Researcher Annotates 7600-Word Diss Track On AI Interpretability Methods

Original post

I grew obsessed and ended up writing 7600 words to annotate this diss track. I learned a lot by doing this. It led me to conclude that everyone is more or less equally (un)deserving of disses: https://web.stanford.edu/~cgpotts/blog/diss/

Christopher Potts@ChrisGPotts

Like any good advisor, I felt duty-bound to defend @aryaman2020 and @ZhengxuanZenWu in this rap battle. However, the SAE diss track I wrote was so devastating as to be unanswerable, so I decided to graciously balance things out with a second verse dissing causal interp.

7:43 AM · Jul 6, 2026 · 638 Views

VIEWS266LIKES2REPLIES1

Christopher Potts@ChrisGPotts

This is a task deemed too dangerous by Claude Fable, so please proceed with caution:

Aryaman Arora@aryaman2020

it turns that @ChrisGPotts's raps are dangerously pushing the frontier of AI research, and so Claude Fable is banned from helping you decipher them

1h26620

BOOKMARKS2

Christopher Potts@ChrisGPotts

I nominally knew all this stuff before doing the annotations (I wrote the disses), but the writing process itself was so valuable, as always. I am reminded of this wonderful parable (written in 2012, despite what the page says): https://croissanthology.com/earring

Christopher Potts@ChrisGPotts

This is a task deemed too dangerous by Claude Fable, so please proceed with caution:

1h13912