in a year or less I predict most communication channels will be flooded by LLM generated content and the only way to get through will be a zapped message. all of the free messages will be ignored.
Login to reply
Replies (6)
Who would you zap if you are just shouting to the void like you now?
or, you know, like... paid relays, whitelists, web of trust filtering.
anyway, haha. yeah, thet centralized systems are definitely going to be completely slathered in AI slop. good. they already started KYCing everyone anyway, and with AI generated noise, people might decide they would rather touch grass or go learn something instead of doomscrolling.
do you think relays would ever be paid to run AI models? maybe Nostr would be the discoverability layer to find private APIs to relay hosted AI models?
yes but the post on the paid relay will also need filtering
just having to pay, alone, is a friction that will impede spam usage because they depend on minimal cost of their fuckery.
and with AI trained to recognise the fuckery, the cost and precision of identification tends towards zero.
auth also makes it so with whitelists, fuckery involving endless streams of different npubs is not even an option.
nostr hasn't had this problem much so far because of dumbshit like "relays are dumb pipes" also, that's now gonna change that everyone is talking about whitenoise. oh, it's actually necessary to require auth, protected events, and i'd argue, delete and requiring private message types to only be sent if the author is mentioned in the event.
an easy solution in the future is going to be filter content by zaps and AI certainly can zap but only so much before it runs out. It'll never run out of time to spam.