AI 2027 Tracker
Tracking predictions from the AI 2027 scenario against reality.
Status Breakdown
Visual Overview
Status Distribution
Category Coverage
By Category
Recently Updated
It could offer substantial help to terrorists designing bioweapons, thanks to its PhD-level knowledge of every field and ability to browse the web.
China has aggressively hardened security by airgapping (closing external connections) and siloing internally.
Agent-2 is 'only' a little worse than the best human hackers, but thousands of copies can be run in parallel, searching for and exploiting weaknesses faster than defenders can respond. (page 10)
CCP leadership recognizes the importance of Agent-2 and tells their spies and cyberforce to steal the weights. (page 11; Appendix D provides detailed theft mechanics.)
OpenBrain's security level is typical of a fast-growing ~3,000 person tech company (RAND's SL2). They are working hard to protect their weights and secrets from insider threats and top cybercrime syndicates (SL3).
Current Assessment
March 2026 Summary
The AI 2027 scenario remains remarkably on track. All qualitative predictions about agent emergence, coding transformation, infrastructure investment, and institutional skepticism have been confirmed. The quantitative predictions are tracking at roughly 70% of predicted speed — behind on raw compute milestones but ahead on agent time horizons and labor market impact.
The most interesting divergence is in agent autonomy: METR time horizons are doubling faster than predicted (every 3 months from 2024+, vs the 4 months predicted), while the absolute compute scaling is lagging. This suggests the scenario may have underweighted algorithmic progress relative to brute-force scaling.
Bottom line: Directionally correct, with timing ~6-12 months behind on hardware but potentially ahead on capability milestones. The overall trajectory remains concerning for anyone who thought the scenario was too aggressive.