What we get wrong about evaluation sets
The tooling is better than it was. So are the expectations.
The pitch for evaluation sets is straightforward. The part that gets left out is the operational bill, which arrives in a different department.
Subscriber story
Continue reading with a subscription
This piece is part of the subscriber archive. Subscribe to the daily brief for full access — one email, no card, cancel any time.
Already a subscriber?
#licensing#evaluation#infrastructure
More in this section
ai
Why small models are winning on-device
Four-bit quantization plus better distillation is making 3B models genuinely useful in production.
🔒aiSubscriber
Synthetic data is hitting a wall nobody wants to name
Model collapse is real, but the more immediate problem is boring: attribution and licensing.
🔒aiSubscriber
The open-weights licence split is becoming permanent
Two families of licence, two incompatible expectations, and a lot of downstream confusion.