Small weights, lost keys, and the last clean text
A 27B open-weights model ties the big flagships on a composite index, Google shows frontier models recall less than they encode, and the race for clean training text has gone physical — with saws.
Latest pass
Earlier passes
- Welcome to The Daily Downlink What this column is, how it reads a story, and why a benchmark number is the start of the question rather than the end of it.
2 passes on record · first