Business · Economy 51ZDNET’s key takeaways New AI tools make modifying gadgets and device firmware easier. People use them to get more from their smartphones and smart TVs. Easy, useful modifications can pose security and reliability risks.
The reverse engineering of popular software, such as Adobe Premiere and PowerPoint , has become popular lately — and now, hardware is being reverse-engineered with the AI-powered Reverse Engineer Anything (REA) . This open-source project has exploded by enabling hobbyists to reverse-engineer pretty much anything, such as your obsolete smartphones , smart TVs , and smart picture frames . Even with this tool, you still need some programming chops.
But here’s the important bit: Anyone can use AI agents and borrow technically skilled users’ work to easily modify their own hardware. Also: GridCraft is the open-source Excel clone for Linux I’ve been wanting More from ZDNET The concept itself is far from new. In 2013, the alternative Android firmware CyanogenMod, which brought newer versions of Android to obsolete smartphones , enjoyed several years of popularity before dying in 2016 .
The project technically still lives on as the open-source project LineageOS . Other alternative Android operating systems, such as the Free Software Foundation’s (FSF) LibrePhone and /e/ OS , are alive and well. All of those were made by expert open-source developers, while this new generation of alternative hardware firmware is largely the product of vibe coders.
This new firmware brings new tricks to old hardware. For example, it can remove smart TV advertising, replace cluttered interfaces, and restore functionality to abandoned devices. This emerging hacking movement empowers old-school device-hacking techniques with AI tools that investigate software, generate replacements, and automate changes.
Also: After the vibe-coding rush comes the debugging hangover The activity has gathered attention alongside REA, an open-source project that connects AI agents to software-analysis tools. But the name has become an umbrella for many hardware reverse-engineering products that have nothing to do with the REA project. Several popular device modifications were built using general-purpose coding assistants , such as Claude Code and ChatGPT Codex .
A smart TV makeover is one viral example One of the most visible examples began when a user got annoyed that their television showed ads before they could reach Netflix. On Oct. 17, a vibe coder posting as Gathin described using Anthropic’s Claude Code to modify his parents’ Xiaomi TV. “Gave Claude Code the model number, it found the TV on Wi-Fi, removed 25 Xiaomi packages, and wrote a new launcher.” The project he used was CleanYourTV .
According to project documentation, it backs up package lists and settings and then disables selected manufacturer packages. It doesn’t uninstall them. It also installs a replacement home-screen launcher and includes a restore script to reverse the changes.
Also: I tried the free Microsoft Word clone on Linux – and it’s my new default That is a significant customization, but it’s no wholesale replacement of the television’s operating system. It uses Android Debug Bridge (ADB) to work through debugging access that the owner enables and authorizes. Another project, dumbtv , pursues a similar goal.
It’s been tested on Fire TVs, Android TVs, and Google TV platforms . Both projects still face device-specific limitations. Simplifying the home screen does not mean every advertisement or tracking mechanism has disappeared.
Beyond TVs The same impulse is appearing across other connected products. Also: I tried the free Acrobat clone – and now have a pro PDF editor In an Oct. 20 report, 404 Media’s Jason Koebler described people turning smart picture frames into calendars, removing advertising from baby monitors, and installing new software on abandoned Lululemon Mirror exercise displays . Those are reported individual projects, not evidence of a universal method for restoring every affected device.
What REA brings to device research REA provides a clearer link between AI coding agents and established reverse-engineering tools. Its documentation describes a Model Context Protocol (MCP) interface, which agents can use to inspect software and receive findings with supporting evidence. Native-binary analysis then uses Ghidra , Hopper , or IDA , a software reverse-engineering (SRE) framework, to allow the agent to ask follow-up questions, explain behavior, or write and test an implementation.
For device researchers, the project’s documented capabilities extend beyond desktop applications. Its firmware workflow uses Linux’s Binwalk or Unblob to dive into the code, extract its content, and then hand the material off for analysis. Other integrations support Android device inspection through ADB, recorded-packet analysis through TShark , and decoding recorded logic signals through sigrok-cli .
All of these may sound scary if you’re not a developer. Also: Someone made a free Photoshop clone – so I had to try it on Linux They’re all interfaces to specialized tools. They won’t enable AI to automatically unlock any device.
The documentation specifies external dependencies and target-specific requirements. Its logic-signal workflow, for example, analyzes supplied recordings rather than promising autonomous access to a circuit board. Users, however, don’t need to worry about the details; they can rely on the work of others and their AI agents.
The project also distinguishes analysis from source-code recovery. Native analysis returns pseudocode and assembly. The agent uses those findings to develop and test a new implementation.
That is not the same as retrieving the manufacturer’s original source tree, which would be true reverse engineering. Quick modification or long maintenance problem? The immediate attraction is easy to understand: fewer ads, a cleaner interface, or returning an old favorite device to life.
Reliability and security are another matter. Just because you can create or download an alternative for your hardware doesn’t mean that it’s in the least bit reliable. That may be fine for reviving an otherwise useless Samsung Galaxy 9 phone, but do you really want to do that with your year-old smart TV?
I don’t think so. Bricking old gear is one thing; bricking your almost-new LG TV is another. Also: How I made a paid Mac app in 11 days with Claude – no writing code There is a security trade-off as well.
ADB grants powerful access to a device, and network debugging can introduce exposure if left enabled or inadequately protected. Separately, installing unreviewed replacement code creates the possibility of new vulnerabilities. CleanYourTV and dumbtv both sound great, but they also might contain malware.
As my colleague Jack Wallen put it recently in an article about rooting your Android phone, “You could, however, wind up installing a malicious OS , which can in turn lead to your information being sent to a bad actor. You need to ensure the OS you’re installing is safe.” Right now, there’s no way to be sure you’ll be safe. For now, the clearest development is not that AI can “reverse-engineer anything.” It’s that coding agents are making some existing forms of device modification more accessible while exposing owners to the responsibilities of maintaining the software their manufacturers once controlled.
Me? I’ll try it, but then I have an office full of no-longer-supported gadgets and devices. If all you have is gear you use every day, I wouldn’t touch these programs until we have a lot more answers about their stability and safety than we do today.
Steven Vaughan-Nichols Senior Contributing Editor Steven J. Vaughan-Nichols is a freelance writer and technology analyst. Besides ZDNET, he works with Foundry (Formerly IDG Communications), The Register, The New Stack, TechStrong, and Cathey Communications.
He does not own stocks or other investments in any technology company. See full bio
Source: ZDNET
