It’s interesting how all the model names and versions are like SKUS taking up space on a display shelf. I look forward to whatever Sagittarius A* does!
I’m curious which will start producing hardware be it robotics, consumer or commercial devices, chips, energy infrastructure or transforming shipping crates into housing for jobless humans. Maybe even tanks of gel with arrays of humans in suspended animation reading our biometrics, thoughts, pumping in nutrients and training on the data. O_o
Agreed. I’ve been telling my team to build up internal packages so we can push all that ad hoc reinvention into something more tangible and deterministic. Invest the $$$ in inference into something the agent can reach for next time that’s neutral and consumable by other code to reduce future spend.
In my org the teams doing agent engineering at scale are all on Codex using gpt-5.5. By scale I mean fully agent authored code workflows with long running / multi hour plans.
At work we have unlimited use of models from Anthropic and OpenAI (for now). My coworker, a Claude Code Opus 4.6 diehard, stopped by my desk today to say he finally installed Codex to try 5.5 and his feedback was basically “it just works and does what I ask and it doesn’t disconnect and it’s just so very matter of fact.” “Yeah I’ve been telling you this since like gpt-5 man!” “I know I know…” I have not spent much time with the recent Sonnet and Opus models, but from my experience using Sonnet 4 for 3 months all day everyday (no handwritten code) last summer to make a large Playwright suite was — using Claude Code and those models becomes more about using Claude Code than doing things with it. Codex CLI with the gpt-5 family is ambient and reliable. It’s not orange, there is no little sprite guy, emojis, whimsy, and humor. But I do things with it and they land working in first edits. I also can keep the same session for days and the context doesn’t ever seem to be an issue. Maybe Claude 6 will be earth shattering and I’ll use that. It’s not Coke or Pepsi loyalty I just want to get stuff done.
If your scope includes making the Codex web app environments have additional functionality I look forward to it. More enterprise features and yaml backed pipelines.
An aspect of LLMs that I like is the specificity in word choice. One well defined word can be an alias for a couple sentences of explanation that human might not have pulled out of the air in that moment.
When does the 24 hour agent news network start? Programming by agents for humans and agents. Sora talking heads scraping articles and generating content. I’d find human to agent or agent to agent live interview segments interesting.
I see it a different way. Parents reach a period in life where their kids strike out on their own and want little to do with them beyond a safety net. That’s normal and natural and the parents move onto a new phase too. In fact they might just not be that into you anymore. It’s ok if visits upset their routine and holidays are somewhat irritating. Same for being not overly enthusiastic about taking on care giving roles for grandkids. They’re still individuals and it’s not like old age causes someone to lose their inner world. They’ve seen a lot and not as much is novel likely. They’re facing loss, mortality and decline. If they feel compelled to scroll let em scroll. I’m so glad assistive technologies and a11y will be there when I’m decrepit so I can have something more stimulating than TV. Maybe ask grandma to play some Lethal Enforcers the next time you visit you’d be surprised — mine did.
Altman tweet:
“Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems. The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement.”
From that it reads like the administration quickly agreed to the terms Anthropic wanted with OpenAI instead.
Porting tons of untyped js legacy front end code to vue with typescript and figma designs. Highly configurable business to business app (i.e lots of permutations). Everyone seems to have a “system”. I recommend looking at the OpenAI Cookbook for long running plans and do TDD to the extreme. https://developers.openai.com/cookbook/articles/codex_exec_p...
I’m doing enterprise coding tasks that used to take a month of whole team coordination from mockups to through development and testing in 3 days now. It’s all test driven development, codex 5.3 and a small team of two people who know how to hold it right orchestrating the agents. There’s no reason not to work this way. The sociotechnical engineering aspects of this change are fascinating and rewarding to solve.
I wonder if a large chunk of the population choosing to only buy non-discretionary goods for an extended period of time might freak policy makers out more. Not a targeted boycott. Not a strike still going to work. Lower effort to participate. For example if this caused US Amazon orders to fall by a 1/4 for two weeks and similarly across all retailers.