AI-powered "vibe hacking" reaches TVs, cameras and synths
Users are increasingly employing large language models such as Claude Code to reprogram their devices, including TVs, cameras, music gear and game consoles. Manufacturers are split between resisting the trend and embracing it.
A growing number of people are using large language models (LLMs), including coding tools such as Claude Code and Codex, to change the software on their own devices. The practice is called "vibe hacking": a user gives plain-language prompts to an AI assistant and gets results that once required experienced programmers. According to a column in the tech publication The Verge, this is fuelling a market for developer-friendly gadgets, but it also brings new risks.
From TVs to cameras
Gathin T., a developer in India, was unhappy with the ad-filled interface on his Xiaomi TV, which runs Android TV. Using Claude Code, he redesigned the TV in about an hour and a half: he removed ads, simplified the interface and made it easier to reach his PlayStation console. After a post on X went viral, he released a tool called CleanYourTV, which became one of the most popular repositories on GitHub. This approach works this easily mainly on Android-based TVs.
Camera enthusiasts are using AI to extend camera firmware with new recording features, higher resolutions and new colour profiles. Musicians are modifying cheap synthesizers; the maker of the M-Vave FM-1 even released open-source firmware. Gamers have used similar methods, for example to play PlayStation 2 discs on a PlayStation 5.
Manufacturers respond, risks grow
The games industry is trying to defend itself. Luke Dicken, a former head of artificial intelligence at Take-Two, predicts more invasive DRM (copy protection) and a stronger push towards streaming. Users, meanwhile, risk voiding warranties and bricking their devices.
A safer route is offered by ESP32 modules, cheap programmable hardware components. Meta already runs a Muse Gadgets programme. Gathin advises making backups and experimenting on a separate computer rather than on devices that matter.

