The Shape of AI: Jaggedness, Bottlenecks and Salients
3.00T1 sourceOne Useful Thing (Ethan Mollick)
Source record
Published by One Useful Thing (Ethan Mollick) (T1 source). The original is at https://www.oneusefulthing.org/p/the-shape-of-ai-jaggedness-bottlenecks.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryEthan Mollick revisits his 'jagged frontier' framework for AI capabilities, discussing where bottlenecks and salients appear, and uses it to explain why Google's 'Nano Banana Pro' (Gemini image generation model) marks a meaningful capability jump rather than a typical incremental release.
Why it mattersMollick's jagged-frontier framing is a working mental model many readers already use; the post extends it with bottlenecks and salients, and grounds it in a current model release.

Cited by
No citations on record.
