ICYMI - There are great demos of GPT-6 Astra running 24-7 on GPT TV: I'm gonna be spinning up a laid back, put it on in the background stream, of some live Astra improv. Kicking it off here at the hour. Come hang out!

I built my first native Mac app! It's called Coverage. It connects with whatever you're playing on Spotify and makes effects play on the album cover. This effect is called Sheet. And when I say I, you know of course that I mean Astra and I. I don't know how to make Mac apps. That doesn't matter anymore. I'll make a demo video soon about how it works and how I prompted Astra. But I can't now because I'm on a plane, and that would be weird for my seat mates. So I just asked Codex to add a new feature to Coverage. I had it make a video capturing how rad it looks.

Y’all remember the first time you ever touched a Large Language Model? I vividly do, I’ll never forget it. This was pre-ChatGPT and at Twilio we got early access to this super bizarre API where you gave it some examples and then you would start a sentence and it did this magical thing where it generated what should come next. In the body of a POST request you would send something like this: Q: What is the capital of the France? A: Paris Q: What is the capital of the United States? A: And the response from this magical API would somehow contain the answer. Now if you are like, “yeah big whoop, so what dude?!”, I beg you to remember that this did not exist aside from science fiction. You just happen to be used to this world now. And at one point in your journey, you’ve had a relative whoa moment. My team and I at the time were working on a main stage demo where we were trying to build something interactive that showed off how powerful this could be. We had the wild idea that maybe we could have it generate TwiML, or Twilio Markup Language, a relatively simple tag based declarative language. It kinda worked! So we made a site where our customers were able to write what sort of telephone app that they wanted to build from English and wire up a telephone number. (And if it failed I made it Rickroll them. Demo magicians always have to have an ejector seat button.) I remember watching in disbelief as the requests came in and our application was generating actual working code. eg: "Make me the sound of a whale" and it generates some markup that makes it so when you call the phone it answered with pretty robotic text to speech that said "Cachunk", onomatopoeia for a whale coming up and splashing down. It was creative. I could not believe my eyes. And honestly since that day, I’ve had that feeling. I’ve been chasing it, finding it, and sharing it. This needs to be experienced, touched, and played with. By everyone. This time we are living in is unlike anything we’ve seen before. The abilities that this unlocks for humans to create and build things is untapped. And sure, I know you as a human can also do autocomplete. Like, here, give it a try based on the context I just fed you: Q: Craig now works for A: . . . . . Did you guess @OpenAI ? You're absolutely right! I'm so incredibly excited and honored to be able bring these tools forward and present that feeling of wonder and empowerment to all’y’all daily. Here’s to many more whoa moments! Let’s do this! (obvs the API and tooling are a lot better now, follow along at @OpenAIDevs)

Sometimes the best cure for AI is AI. I built a @flueai Agent on @CloudflareDev to treat slop. Here’s how it works: 0:00 The commercial 0:46 Why slop gets spotted 1:32 Testing Slop Doctor 2:18 When design gets dated 2:39 Website exam 2:46 How it works 3:00 Flue agents 3:51 The diagnosis 5:32 Prescriptions + checkups 6:27 What’s next?

I took advantage when making an educational video about Python Workers to make the motion graphics be in the style of the muse for the language, Monty Python. I used SeeDance the AI generation model and I kinda dig how it turned out. Wondering how y'all feel about it? Want more like this?