May 24, 2026, 7:04 PMI've brought up the McNamara fallacy before in the context of metrics-driven management, something all too common in software engineering. The basic idea is that just because you don't have the ability to easily measure something doesn't make it not important. Something can be vitally important to your success, but difficult to measure for whatever reason.
Let me apply that to what AI is doing to OSS and to common culture.
May 24, 2026, 7:06 PMOne of the most difficult things to measure is the *absence* of a thing. A survey, as xkcd humorously put it, measures amongst the demographic of people who love answering surveys. NPS scores measure how people who are currently using a product *and love clicking that goddamned button* feel about that product.
None of that measures people who have left the room.
May 24, 2026, 7:08 PMOpen source has, to this point, thrived on people who put in work and passion to make open source a thing. People volunteering in OSS organizations, people making and maintaining projects, people subverting work policies to support OSS, and so forth.
It's a lot of people going out of their way to make OSS a thing.
How do you measure the people who just... give up on the whole thing, because why bother when it all goes to automated scabbing in the form of AI products?
May 24, 2026, 7:10 PMOut of the wreckage of the tech industry trying to solve "wages" and "unions" by setting more than one trillion dollars on fire, I'm willing to predict that there's going to be an effort to make new institutions and structures... some of them positive, some of them the disaster capitalism sort of filling the power vacuum with something even more extractive.
The latter will lie to you with numbers, hoping you'll forget the importance of those negative numbers, of the people who left the room.