"Muse Gadgets" turns AI hardware into an open-source DIY project
"Muse Gadgets" turns AI hardware into an open-source DIY project Meta has announced Muse Gadgets, an open-source project that lets hobbyists build their own AI hardware. It includes firmware for ESP32 boards and a Linux SDK to connect homemade devices to Meta's AI agent Muse.

- Meta has announced Muse Gadgets, an open-source project that lets hobbyists build their own AI hardware.
- It includes firmware for ESP32 boards and a Linux SDK to connect homemade devices to Meta's AI agent Muse.
- The code is Apache 2.0 licensed, with a Discord community discussing the project.
Meta has announced Muse Gadgets, an open-source project that lets hobbyists build their own AI hardware. It includes firmware for ESP32 boards and a Linux SDK to connect homemade devices to Meta's AI agent Muse. The code is Apache 2.0 licensed, with a Discord community discussing the project. Meta also built its own device. The "Muse Home Link" is a small USB-C gadget that connects Muse to a home network, letting it control TVs, speakers, and anything else with an HTTPS interface. According to Meta, 5,000 units have been produced and should ship within weeks, free for Muse subscribers while supplies last. Muse Home Link connects Meta's AI agent to smart home devices. | Image: Meta The open-source approach likely serves a second purpose, too. By watching what the community builds, Meta can learn which form factors people actually want, since nobody has settled on what AI hardware should look like. Apple and OpenAI are both working on their own devices, while Meta's Ray-Ban smart glasses are arguably the most successful AI gadget on the market so far.
Sources
Related stories

Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions
Google froze its open source bug bounty program due to a significant rise in AI submissions | TechCrunch Last day to exhibit your breakthrough to 10,000+ tech leaders at Disrupt is on Oct 2 . Book Exhibit Table Now.

NASA and IBM's open source lunar model turns 17 years of orbiter data into a foundation for lunar science
NASA and IBM's open source lunar model turns 17 years of orbiter data into a foundation for lunar science The NASA-IBM Lunar Foundation Model makes decades of lunar observation data usable for machine learning. It's especially strong at predicting ice deposits at the poles and detecting craters.

FlashML Runs MiniMax H3 Video AI on 8 GB Consumer GPUs
Takeaways − FlashML-org released FreeVideo , a local inference engine for MiniMax H3 video generation. Runs in 8 GB VRAM and 16 GB RAM via aggressive weight offloading and streaming.

The Agent Said It Was Done. The Database Disagreed.
The Agent Said It Was Done. The Database Disagreed. The Agent Said It Was Done. The Database Disagreed. Microsoft ThinkingBox grades AI agents on the records they leave behind, not the sentences they generate, and then asks whether they can do it twenty times in a row.