[A_Venom_Roll]

Perfect timing - Sharopolis just released a video [0] using this exact disassembly to mod Super Mario-style physics into the game.

He used Claude Code to 'vibe code' the assembly changes, leveraging the fact that the disassembly identifies about 2KB of unused memory. It’s a fascinating look at how LLMs can now navigate and modify 40-year-old Z80 assembly when provided with a well-documented codebase like this one.

The video sparked a lot of discussion in the comments, with some people being very upset he used AI for this.

[0] https://www.youtube.com/watch?v=gxxNgZgd88I"

[s-macke]

AI is very effective for reverse engineering. Unless you’re doing it purely for fun, it makes sense to use AI where it helps.

I’ve tried to visualize the “navigate and modify” process you mentioned in [0]. It’s mesmerizing.

Because reverse engineering outcomes are comparatively easy to verify, it’s a good fit for training for AI. I expect major progress in the next few years, potentially to the point where reverse engineering many binaries becomes highly automated.

[0] https://github.com/s-macke/OpcodeOracle

[pjmlp]

My first agent test was pointing it out to my toy compiler repo (L98) and ask to translate the AT&T Assembly files that gave me so much trouble to come up with (my head has Intel syntax burned into it), and translate it back to Intel syntax.

In a couple of seconds I had it back.

Didn't bother commiting the changes, because it works and was a toy compiler anyway.

[qingcharles]

The newest February Google Deep Think update is excellent at writing 16-bit x86 CGA code. It very quickly put together a complete 3D software renderer in assembler. I'm trying to replace the rendering functions of the original 8086 Elite PC since there have been significant improvements in how to write video RAM since it came out.

I'll have to try Opus 4.6 now.