Heidelberg AI News

Video · Two Minute Papers ·

Is Software Over?

In short

AI systems are rapidly improving at reverse-engineering software from compiled code, behavior, and visual output, making near-identical reimplementations increasingly feasible. The video argues software is not “over,” but personalized, low-cost software could become common as models improve, despite legal, professional, and access restrictions.

Key points

  • MirrorCode improved from 15% to nearly 100% test pass rates in nine months when reconstructing software without source code.
  • A 12-million-parameter model recovered mathematical equations with about 90% accuracy from software behavior.
  • WebUI-95 achieved a 0.938 visual similarity score while using 95% less code than the original website.
  • AI can iteratively compare its recreation with the original software and refine differences using a ground-truth testing loop.
  • Software functionality may not be copyright-protected, but assets such as models, textures, and characters can be.
  • Anthropic’s Claude 5.5 Opus reportedly discourages some game-recreation and legitimate reverse-engineering tasks.
  • Tailored homebrew software may become widespread if it supports interchangeable formats and collaboration.

Written by an AI model from the video’s captions and description.

Watch on YouTube The opening of the article at youtube.com.

In this story

A tool or model opens in the Heidelberg AI Curriculum: what it does, what it costs, how to start. A repo opens on GitHub.

Related stories

All stories

VideoTwo Minute Papers

Funkiest Zoom meeting EVER (Surprise video)

A surprise musical performance about learning through mistakes, staying present, and becoming better than yesterday. Its message matters for AI builders because progress depends on persistence despite noisy, imperfect results.

VideoTwo Minute Papers

DeepMind’s AI Just Did The Impossible

DeepMind’s AlphaProof Code used AI agents and Lean to formally verify that Signal’s cryptographic implementation correctly encrypts and decrypts messages. The work matters because mathematically verified software could reduce security vulnerabilities, though the proof covers only specified assumptions and excludes AES, group chats, and the rest of the app.

VideoTwo Minute Papers

The OpenAI Astra Weight Hack, Explained

The video explains how OpenAI Astra can inspect neural-network weights, diagnose failures, and sometimes improve model behavior through weight-space programming. The results are promising but limited: these capabilities were demonstrated in separate, narrow experiments rather than as a universal system that automatically teaches or upgrades AI.