161–170 of 194 posts
This was tweeted about when it happened, with some explanation from Tibo here: https://x.com/thsottiaux/status/2076543065045795309
How can the overall trajectory length be the same across reasoning efforts? I don't see how this is possible even if reasoning is not included in the trajectory length calculation.
Big context windows made everyone lazy about deciding what actually deserves to be in there... compaction makes it worse, it's lossy compression applied to everything at once so the details you actually needed probably get lost. I always thought that retransmitting the full chat every turn is a foundamental problem... in my experience with conversational agents I clear the context every turn and reinject all the rele…
This is a shame, and hopefully it reverts soon. I switched from Claude to Codex just before the Fable release because Claude doesn't follow instructions and goes off the ranch too frequently. I really don't want to have to switch back to what I consider to be a fundamentally worse model.
Big context windows made everyone lazy about deciding what actually deserves to be in there... compaction makes it worse, it's lossy compression applied to everything at once so the details you actually needed probably get lost. I always thought that retransmitting the full chat every turn is a foundamental problem... in my experience with conversational agents I clear the context every turn and reinject all the rele…
So did many NLP experts who worked on LSTM, GRU, etc. But transformers empirically won out and here we are.
It will be interesting to see if future model architectures need to take this into account again.
If humans are any indication, one thing we're still missing is efficient movement of information from short term to long term memory. Fine tuning in principle does something akin to this, but not efficiently.
Earlier quoted context omitted.
Claude does the same? It asks you to create new chats (Claude Design) and Claude Code nudges you to start over too, I think `/compact` is optional though?
Claude does the same when you hit the context limit, but the difference is that claude's context limit is 1m and Codex's is 272k now. Claude lets you get a lot more done before you have to worry about the start getting compacted away.
I know a lot of people like to say that compaction makes this moot, but the level of detail you lose across compaction is wildly too much for most things that I do, unfortunately. Perhaps if your plans don't have as much detail, or if you're not, for example, having a discussion with a lot of nitty-gritty then it's fine? The lack of long context is the main reason that I still end up using Anthropic. The worst is whe…
Earlier quoted context omitted.
Note that GPT-5.6-Sol is ~2x as token efficient as Opus/Fable, so a 258K max is equivalent to ~516K on Claude models. You also may just be fooling yourself, as context rot is still a problem[1][2] and there's evidence compaction can be equivalent or better than longer context windows for agentic tasks[3]. Obviously it would be best if the models could reason over 1M context like they do over 256K context, but they ca…
The improved token efficiency would only account for the output and reasoning tokens occupying context, right? What about the user messages and other context like file contents? I’m not sure how these measure up in a long conversation
Earlier quoted context omitted.
[flagged]
I did find a strategy, thank you very much, and it was easy: npm uninstall -g @openai/codex && curl -fsSL https://claude.ai/install.sh | bash This was the final straw that got me to downgrade my OpenAI subscription and move the bulk of my spending to Anthropic and Claude code. OpenAI devs, if you're reading, this is how you lost a customer.
^ this is literally what the op did. Don’t flag facts.