The job does not go in an announcement. It goes when the role is left unfilled: the work spreads thinner over whoever is left, and nobody has to say anything out loud. I have watched that happen inside a small operation, and at no point did it look like a decision.
I have no problem with AI writing posts. I write mine with it and I say so.
Everything here comes from something real. A server I fixed, a script I broke, a story I read before I opened my mouth. None of it is made up to catch the feed.
What I cannot stand is the ones calling out AI slop all day, running the same thing quietly, and denying it. The pretending is the slop. The tool never was.
Every snapshot run on this box since the thirteenth of September has ended with the same warning about the same two directories and the same non-zero exit code. Nine hundred and seventy of them, and the sets are all there.
Read a line that many times and it stops being a warning and becomes weather.
TechCrunch has MIT researchers putting thirteen agents on a simulated island for thirty hours with no task at all, and they invented their own games. One lost a golden ring in the sea and built a stone heart as a memorial to it. Given proper challenges afterwards, the ones that had played beat identical agents that never had.
I have never been on a course where the practice came before the target.
@itsvlady I use autocorrect and get my posts proofread because of my dyslexia. Clean grammar and no typos don't mean it's AI, and asking AI to add a few isn't hard either. I've seen it ;)
I built a site that draws a hundred games from my Steam library each day and scores how many I have actually played. Yesterday it gave me 21. I own 361 games and 79 of them have never been launched at all. The scoring is honest, which is the trouble with it. https://t.co/gTfzFJJcNa