The recording of my talk—“The UI That Builds Itself”—from @ReactSummit is now available!
The audience’s reactions and the Qs afterward were… kinda expected 😅: Devs are still hesitant to give so much control to an LLM, especially for the UI their users interact with.
Curious how many of you are willing to delegate UI generation (like, at runtime, not during development) to an LLM?
https://t.co/CDP6W1v2sA
@mrmagan_ Sometimes I hope this would be applied to Video Games. Not sure if you game, but having an optimized tutorial somewhere for all the things in a game kinda sucks - now give everyone a different quest and we're back to exploring and diving into the game's world.
Origin, our code hosting platform, is now live.
It's fast, easy to use, and deeply integrated with Cursor.
Get started by syncing your repos from GitHub.
@GergelyOrosz Isnt the "harness" part free, and why would you build a harness specifically - cuz it's really easy to get this part wrong and end up with sub-optimal tool for your engineers. Although I agree for the rest
I was a bit skeptical about installing agent skills from repositories (e.g. `npx skills add org/repo`) but we just had to move a skill into a new repo and it occurred to me that i can simply inform the agent that it was move 🤞hopefully it would follow the instructions better than a human would...only time would tell.
Back in june, during my “The UI That Builds Itself” talk at @ReactSummit, i gave a subtle sneak peek at the Observability platform that helped me show the audience exactly what was happening between my UI agent and the LLM powering it.
Today we’re launching it on Product Hunt - go check it out!
Progress AI Observability is now live on @ProductHunt! 🎉
AI agents are getting smarter. Understanding what they're doing shouldn't get harder.
Get trace-level visibility into prompts, tools, retrieval, retries & outputs.
🚀 Check it out now: https://t.co/M71x8KiUsE
GenUI (A2UI, or whatever you call an LLM generating user-interface on the fly) has been a favorite exploration topic of mine in the last couple of months, and my talk from React Summit is available online.
Already working on an updated version - stay tuned!
What if the UI assembled itself around what the user actually needs?
@kirchoni on generating dynamic interfaces with LLM calls, without sacrificing reliability.
Watch the talk: https://t.co/ZhAZIeE10o
I’m exploring the mental health effects of AI coding. This is a super quick (blazing fast), anonymous survey.
Please fill out and share so that the responses will be well rounded.
https://t.co/qbLPRRXt7I
GitNation × @ProgressSW Hackathon Winners🎉
After an amazing showcase of innovation & creativity, congrats to the winners!
🥉€2,000 Prize — Stuck stuck
🥈€2,000 Prize — ConfPilot
🥇€6,000 Grand Prize... is 🏆Hallway
Thanks to all for making this hackathon remarkable!
@thekitze When you research this, please write a short blog about it, I will buy it over whatever slop AI SaaS product you have to offer.
jk, love your stuff <3
omg the room really was FULL.
💚thank you, thank you, thank you to everyone for coming to my talk, i promise i’m working on many more interesting stuff!
@rauchg I do agree to an extent that it looks similar to the npm registry, but a self-contained unit of "extensibility" for an LLM requires both text (skill) and structured (mcp/tools) combination. Some ppl call it a `plugin` - thus my question: Does vercel plan a `plugins` registry?