A mathematician chronicles AI's shift from toy to near-indispensable between 2024 and 2026, estimates it made him about 50% more productive, and argues what AI really lacks is abstraction-making.
The author says AI can now do in minutes what used to take him hours, yet his overall productivity is up only about 50%.
What struck me most is his complaint about AI writing: the problem is not formulaic style but that its proofs make no goddamned sense. It has the technical skills but cannot break an argument into modular pieces with clear assumptions and conclusions, so it descends from simple into a rampage of complexity.
I also nodded at his self-warning: every time he lets AI run until dawn, ends up with nothing, and has to start over, he feels foolish.
Curated from high-quality sources, with concise summaries and key takeaways.
