Was busy doing some edgecase analysis for my new AMM, asked o1 deep research for some review on a fairly generic model... Saw it started reading my own research unprompted... this won't end well...