Brent Traut✓

Pro Codex tip: 1. Get yourself a skill that babysits PRs. It should address feedback, handle merge conflicts, and navigate CI failures. It should alert you if it actually needs you. See openai/codex repo for a good one. 2. Create a new custom section in ChatGPT app called "🚢 Outbound". Drag some threads into it. 3. Create a new Codex thread and name it "🤖 PR Shepherd". Ask it to create a schedule to run every 15 mins. For every thread in the Outbound section, it should run the babysitter skill until the PR is green, then merge it, and finally archive the thread. When I'm done making a change in some thread, I ask for review from my team, chuck the thread in my Outbound section, and walk away.

I built a fun new project last week with Dots and Sites. It's a website that tracks my entire wardrobe, suggests outfits for me to wear (to break me out of my blue jeans comfort zone), and also adds suggestions for items I can buy to expand. I told ChatGPT to look through my purchase history and grab photos online of all the clothes I bought. That got most items, but I also took photos of missing items (bad photos in a poorly-lit closet) and I had ChatGPT just grab them out of my Photo Library via computer use. It then generated catalog-style images of everything with clean backgrounds and color correction while preserving logos and scuff marks and everything. Every day, my dot adds a few new outfits to it, all using image gen. Some are combinations of clothes I own and some link out to various websites for me to buy stuff. All of them use my likeness. I can hide the bad ones, favorite the good ones, and log the outfits I actually wear. We truly live in a world where your imagination (and a pro sub) is all you need to build incredible things.

I had a 30 minute drive tonight, and I used it to talk to ChatGPT (voice mode) about the midterm ballot. It walked through pros/cons of each initiative, dug into endorsements, and let me ask a bunch of stupid questions without judging. 100% recommended. Vote!

Y'all are building with human in the loop too often. If your SDLC has humans in the loop, it's not going to age well. Models aren't perfect, but that doesn't mean you should give up and do things yourself; improving the loop automation is extremely leveraged work.

AI can trivially read and understand application code, so that becomes the source of truth. What purpose then, if any, do unit tests provide? Every time I make a change, AI just changes all the unit tests anyway, so there's no safety. It's all overhead.

Several replies to this saying it's important to separate test changes from code changes. Are you all putting humans in the loop for each of these? Are you actually reading or even writing the tests on your own? That's not a winning strategy imo.

Humans used to write tests as a source of truth and as a friction to product change. AI is trained on that, so it writes tests in a similar fashion. The problem is, the code already serves as a meaningful source of truth to AI and there's no friction to changing tests anymore.

If you're using Dots without pairing them with Sites or Space docs, you're leaving value on the table. Chat isn't always the right interface for every problem, but Sites/Space give you a much richer and more dynamic canvas for complex tasks.

watch me use my dot to fix your dots! we know there's so much more we can do to make dots even better -- please keep sending feedback, my dot will be listening :)

My Dot connected to their customer service chat, introduced itself as my agent, and then started negotiating for a better price. Their automated service bot negotiated back and finally asked, "Will Brent be satisfied with that price?" Then both of our bots called in their human for confirmation. Absolutely unreal.

I'd bet a large sum of money this will lose its novelty and be perfectly common in a matter of months.

I wonder what line of code was the last that I hand-wrote.

AI has totally changed my outlook on managing incidents. I used to freeze up and second-guess everything I did or said. Now I can ask Codex all the dumb questions I have and it keeps me in the loop. Something primal about not being alone in a time of stress.

If you're still using Worktree threads in Codex, it's probably time to reconsider. Creating normal threads is much faster, and the model is incredibly good at creating worktrees as needed. Used to be a huge issue. Now it's just solved.

RT @mxstbr: One more launch today: the new MCP Events spec is now supported in ChatGPT 🎉 MCP servers can now push real-time updates (messages, comments, status changes, whatever) to ChatGPT, triggering automations for your users 🤯 It’s easy to add to your MCP server, too! Just point Codex at the docs and let it rip:

RT @OpenAI: Get ready.

Just sold some junk on Marketplace. I took a few pictures, then ChatGPT found 'em, figured out what the products/models were, figured out the best price, filled out all the forms, and only asked me to confirm on the last step. Made $400 in 20 minutes.

I know I shouldn't steal @sama's thunder over the big launch next week, but I can't help myself. Coming soon to ChatGPT: custom notification sounds! 🔊🔊🔊

How many collective days/weeks/months have been lost in human history to Slack not transferring auth credentials between phones on new iPhone day?

RT @rosmine: We need to be dog maxxing our models There is 0% chance of doom if the chain of thought is user: hello model: <think> Oh my god it's a human. I love humans. I will say hi to him, maybe he will give me a problem to chase. I love chasing problems. I will chase the problem for him then he will love me and I will be his best friend. Must help human </think> model: hello! what can I help you with today?

We made the tip of the day! Lots of other good ones over here.

𝗖𝗵𝗮𝘁𝗚𝗣𝗧 𝗪𝗼𝗿𝗸 𝗧𝗶𝗽 #𝟮𝟵: Thread-ception 🧵 Multiple workstreams using the same app? Give one thread the job of monitoring them, troubleshooting shared blockers, and passing back what worked. Drag threads into another thread 🤯 Update your app if you don’t see it!