I would like to use something like Wiz for sure. Been thinking of building parts of it, ticketing system that has implement fix and just does the work.
As your post suggests - there is great value in using both. I have only recently come to that perspective. Just ran CODEX against my project and got some good inputs for improvement. I'm listening ...
Yeah - that is something I "discovered" as well. And don't get me wrong, for agent work I'd prefer Claude, but for improving my AI Agent - Codex is so good. Less back-and-forth.
Good analysis and write up. I have also covered it https://globalrule.substack.com/p/cursor-vs-github-copilot-in-2026
Oh I already have :D
Thanks! And I will read yours as well :)
I would like to use something like Wiz for sure. Been thinking of building parts of it, ticketing system that has implement fix and just does the work.
I am working on open sourcing it. Needs time as I started for my own needs only : D
Best way to build
But not easy, when you started WITHOUT such intention. Hence - Codex will have some tokens to burn xD
Oooo, interesting development with Peter joining OpenAI.
It is! I hope that doesn’t mean OpenClaw will go “closed” :/
Pawel what hardware are you running, if you don't mind me asking. I'm looking at upgrading.
Naswered on your resack - Macbook Pro 16 with M1 Pro + 16 GB RAM. Enough for 90% of things. That’s why I am thinking about upgrading as well :)
Oh, I'd love to know more about Wiz! How do you run it, where, etc.?
Amazing breakdown in this article though!
Thanks!
Wiz is on Claude Code with a lot of custom things...a long story! I think open-sourcing would be the best move :D
I like the idea that OpenAI is not dead yet.
There is a lot of people that want to see its fall, maybe because of Sam Altman's personality.
Me, I personally prefer less cheerleading and more business.
Far from dead + today’s info about OpenClaw joining OpenAI - interesting!
BUT - yeah, I just stick to Claude for now. It’s good!
As your post suggests - there is great value in using both. I have only recently come to that perspective. Just ran CODEX against my project and got some good inputs for improvement. I'm listening ...
Yeah - that is something I "discovered" as well. And don't get me wrong, for agent work I'd prefer Claude, but for improving my AI Agent - Codex is so good. Less back-and-forth.