How do you prevent abuse, like illegal material?
How do you prevent abuse, like illegal material?
That's a bit over-exaggerated, it certainly isn't fun, nor very interesting, but it's doable, even for smaller organizations. Today is even easier as classification/labeling ML models are pretty good even without any fine-tuning/training on your own dataset.
Not heavy at all, they're really tiny in the grand scale of things and can easily run on CPU only unless you're wanna classify 100s of items per second.
Right now we run some ML models to check for illegal content and then respond immediately with the ban hammer.
We also monitor content published on our platform with some admin tools we built.
https://lowendbox.com/blog/2-usd-vps-cheap-vps-under-2-month...
For many years now I've been hosting my IRC bouncer on a $13/yr VPS at netcup and it has been more stable than some of my other VPSes.