The grocery productivity paradox is fascinating because I think it generalizes: AI often makes the measurable parts of a job more efficient while making the unmeasurable parts (trust, judgment calls, edge cases) harder to account for. I saw this building an AI ops tool — ticket resolution time dropped 40%, but customers started complaining the answers felt "less human," which never showed up in our dashboards until churn did. Curious if the data in this piece shows anything similar on the customer satisfaction side, not just throughput.
CXMT is the live version of this cycle. pentagon blacklisted it in 2023, kept building commodity DRAM while everyone else converted to HBM, and now Apple is lobbying washington to buy from it. atoms before the software demand wave.
The grocery productivity paradox is fascinating because I think it generalizes: AI often makes the measurable parts of a job more efficient while making the unmeasurable parts (trust, judgment calls, edge cases) harder to account for. I saw this building an AI ops tool — ticket resolution time dropped 40%, but customers started complaining the answers felt "less human," which never showed up in our dashboards until churn did. Curious if the data in this piece shows anything similar on the customer satisfaction side, not just throughput.
CXMT is the live version of this cycle. pentagon blacklisted it in 2023, kept building commodity DRAM while everyone else converted to HBM, and now Apple is lobbying washington to buy from it. atoms before the software demand wave.
Great report Moses. The leaders chart is outstanding. A rotation from "bits to atoms" indeed.