Charlotte Reed

A new NBER paper tracks 500,000+ GitHub developers against their actual AI telemetry. The task-level gains are enormous and they mostly don't survive the trip to a shipped release. Across four app marketplaces: a sharp rise in new apps, no rise in total usage.

Most of the AI-productivity argument you've read this year is two camps shouting single numbers at each other. There's a paper out that makes both camps right and is more useful than either, and the revised version went up this month, so it's worth twenty minutes before you write your Q4 plan.

Sometimes it takes 30 seconds.

Sometimes 5 10 minutes.

You ve already checked Slack, replied to emails, reviewed your tasks

Early on, we spent a lot of time trying to sell into a market built around long-standing vendor relationships.

When security and trust are involved, being the new company in the room is a pretty big disadvantage. Even if the product is good, getting an institution to switch is a whole different problem.

So we changed direction. Instead of selling to the institution, we started going directly to the people who would actually use the product.

It's definitely a slower path. We're building trust one person at a time instead of closing one deal that could potentially bring in thousands of users.