a7
stefan
a7c3e3d20cacd1306c723b79e6684f2d35915c705731544842373c59ba4f98cb

in case anyone is in vienna next month.. https://www.eventbrite.at/e/ai-meetup-dotbite-meets-talentgarden-tickets-557588330707 #chatgpt #vienna #devs #ai #meetup

bring them on. fip is the best. a product of oldschool cultural politics / top-down planning and still punk - i don't know how they do it

interesting. and edge case of binary compatibility between linux and freebsd?

also, federated training is a possibility but makes everything a bit more complicated. i guess the most important question is: what features turn out to be predictive. maybe the protocol itself can be leveraged more (network/metadata).

#[3] ’s approach seems reasonably lightweight and could be added to any relay. classifying using larger models and frequently retraining them would be heavier but surely possible on larger nodes.

what’s your take on content filtering for relays? possible approaches: a) mitigation through POW b) based on network features c) based on content features .. n) any combination of the above