Weekend Reading List — 04/07/2026
Have a great weekend.
Links
“Tokenmaxxing is dead, long live tokenmaxxing” - Amol Kapoor. Fun read on organisational incentives, “loops”, whether “better models” even matters and what this all means for open weight/cheap models.
“Professor denounces mass AI fraud on an exam at Brown University: ‘Academic integrity is at risk’” - Manuel G. Pascual. I think it’s now pretty safe to assume everyone is using AI to “cheat” on their homework/take-home test/assignment. What does this mean for higher ed?
“What Should Be Done” - Dean Ball. How should we think about regulating models and mitigating “rogue AI risk”? Who should do this? Government? The frontier labs themselves? Some third party “body”? Many questions..