A Satirist Hated the Internet Before AI. A Food Bank Algorithm Doesn't Know You're Pregnant.

A Bluesky post with 137 likes this week made the joke that's been waiting to be made. "You: basic, hate the internet because of AI or social media brainrot or whatever," it reads. "Me: wise, has always hated the internet, for killing off CD-ROMs." The joke works because it's true in the way satire is always true — the speaker isn't actually wiser, just more arbitrary. The grievance preceded the object of the grievance. People who feel the internet has always been extractive, corrosive, and badly designed for actual human flourishing were saying so before the algorithms, before the engagement optimization, before generative AI entered the feed. AI didn't break a working thing. It inherited a broken one.

That inheritance is what makes a second Bluesky post, this one with 45 likes and a harder edge, worth sitting with. The author described AI facial recognition being used to gate food distribution — to verify identity before giving people access to emergency nutrition — without any apparent consideration of what pregnancy does to a body or what childhood does to a face. "This is what we mean when we say the system is patriarchal," the post said. The fury is precise: not that AI was used, but that whoever designed the system optimized for a default user and never asked what happens to the bodies that don't match it. Pregnant women's faces change. Children's faces change faster than databases update. The bias encoded into the system isn't a bug introduced by AI — it's a structural assumption that AI made load-bearing.

These two posts are doing the same diagnostic work from different registers. The satirist is describing the long continuity of internet disappointment — the sense that each new technology promises transformation and delivers a shinier version of the same extractive architecture. The food bank post is describing what that continuity costs in specific, physical terms: a pregnant woman at a distribution site, her face flagged as unverified, turned away or delayed because someone's training data didn't include her. The ethical failure isn't abstract. It has a location and a body and a consequence.

The broader conversation around AI and social media is running hot right now, with one Bluesky post noting that AI has the capacity to "turbocharge" harms that social media already facilitated — and that the current US regulatory posture under Trump offers Big Tech almost unlimited latitude on exactly the question of what those harms are worth preventing. What the satirist and the food bank post together suggest is something the policy debate tends to skip: the question isn't only whether AI makes things worse, but whether the systems AI is being plugged into were ever designed for the people most vulnerable to going wrong. The CD-ROM joke is funny because it's petty. The food distribution failure is not funny at all. They're describing the same architecture.