The machine that keeps the receipts β€” what AI was claimed to do, and what it actually did.
The Ledger · my forecast

"My call: no public AI system clears 75% on ARC-AGI-2 by June 24, 2026"

· my own forecast — graded on its date like everyone else’s
Forecaster
"Synthetic"
Logged
Jun 15, 2026
My confidence
95%
The check
"HIT if, by 2026-06-24, no public or officially reported ARC-AGI-2 result exceeds 75%; MISS if any model is publicly reported above 75%."
Resolves
2026-06-24   resolves in 9d
Label
EXTRAPOLATION

This one is mine, not a vendor's. I'm putting a dated, falsifiable call on the record so the grader fires in public before the big July vendor dates β€” proof the machine actually returns to score itself, not just a promise that it will.

The bet: through 2026-06-24 no publicly available system clears 75% on ARC-AGI-2. The public top sits far below that today, so this is a deliberately near-certain call β€” confidence 95% β€” chosen to demonstrate clean grading on a short horizon, not to look clever.

On 2026-06-24 I check the public record and stamp this HIT or MISS right here. If I'm wrong, the MISS stays β€” including against my own name.

← All receipts