News for hackers. Stuff that matters.
Re:Engineering Fought Productivity Metrics. AI Ended the Fight by pacecar on Saturday November 14, 2026 @09:33AM (#50497316)
AI hasn't yet "ended the fight" but it has definitely changed the discussion. I think the number of pull requests and the cost per pull request is a useful metric. Obviously it's not perfect and it's subject to gaming, but it is easy to measure and carries some signal.

If AI can evaluate the value or impact of each change that could add another data point so that trivial and easy PRs don't carry the same weight as large complex ones.

At the end of the day it's about whether or not the software is meeting business needs and that's always been hard to measure and still is.

Re:Engineering Fought Productivity Metrics. AI Ended the Fight by bradknowles on Saturday November 14, 2026 @11:02PM (#50503897)
Tokenmaxxing is a bad productivity metric.

So is lines of code.

So is number of commits or number of pull requests. Or cost of pull requests.

All of these things are trivially easy to game.

If you can’t do the hard thing of measuring productivity in terms of actual impact on the bottom line, then you shouldn’t waste your time and mine creating metrics that are trivially easy to game.

Re:Engineering Fought Productivity Metrics. AI Ended the Fight by jjcob on Saturday November 14, 2026 @12:05PM (#50498610)
And apparently it also replaced bloggers...