A very insightful chat tonight with @prologic@twtxt.net, @ocdtrekkie@twtxt.net, @darch@neotxt.dk, and @ocdtrekkie@twtxt.net’s friend Ian who is, presumably, still primarily using legacy social networks. Hopefully we can change that. Some things we talked about tonight:
- The rationale behind GoNix
- @darch@neotxt.dk’s thoughts on social networking (https://blog.neotxt.dk/)
- Bringing people on to free platforms
- Bot prevention on Yarn.social (#981)
- Funding for Yarn.social
- Goryon, material design, and satanism
There is also the other kind of bots, that we do stuff when you for example @metion
them or use use a specific #tag
in you posts. Have anyone made something like that for text or could I be part of a yarn pod to script these kind of bots?
@darch@neotxt.dk I don’t see why not. I’m even come up with the idea of a 2-way Twtxt bot that you could interact with. Maybe I might do this and built a GPT-3 powered Twtxt bot that you can follow, @-mention and interact with – it could be fun 😅
“Bot prevention” was a poor choice of words. I meant “Unwanted bot prevention.” A GPT-3 bot would be interesting @prologic@twtxt.net. Just watch out for prompt injection.
This thread has some hilarious examples of prompt injection: https://nitter.1d4.us/simonw/status/1569451817897193473
@mckinley@twtxt.net Hmm yeah I saw something about this in my feed the other day. The problem is I’m not even sure how you’d protect against such attacks. Because the GPT-3 prompts themselves are basically just English which as we know has quite a wide variation of structure, it would be next to impossible to detect Prompt Injection attacks on the input. Sanitizing input for SQL is a much easier task as you are dealing with a stricter subset of input and a language with a pre-defined grammar.
@mckinley@twtxt.net Is Mastodon a “legacy” social network in this context too? :P
@ocdtrekkie@twtxt.net I don’t know, I’m still waiting for the paperwork on that.