The Life and Death of Direct File [pdf]

I was surprised how much I enjoyed reading this. Although it's necessarily opinionate d(the lead author was part of the Direct file team) it comes off as very even-handed, giving equal consideration to successes and failures, and addressing the partisan political environment in factual terms.

It's unusually well-written, providing lots of key details or dropping in relevant context without ever losing sight of the big picture and the looming deadlines. I expected to just read the executive summary and then skim the rest, but it held my attention the whole way through.

There isn’t much to study - it didn’t succeed or fail on the merits - someone didn’t like it because he didn’t create it so it got axed.

Very basic analysis by da bot says it costs the government roughly 226$ per filing vs a private company 40$.

Would love to hear the opposite argument but the point I first saw online is saying the same, that it cost more tax payer money to run than it saved those tax payers.

Scale problem? Open to hearing the opposite argument.

A fascinating , yet sad, post-mortem.

Anyone wheeling and dealing along in the shadows between "politics" and "deliverable, functioning product" should read it. Or, I guess, paste the link into Claude or ChatGPT and ask it to summarize for them, lessons learned in bullet points and how it could apply to $their_job.

But I'm generally a weirdo who finds these kinds of reports fascinating, so don't purely go on my recommendation.

Publication homepage with some intro/context instead of direct pdf: https://www.ischool.berkeley.edu/research/publications/2026/...

Eh, government should not be in the business of maintaining digital infrastructure, regardless of how capable it might be. As much of it as possible should be contracted out to 3rd parties.