quick hack would be to just install codex twice and use one to control the other
CODE
-
Codex versus Claude: Code Generation Performance Comparison
By
–
1. we don't run undisclosed ads
2. curious to see where/ how and in which situations does codex produce inferior code vs claude like any other coding agent, it's worth spending couple minutes ahead of kickstarting a project to build a solid agents md (build it w/ codex) -
Comparing Codex Performance Against Other Coding Agents
By
–
I doubt so, specially find it quite hard to reconcile with how teams at OpenAI have been pushing GPTs to the extreme for exact same tasks mentioned above! That said, I’d be keen to look at examples where other coding agents perform better at tasks which codex doesn’t
-
AsyncGRPO TRL Library: Scaling Inference Training Decoupling
By
–
**Deep content post alert** A technical deep dive for your Sunday morning, somewhere between a short detective story and a tutorial on RLHF We recently added AsyncGRPO in the TRL library to decouple inference and training and scale much faster and harder. As a sanity
-
CLAUDE.md Framework Makes Claude Code 10x More Powerful
By
–
A single 𝗖𝗟𝗔𝗨𝗗𝗘.𝗺𝗱 file that makes Claude Code 10x more powerful! This repo distills Andrej Karpathy's observations on LLM coding pitfalls into four actionable principles. The problem Karpathy identified: LLMs make wrong assumptions without asking. They overcomplicate
-
Critical Software Vulnerabilities Found in San Francisco Tech
By
–
No the real danger is a hacker I met with last night found critical vulnerabilities in most San Francisco-made software.
-
Solving Data Analytics Problems with Python
By
–
How Do You Solve #DataAnalytics Problem?
by @Python_Dv #DataScience #BigData -
10 Tools to Slash Claude API Token Costs
By
–
STOP BURNING YOUR TOKENS! If you use Claude Code, you are probably wasting 80% of your context window. I found 10 ace tools that will completely rescue your API bill. 1. Caveman Claude
– Literally makes Claude talk like a caveman
– Slashes 75% of output tokens with zero -
Fill-in-the-middle completion used by old code models like GitHub Copilot
By
–
Fill-in-the-middle completion was the closest variant of this idea that was ever widely used AFAIK. Old code models (e.g. the original GitHub Copilot) depended on that.