I've had a tiny script running since around 2012 that takes whatever I post here and pushes it to Twitter automatically. Nothing fancy, just a cron job and a handful of lines that hit the API every hour. It's been quietly doing its job for over a decade without me thinking about it once. That's probably about to change, and it's not really my fault.
Twitter rolled out its new API pricing at the end of March, and it's brutal if you're not a company with a business card to show for yourself. Free access, the kind hobbyists and weird little bots have used forever, is basically gone. If you want any real access now you're looking at $100 a month for the "Basic" tier, and if you need actual scale you're suddenly staring at a number with a comma in it, something like $42,000 a month for the enterprise tier. There isn't really a middle option for someone who just wants to run a bot for fun.
And the bots, man. That's the part that actually stings. There's a bot called @big_ben_clock that's tweeted "BONG" thirteen times a day, timed to the actual chimes of Big Ben, for something like nine years now. Whoever runs it posted this week that they can't justify $100 a month to keep a joke account making bell noises alive, so it's done. Gone. Same story for a pile of smaller accounts I've followed for years - earthquake alert bots, transit bots, one that tracked edits made to Wikipedia from congressional IP addresses. None of these things made money. None of them were ever supposed to. They existed because the API used to be free and somebody thought it'd be funny or useful, and now that math doesn't work anymore.
I get that Twitter needs revenue, obviously, it's not exactly swimming in cash right now and I don't think anyone's under the illusion the old free-for-all API model was sustainable forever. But there's a difference between charging real money to companies pulling millions of tweets for sentiment analysis dashboards, and pricing out a guy who wanted his blog to auto-post. The $100 tier reads like it was designed for absolutely nobody in between those two groups. Either you're a hobby project that gets nothing, or you're a company that can eat a five-figure invoice, take your pick.
My own script is small enough that I'm not sure yet if it'll even get cut off or just quietly stop working one day when a key expires. I haven't gone and paid the $100, and I probably won't, honestly. It's kind of silly to spend that much a month just so this blog auto-tweets a link that maybe four people click. I'll probably let it die and just post links manually like it's 2009 again, or not bother at all. Feels a little bit fitting actually - this blog's outlived a lot of platforms coming and going since 2011, it can outlive one broken cross-poster too.
What's getting lost here isn't really about Twitter specifically, it's about the kind of stuff that used to be possible on the open web without asking anyone for permission or a credit card number. Somebody somewhere built the Big Ben bot just because it was funny. Nobody at Twitter approved a budget for it, nobody sat in a growth strategy meeting about it. It just existed because the barrier to trying something dumb and small used to be basically zero. That barrier's gone now, at least on this particular platform, and I don't think we're getting it back.
The Mastodon crowd has been saying "told you so" for months about exactly this kind of thing, and I get why, though I still find the whole federated timeline experience kind of a chore to actually use day to day. Maybe that's a me problem. Either way, if you're running a little bot or a little script that touches Twitter's API, go check on it this week. It might already be quietly broken and you just haven't noticed yet.