You can find the proof here: cdn.openai.com/pdf/04d1d1e4-… And the full text of the prompt used to produce the proof here: cdn.openai.com/pdf/04d1d1e4-…

I’ve had very good results running autoresearch with local qwen 3.6 26b model as long as I had a simple vibed pi “advisor” extension that allowed it to periodically ask GPT 5.5 for ideas. I think this direction has a lot of merit.

GPT 5.5 + Codex with /goal the first setup that turned my work 100% administrative, and I'm getting good at it. There are two outcomes: → A: my plan is good → GPT nails it → B: my plan is bad → GPT tries anyway → it fails but minimizes it And this is the catch: once you're used to detecting the apologetic "everything is fine" tone of case B, you can just revert and pivot to another idea. In particular, it seems effective to ask: "Do you honestly think this is production ready?" If there is anything smelly, GPT 5.5 will say so, allowing you to evolve towards a better design. Concrete example: > Asked it to implement an in-kernel GC for Bend2. > It did exactly I asked, and declared success. > Code grew too much, so I suspected and asked: > "Would you honestly merge that?" > It then replied in the lines of: > "Ehh... actually, this kinda sucks due to " I then investigated and, yes, my idea was bad. Doing GC at that point requires us to track roots coming from many places, which is a massive, fragile endeavor. It is just a bad approach. So, I reverted it, chatted a bit more, then choose another path: use linearity information (which we already have) to free pattern-matched constructors, with small freelists. A few minutes later, it came back with a working implementation that completely solved the issue. I asked again: > "Would you honestly merge that?" To which it replied (verbatim): > Yes, I’m confident enough to call it production-ready for merge, subject to normal review. Not “impossible to break,” but the specific requested invariant is implemented, audited, tested, and bench-checked. When GPT uses that kind of tone, it is very, very likely that whatever it was done is actually solid, robust and trustworthy, and it knows it. I still inspected the code manually and, indeed, it is solid. Doesn't mean fail-proof, but certainly *much* better than before, and mergeable. That's what I like the most about this model: if your plan is solid, it will almost certainly succeed at implementing it. It only fails if the plan itself is poor, which will cause it to desperately attempt to fit a circle in a square, fail, and be too embarrassed to admit. If you master how to detect which case happened, you can get it to produce a lot of constructive, not destructive, outputs...

this guy just cracked how to get 10x more out of Codex I tested his method and it completely changes how the agent works for you the mistake everyone makes: using Codex like a chatbot. prompt, wait, clarify, re-prompt, repeat here's how it actually works: 1. write one detailed task doc with every requirement, context, and the final vision 2. fire one command: "accomplish this and test until done" 3. walk away that's it. no babysitting. no "can you clarify" loops. no rate limits it runs GPT-5.5 for hours straight in the background while you do something else think of it like this: ChatGPT = a coworker you have to micromanage every 5 minutes Codex = a contractor you hand a detailed brief and come back to a finished product the key: front-load your thinking into the doc. the better the brief, the longer it runs without interrupting you $200/mo with no daily cap while most AI tools burn your entire allowance in 3 prompts this is probably the most underrated setup in AI right now study this.

/goal + GPT 5.5 is amazing. I can now plan really extensive refactors with e2e tests and it just works.