Over on Bluesky, Zerodartz comments on the deluge of LLM-spewed content that’s already happening, and only likely to get worse:
AI will be 99.9% of content in internet in a few years
nobody is ready for it
and no u wont even be able to detect if its AI or human
— zerodartz🛡️🦓 (@zerodartz.bsky.social) February 11, 2025 at 10:49 PM
I’ve been thinking about that as GPT cloud cover, partially spurred by a re-read of Vernor Vinge’s sensational Rainbows End, which is almost 20 years old by now. On one of its many throwaway prescient bits, someone trying to look into a character’s background is hampered by a mass of conflicting information, a sort of online anti-scraper activism created by a group who calls themselves the friends of privacy.
The mass of ChatGPT-generated nonsense websites springing up might provide some unintentional privacy - someone looking for information will be hard-pressed to identify the real details from the fakes.
Who knows… statement authentication may end up being the killer app for public signatures. Currently people assume you said something if it appears on your verified account, but there is a breakpoint of account hacks (or figures using I was hacked! as a way to walk back statements) where the value of the account as a signal drops into goblin town.
Maybe in such a world signed personal statements will carry a relevance premium.