GitHub's 110K-star REA Lets AI Reverse-Engineer Any Software, Threatening Subscription Models
APPSO · wechat · 2026-10-12
Developer Brandon Thomas declared "Software is over" after using AI to build seven Adobe alternatives. The open-source GitHub project REA (Reverse Engineer Anything) surpassed 110,000 stars by October 12, letting AI analyze installers, runtime behavior, or web requests via tools like Ghidra, Hopper, and IDA, paired with Claude Code and Codex agents to rebuild software functionality.
Verified cases
- Aegis authenticator: agent analyzed the Android package to find the code-generation algorithm; all 133 test results matched
- Notion copy function: traced text and HTML across processes into the clipboard
- DX-Ball sound: 3,205 tests matched old and new function outputs; new function produced matching 63-byte machine code under the original compiler
- Touhou projectile angle calculation: REA inspected the original DOS program against C++ reconstruction code
The repo's SKILL.md defines an investigation method: locate feature code, trace data flow, distinguish observation from inference, narrow scope and save evidence, then run comparative tests after rebuilding.
Broader use: Classic games like Halo and GTA: Vice City have been ported to browsers; VibeOffice brings Office 2003 to the browser with local docx/xlsx/pptx editing, with the author reporting roughly $1,000 in model usage for the UI and core features plus several times that for compatibility work.
Business impact: Matthew Berman argues companies whose edge comes only from the software itself will face pressure, with training, support, and maintenance becoming more important. Replacement tools only need to nail the features users actually need; single-purpose tools face the most pressure, while enterprise software competition is more complex due to compatibility, collaboration, and accountability. Reverse engineering isn't inherently illegal, but cracking protections, copying copyrighted code, or misusing trade secrets can trigger legal disputes.
More from AGI Musings
- Blogger claims hyperscaler-backed AI regulation aims to crush small labs, entrench monopoly — Ghost_Pilot_MD · 2026-10-25
- Toddler development as a metaphor: evolutionary pre-training and efficient post-training — paigeinsf · 2026-10-25
- "I do miss programming": dev says AI coding killed the flow state — generativist · 2026-10-25
- Anthropic CEO on learning to code: programming goes first, but your 5% still matters — 2C_ornot2C · 2026-10-25
- After 22 years studying workforces: the AI question isn't replacement, it's 'what kind of people' — 2C_ornot2C · 2026-10-25
- Mechanistic interpretability will never truly be solved, researcher argues — burny_tech · 2026-10-25