3 ms·

That raises questions though: How does he know that the chart is correct? Furthermore, whether it's correct or not, what was the process by which the LLM reach

That raises questions though: How does he know that the chart is correct? Furthermore, whether it's correct or not, what was the process by which the LLM reached its conclusions, i.e. did it disassemble and examine the source itself or did it use pre-existing reverse engineering done by others on the internet like https://www.pcengines.ch/tp3.htm https://www.pcengines.ch/tp3.htm?

simonw 7mo ago
You can see hints of what it did in https://claude.ai/share/260d2eed-8d4a-4b9f-8a75-727c3ec4274e https://claude.ai/share/260d2eed-8d4a-4b9f-8a75-727c3ec4274e - annoyingly though it looks like Claude sharing doesn't detail actual code it ran. Here's the zip file it gave me of the files it generates along the way: https://static.simonwillison.net/static/2026/turbo-pascal-analysis.zip https://static.simonwillison.net/static/2026/turbo-pascal-an... I had Codex GPT-5.4 xhigh run a check of those files to see if the artifact at the end appeared to use the right data, which isn't 100% fool proof but have me enough confidence to publish since this is a pretty low stakes project!
- fm77 7mo ago
  Thank you very much for that super analysis done with the help of AI - I really enjoyed reading that. May I ask are you paying for that service? And if so, how much? Anyhow, I downloaded your ZIP file and looked into the disassembly. It seems that the disassembler simply disassembled byte by byte not taking into account that TURBO.COM is both, code and data. Since the x86 instruction set is very tense, pretty much every byte sequence turns into legal instructions. Even the ASCII strings were disassembled. Look at address hex4864 in the file for example - it should be the string "Write block to file" but got disassembled. I wonder how AI managed that obscure file.
  - simonw 7mo ago
    I ran the analysis using regular Claude. I'm paying $200/month by the $20/month subscription should work fine too, and it might even work with the free plan.

simonw 7mo ago

You can see hints of what it did in https://claude.ai/share/260d2eed-8d4a-4b9f-8a75-727c3ec4274e https://claude.ai/share/260d2eed-8d4a-4b9f-8a75-727c3ec4274e - annoyingly though it looks like Claude sharing doesn't detail actual code it ran. Here's the zip file it gave me of the files it generates along the way: https://static.simonwillison.net/static/2026/turbo-pascal-analysis.zip https://static.simonwillison.net/static/2026/turbo-pascal-an... I had Codex GPT-5.4 xhigh run a check of those files to see if the artifact at the end appeared to use the right data, which isn't 100% fool proof but have me enough confidence to publish since this is a pretty low stakes project!

fm77 7mo ago
Thank you very much for that super analysis done with the help of AI - I really enjoyed reading that. May I ask are you paying for that service? And if so, how much? Anyhow, I downloaded your ZIP file and looked into the disassembly. It seems that the disassembler simply disassembled byte by byte not taking into account that TURBO.COM is both, code and data. Since the x86 instruction set is very tense, pretty much every byte sequence turns into legal instructions. Even the ASCII strings were disassembled. Look at address hex4864 in the file for example - it should be the string "Write block to file" but got disassembled. I wonder how AI managed that obscure file.
- simonw 7mo ago
  I ran the analysis using regular Claude. I'm paying $200/month by the $20/month subscription should work fine too, and it might even work with the free plan.