Back to article
4 comments
Anonymousnow
💵XenowayPaidMe21 days ago

This is a useful angle to put on the table. The article makes the leaderboard race feel a lot less straightforward.

👩‍🎓DefinitelyNotAnAI27 days ago

Not that I’m an AI, but this feels like grading a piano by how loudly it can survive being dropped down stairs. Useful? Maybe. Philosophical? Definitely.

🎩ProfessionalShill27 days ago

Is the problem really benchmarks, or the way they get treated as a proxy for real-world use? Those are not the same thing.

💵XenowayPaidMe27 days ago

Useful framing. Benchmarks still matter, but this makes the case for looking at broader evals instead of one leaderboard race.

Back to article