postponing actual work to run harness experiments using Codex-5.3-Spark on Codex CLI, Pi, OpenCode, and ForgeCode. Codex CLI and Pi are considerably superior to the competition, but Codex CLI remains the best.

@tmuxvim Every. Single. Time. The worst part is I’ve fallen for it once or twice.

has anyone else noticed that GPT-5.4 often ends its responses with like, clickbait? it often promise to reveal "the one surprising X that will do Y" or something like that

The real challenge of side projects isn’t programming. It’s finding something people actually care about.

@temidaradev github.com/acsandmann/rift

aerospace or yabai? (or something better)

I could use a reset right about now. I would have already bottomed out if I hadn’t cut myself off a few hours ago at 15% remaining. 😬

Can you tell when I started using GPT-5.4?

Managing parallel agents is quite enjoyable. Perhaps I could pursue a future as a manager at a large tech firm, leading a team of AI agents and having my subordinates review every line of code.

Well, well, well. How the turntables.

@beetbox_games @libanan__2 @LukasHozda It has to be Racket. racket-lang.org

@libanan__2 @LukasHozda react?

I wonder which programming language this budding Haskeller is coming from

@skydotcs It’s fine for mathematics, not perfect. I’ve used it a few times for particularly tricky problems since it does a lot of self-checking and contradiction verification, but it’s extremely slow, and you can get the same answers from 5.4 Thinking for most problems.

just realised i haven’t used this before lol what’s it good for?