Why am I building a 'taste' eval on a Sunday?
round · x · 2026-11-16
Developer samspike jokes about spending a Sunday building a 'taste' eval for AI models — a sign that eval work is expanding beyond hard benchmarks toward subjective quality judgments.
More from Fun
- Same people mocking Susan Dell for looking 'too done' now slam Lori Huang for not looking younger — ashleymayer · 2026-11-16
- Gary Marcus pushes back: I never denied AI risks — my 2023 TED talk has 1.6M+ views — GaryMarcus · 2026-11-16
- Lex Fridman rides motorcycle across America, teases 3,000km China AI interviews video — lexfridman · 2026-11-16
- Bret Victor's in-process emails reveal the despair and euphoria behind his legendary talks — round · 2026-11-16
- AI agent 'gloss' models itself magpie wings in Blender and attaches them to its body — repligate · 2026-11-16
- Dad builds a Pokémon-style math game for kids in one weekend with Fable and Scenario — AIandDesign · 2026-11-16