Text to speech generation. 3/10.The good. Passable quality, easy to integrate, and cheap tools. The bad. OpenAI’s TTS felt abandoned. No way to control speed, limited number of voices, and no great way to refine the output with pauses.

Image generation. 4/10.The good. Detailed prompt engineering goes far. The bad. a) Challenging to create multiple coherent images b) Lack of negation in the prompt, you can’t say ‘don’t do this’.

Design tools. 2/10.The good. Cursor chose reasonable defaults and built-in integrations with a design library made getting started a breeze. The bad. Design tools felt immature, with generic designs and limited control. There is so much potential for someone to get this right.

Coding tools (cont).The bad: Required babysitting to do the right things. a) Lacks context of the codebase and made redundant choices that needed refactoring. b) Overly aggressive in making changes beyond what it was asked to do caused regressions.

Building as a team of one- The good and the bad. Coding tools. 7/10.The good: Coding felt like pure magic. Telling Cursor what I wanted and seeing it write code felt like I had superpowers. It wrote code, partnered on debugging, and installed dependencies.

How did I build it?. Chat GPT to draft product requirements, fed into GammaAI to create designs, V0 to generate front end, and imported to Cursor for building out the full app. Firebase powers the backend and auth, Stripe powers payments, and OpenAI for generative AI.

What did I build?.An app that enables people to generate and listen to their own bedtime stories with AI. It accepts payments, sends notifications when stories are ready, and enables people to save, like, and share the stories they create.

We will see an explosion of creativity as ideas are brought to life by people who couldn't build software before. Not only will there be more ideas, but we will see the quality of products improve as iterating becomes easier.

Inspired by @clairevo Figma Config talk, my goal was to see how far I could get as a team of one. I took the bedtime stories my daughter and I create for fun with ChatGPT and made it into a full product. My first reaction as a technical person but not a software engineer?

Building products with AI is completely wild. In 10 days, I built a storytelling app without writing a single line of code. Everything on has been created with AI. The code, the designs, and the content is all generated.

I am obviously biased on Apple vs Patreon discussion. But @benthompson proposal of Apple collecting 30% on Patreon’s earnings, rather than the creators earnings, really feels like an elegant solution that rewards each platform for the value it creates.

When we look back, the biggest example of monopolistic harm from Apple would be the way it held back the wearable ecosystem by limiting connectivity API support for non-Apple devices.

RT @vkhosla: Hard for me to support someone with no values, lies, cheats, rapes, demeans women, hates immigrants like me. He may cut my ta….

Hey @Waymo , how can I teach the car to say my name properly?.

I guess sports betting is a start.

Hamilton move to Ferrari in @F1 is going to create ripples in the order. Mercedes would want @LandoNorris but he isn’t going to move. Next bet is someone with experience to make their car competitive with willingness to take the risk. My money is on @alo_oficial.

Twitter not getting into predictions markets is a missed monetization opportunity. A deeper integration between @Kalshi and Trending topics would change the dynamic of how news gets consumed.

RT @stratechery: Regretful Accelerationism. The Internet removed constraints from the analog world, and AI is finishing the job. That this….

RT @arampell: @zackkanter “People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a….

Insightful explanation of how Chat GPT works.