Actions + generative UI is the next major evolution of AI apps. And Vercel just unlocked this for *everybody* with their AI SDK 3.0 update. We’re about to see an explosion of insane AI experiences. This AI stock bot shows where we’re headed.

By the way… This isn’t the future of AI chat. This is the future of computers.

This style of interaction is going to require people to know what they want and how to communicate. V1: don’t make me think (apps) V2: say what you want (now) V3: it always knows (future?)

Memory gets us to 3 imo. We’ll need an architecture leap to get there. (I swear if I get a bunch of replies from RAG people that we’ve solved memory just no [yes it has its place but it’s not *memory*])

Every AI chat inside an app will have this kind of generative UI and actions soon. We have had this since June of last year in Heights AI. Of course, you can ask how to do something, but it is SO much better when you can just ask the AI to do it.

💯 not just AI Apps. All static web apps will be rebuilt like this or replaced.

this is just adding latency to accomplish something a brokerage app would do 10000x better lol

Stone Age thinking

I’ve been bearish of all the wearable AI tech (eg., humane pin) coming out because I don’t believe people will adopt a strictly voice UI yet. That said, I do see this massively increasing adoption of Siri and assistants. From there, I could see a purely voice UI cross the chasm.

Voice won’t take the place of text. Voice will be useful as its own thing, but it won’t replace what the text interface is

Except that the tech is super super limited. Basically impossible to include dynamic information ATM. Working on this for a hackathon this weekend.

They solve this