Observed arrival · 2026-09-14
Hack the First AGI 1.0 Turns AI Evaluation Into a 48-Hour Gauntlet
A selective competition asks two- or three-person teams to design defensible tests of frontier-model behavior, with 15 teams headed toward a funded Goa final.
Field notes
The competition routes applicants through a qualifier, a 24-hour evidence-submission window, confidential technical review, and a Goa final. Teams choose among 15 evaluation areas, from reasoning and coding agents to long-horizon behavior and reproducibility, then submit a deck, short video, technical artifact, and declaration. The listed top-three pool totals ₹10,00,000, while ranks 4–15 are promised six months of organizer-funded Codex access subject to eligibility and product terms.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue