68% of developers say AI improves the quality of their work, but only 40% of designers say the same.

Teams that said the process of creating a generative AI app is "completely" or "significantly" different from non-AI products were more likely to have been successful than other teams.

Small and large companies are more committed to AI initiatives than medium size companies 🤔 (also note that this is growing quickly at small companies and stagnant at large companies)

Teams are still figuring out how to apply AI. The most commonly cited goals of AI projects are vague: "Enhance customer experience" (41%) and "Experiment with AI" (35%). "Drive revenue growth" is a distant third at 9%.

It also illustrates all the space short of building a full fledged web app. An individual should be able bring an idea to life and share it with others without dealing with all the common infrastructure ... essentially what Blogger did for blogs, but for web apps.

As he points out, the crux of the challenge is these tools are mostly built for the current (previous?) era of software development. By being more opinionated about things like UI frameworks, hosting, etc and taking some choice away from the user they actually become more useful.

I think a UI model with a lot of potential is the AI agent as a document collaborator...making suggestions that the user can approve, reject, or comment on.

This pairs nicely with Adam Aleksic's 'An Elegy for the Pencil.' Both essays suggest how much untapped potential lies in digital interfaces compared to the multi-sensory richness of physical tools. adamaleksic.substack.com/p/an-elegy-f...