@Rasmic heโs openly admitting that Codex has traditionally played catch up with most features. this is an objective example where Claude Code has copied a feature that Codex released first. looks like humility to me, and heโs having fun with it
@btbytes@badlogicgames@trq212@antirez I wouldnโt say he just gets everything from an agent as an html slide show, but it sounds like this video from @lucasmeijer https://t.co/J3KwHd0lz8
the html light bulb turned on for me after watching this
when i voice prompt, i yap for 10 minutes straight and change my mind 3 times in the middle of the yap, and send it without reading
yap enough tokens for the picture to be complete, it understands well when you change your mind in the middle. ai is smarter than you think
@sama creative writing and ability to learn better from personal writing examples and better adhere to stylistic choices, tone, etc
thatโs the main thing Iโve seen somewhat lacking from my personal use cases and workflows - GPT 5.5 is otherwise a banger for everything else
i'm not really a big "one secret that makes ai good" guy
but i will say it does so much better when your prompts are longer
and it's so much easier to produce long prompts with voice, completely changed things for me
@hadesdevs@theo Theo said that the person who won ran it through GPT 5.4 on xHigh in three different harnesses, and the OpenCode harness solved it in 5 minutes