Jonathan Crawford 👾

Mission -> Strategy -> Tactics In that order.

@OpenAIDevs /hatch is not recognized. /pet only toggles. I've updated desktop app and CLI. Distractions are extra distracting when you can't get it to work.

Customize your Codex pet with /hatch

Pets. Now in Codex. Use /pet to wake your pet.

@adxtyahq Uh, sure seems like VC-required. It’s all over the page.

Claude Startup Program is also OPEN btw > API credits for early-stage startups (up to ~$25K) > Built by Anthropic (Claude) > No VC needed (unlike OpenAI) > Selection based on product + real Claude usage > Actually friendly to bootstrapped founders Apply: anthropic.com/startups

@LaurenceBrem @code My bet is that it’s got different instructions. It seems to be more thorough as if it’s being told to ultrathink and one-shot the solution.

@code Using Codex in Web still feels like it has more reasoning power than in CLI, or within VSCode but don't know if it's running a fundamentally different model?

OpenAI Codex is now integrated directly in @code through the new Agent Sessions view - and can be powered by your GitHub Copilot subscription. Try it out now with VS Code Insiders and a Copilot Pro+ subscription. Happy coding!

Hooray! This week’s top AI coding model has been released! 👏 👏 👏

@ah20im - It constantly thinks it doesn’t have tool access that it has - It often attempts constructing raw API calls rather than use clearly defined MCP servers - Planning mode! - Sub-agents!! - Keyboard shortcut to toggle perms - pwd in prompt box - Faster editing previous messages

What are some features/enhancements you would like to see in Codex?

3. In a 2nd Codex instance, same dir: /dhh-linear-plan-review - Fetches issue from Linear again (Linear MCP) - Scans project - Reads the md (named after dir in ./planning/) - Provides savage feedback in the md file Back to 1st instance: /read-plan-feedback - You know the rest

2. I have a few Codex prompts that make all this a breeze: /dhh-linear-coder - Sends the DHH bit - Pulls a Linear issue based worktree name - Reads the project - Thinks about a plan - Calls Gemini 2.5 Pro & Opus for feedback on its plan - Saves plan in a md file named for issue

Not to fangirl, but I’ve gotten the best #AICoding results by telling one instance of Codex to code like they are @dhh (DRY, expressive, self-documenting, etc.) and another instance to review the work like DHH deciding if a PR should make it to Rails core. Reviewer is merciless.

When I was a kid, I thought this was about the life story of Yankee Doodle #fourthofjuly2025

I was just informed that it's "Bard Knock" (singular). AI wrote that post, and so I cannot be held responsible.

It's a Bard knocks life.

If you don't love me at my early Gemini, you don't deserve me at my o3-Pro.

@zeekay Maybe we're all just a skilled fine-tuning away from being amazing. Maybe I've only been dating people who are very bad at fine-tuning. Honestly, reinforcement fine-tuning, where the reinforcement is snacks, would work on me.

@JonCrawford In the words of Sun Tzu... "only the foolish abandon early tech adopters when greatness is in fine-tuning.”

ChatGPT's brain now has a full body. Massive.

Who are the founders & #AI enthusiasts working on health & wellness products?? #wellnessAI #healthAI #digitalhealth

@ionleu Yes, I’m calling it OpenAI

no cursor. no chatgpt. no claude. can you still build a startup?

It’s also not great at working in MVC frameworks and OOP. It’s constantly having to go through and orient itself to avoid reinventing the wheel in disastrous ways. Frequently this “research” goes on so long, that it ends up just telling me what my application does and stopping because it totally ran out of context and forgot what it was doing. I haven’t determined if the better approach is sending way too many instructions (which can go out of date fast) or having to manually tag 15 files every time I start a new chat.

what should i add to @cursor_ai that would make u 10x more productive? feature requests, tiny nits, anything :)

It would be awesome for there to be a first-order feature to specify the tech stack as part of every prompt (including versions). Then it auto-fetches docs for each. The model would get instructions to check those frequently. Bonus points for summarized changelogs for common frameworks for changes after model training dates. I’m on versions of Rails and tailwind that were released after training dates and it’s a LOT of babysitting, backtracking and tagging “@docs”. It’s also pretty bad at discovering/using modern libs/gems. Just had to tell it not to use a rubygem that hasn’t been updated since 2012. Tech moves fast. It would be nice to figure out a way to keep the toolbelt more current.

OpenAI learned nothing from the Pepsi Challenge and Cola Wars of the 1980s Poorly-contextualized marketing data & product feedback can lead you to make the wrong decision. People prefer a taste of something sweeter in small amounts, but will always end up buying the real thing.

So OpenAI learned nothing from the Pepsi Challenge and Cola Wars of the 1980s? People prefer a taste of something sweeter in small amounts, but will always end buying the real thing.