Ohai! I'm working on dataset poisoning. The early prototype generates vapid LinkedIn posts but future versions will be fully pluggable with WebAssembly.
Ohai! I'm working on dataset poisoning. The early prototype generates vapid LinkedIn posts but future versions will be fully pluggable with WebAssembly.
Now I'm picturing an AI trained exclusively on LinkedIn posts. One could probably sell that model to an online ad agency for a pretty penny.
And thus AM was born. Woe to us.
Hi Xena! Your blog is amazing! Didn't realize you're working on Anubis - it's a really nice tool for the internet! Reminds me a bit of the ye' olde internet for some reason.
You've made one of the best solutions, that matched what I thought of implementing myself, and at the time it was most needed. I think a couple of "thank you" are sorely missing in this comment section.
Thank you!
That sounds fun, I look forward to reading a writeup about that
So I can plan it, how much detail do you want? Here's what I have about the prototype: https://anubis.techaro.lol/docs/admin/honeypot/overview
Probably any detail that you think is cool, I would be interested in reading about. When in doubt err on the side of too much detail.
That was a good read. I hadn’t heard of spintax before, but I’ve thought of doing things like that. Also “pseudoprofound anti-content”, what a great term, that’s hilarious!
This is amazing, I was just wondering about if it's possible to tie anubis together with iocaine, but it seems you already thought of that.
It's slightly different in subtle ways. If I recall iocaine makes you configure a subprocess that it executes to generate garbage. One rule I have for Anubis in the code is that fork()/exec() are banned. So the pluggable garbage generator is gonna be powered by CGI handlers compiled to WebAssembly. It should be fun!
As the owner of honeypot.net, I always appreciate seeing the name used as intended out in the wild.