This might be the most disturbing AI paper of 2025 ☠️
Scientists just proved that large language models can literally rot their own brains the same way humans get brain rot from scrolling junk content online.
They fed models months of viral Twitter data short, high-engagement posts and watched their cognition collapse:
- Reasoning fell by 23%
- Long-context memory dropped 30%
- Personality tests showed spikes in narcissism & psychopathy
And get this even after retraining on clean, high-quality data, the damage didn’t fully heal.
The representational ��rot” persisted.
It’s not just bad data → bad output.
It’s bad data → permanent cognitive drift.
The AI equivalent of doomscrolling is real. And it’s already happening.
Full study: llm-brain-rot. github. io
@warmwind_OS Looks fantastic! Consider removing the password requirement for waitlist signups. It would improve the user experience and make it easier for more people to join..
@Chobani you should change the way you print expiry dates. I trashed a bunch of your products because I couldn’t read the dates on them and I’m not visually impaired. Imagine how much this would fail accessibility compliance checks!
Climate change is a crisis — but we have solutions to start addressing it. I’m now funding a project on @project_wren to offset my own CO₂. Your friend is inviting you to Wren! Join them by calculating and offsetting your carbon footprint. https://t.co/3PjJXiAkO6