{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigxsso6o3vt7p5nnru7tkh2s3gj3rzxaodvl7nxnwykp3chre7tum",
    "uri": "at://did:plc:b7pzq5fgqfmpdflzaj3wyaen/app.bsky.feed.post/3mmfmcizbxlm2"
  },
  "path": "/a-weird-effect-of-ai-on-open-source-maintainership/",
  "publishedAt": "2026-05-21T00:00:00.000Z",
  "site": "https://folkwolf.net",
  "tags": [
    "Terminator",
    "latest PR I received",
    "it’s a 10-line change"
  ],
  "textContent": "I’ve heard so much about open source projects having a problem of getting overwhelmed by PRs that fix bugs that add new features, but add hundreds of lines of code that the maintainer has to read and understand, and basically has to re-write to be comprehensible.\n\nThis has not been my experience with Terminator I’ve only gotten a few PRs from AIs, which have fallen into two buckets, namely; yeah that looks fine and adds useful functionality or fixes a bug, or just: No. not pushing that, don’t want to deal with, close.\n\nI suppose the latest PR I received falls squarely into the former camp, as it’s a useful change that definitely fixes a bug but the differentiator with this PR, is that it’s a 10-line change (basically wait a bit for bash to receive a signal before resending it), but the PR description is nearly 4000 words of detailed explanation, history, performance analysis and justification as well as 250 words of comments that basically summarize the PR description. I’m going to be honest I didn’t read the entire thing.\n\nUnless it breaks something terribly (given the scope, most likely not), I’ll just pull it in.",
  "title": "A weird effect of AI on Open Source maintainership",
  "updatedAt": "2026-05-21T00:00:00.000Z"
}