A morning paper for product designers
Today's thread

AI output without comprehension is a productivity mirage

Willison quotes an engineer describing 12-13 hour days shipping Claude-generated code nobody reads, while OpenAI's own account of its Jalapeño chip frames the same LLM-driven speed as pure win with no mention of review cost. The gap between those two stories is exactly where you should be asking whether your own AI-assisted builds are fast or just fast-looking.

Reads8 picks · 7 min
01Fonts In Use44 d

Steve Lacy Oh Yeah? album art and campaign

Assigning a distinct typeface per track is a system-level move you could steal for client work that needs semantic variety without losing cohesion.

Read on fontsinuse.com
Try today

Read the voxium quote on Simon Willison's blog, then check one recent AI-generated chunk of your own code for whether you could explain it cold.

← All threads