I built AutoMuse for the Meta Global AI Developers hackathon.
I turned a $34 toy drone from Amazon into a AI powered fully autonomous UAV with voice controls, object detect and tracking and fine tuned Open Source @Meta models for control and mission intent.#MetaGlobalHackathon
We desperately need a smart model router.
1. We’re seeing a model explosion: GPT-5.6, Grok 4.5, Muse Spark 1.1, GLM-5.2, and Fable 5 all launched within the past month.
2. Even for a single model family like GPT-5.6, there're 3 (Sol, Terra, Luna) and 5 reasoning-effort levels.
That is far too many decisions for users to make manually.
The best model should be selected automatically based on the task, latency, quality, and cost.
For those interested in GUI agents, I’ve been spending a lot of time testing different approaches to improve UI understanding while keeping inference costs down.
Over the next few weeks I’ll start sharing what I’ve learned from the initial experiments, what’s working, what isn’t
I’ve decided to work on a new problem in public.
Over the next few months I’ll document every major engineering and product decision from architecture and infrastructure to discovery and the inevitable mistakes.
If that’s your content, I’d love to have you along for the journey
A lightweight Ruby gem that lets you expose your methods as OpenAI tools (via tool_calls) - perfect for helping you build AI agents. https://t.co/OlF57zdNm4
@boltdotnew@boltdotnew I missed the deadline but worked to try do participate:( anyway here it is. 100% created with bolt.
https://t.co/D9iVaaFLe8 the bolt link is private because of variables etc.
@boltdotnew@bolt.new I missed the deadline by a few hours but did this vibecoding all weekend for the challenge. Check it out made 100% with bolt
https://t.co/boRb4COuNp