Feasibility of LLM decompilation of x86/Motorola binaries back to Python
Key claims
- A user asks how difficult it would be to compile Python to x86/Motorola binary code and then train an LLM to decompile the binary back into the original Python.
- A user asks when a fundamentally new paper will arrive that transforms LLMs and yields an absurd multi-thousand-fold improvement.
Volume over time
Peak 1 mentions/hour · 2 hourly buckets
Source threads (2)
Every thread this story was extracted from, with live and archive links so the evidence is verifiable.
- /g/109566612 ↗ live · archive impatient evergreen 7 replies A user asks when a fundamentally new paper will arrive that transforms LLMs and yields an absurd multi-thousand-fold improvement.
- /g/109563151 ↗ live · archive curious evergreen 3 replies A user asks how difficult it would be to compile Python to x86/Motorola binary code and then train an LLM to decompile the binary back into the original Python.
Related stories
- Crash-learning x86 assembly in three days 1 mentions · shared: x86
- Skepticism that HuggingFace incident justifies AI panic since models are just LLMs 1 mentions · shared: llm
- Von Neumann memory bandwidth bottleneck questioned for LLM inference 2 mentions · shared: llm
- Debate over AI code generation capabilities and limitations 122 mentions · shared: llm, python