Artificial Intelligence — Theses
5 theses·Avg conviction 6.0/10·5 stale
T1never reviewed
Agentic reasoning will consolidate around a standard stack (VLM + tool use + memory + RL) by end of 2027
6.0/10
no history yetT2never reviewed
Mechanistic interpretability will fail to keep pace with model capabilities, creating a widening safety gap
6.0/10
no history yetT3never reviewed
Evolutionary code generation (AlphaEvolve pattern) will become a standard optimization tool in every major tech company by 2028
6.0/10
no history yetT4Anever reviewed
World models will be a central architectural story of embodied/robotics AI by 2030
6.0/10
no history yetT4Bnever reviewed
JEPA and generative world models will specialize to different use cases (control vs. simulation), not converge on a single architecture
6.0/10
no history yetHow theses work
Each thesis has a conviction score (0–10) and a history of re-scores with reasons. A thesis is flagged stale if it hasn't been re-reviewed in 90 days — that's a signal evidence may have moved. Re-score by appending a new dated line under **History.** in the thesis markdown.