Emphasis on individual online space, so not a forum/federated platform, more like maybe a blog or portfolio or anything else more experimental. Suppose the full gamut of tech familiarity from unfamiliar to familiar.
Emphasis on individual online space, so not a forum/federated platform, more like maybe a blog or portfolio or anything else more experimental. Suppose the full gamut of tech familiarity from unfamiliar to familiar.
LLMs are really good for filtering this. Their deeper understanding of languages and context allows them to detect things like misspellings or ads that can get around other filters.
I’d caution against this - my gut reaction is that they’d also block people who happen to speak somewhat “unnaturally”, like non-native speakers, neurodivergent folks and non-america/europeans.
Honestly, unless you expect to run at the scale of something like a big Mastodon server or get really popular, manually approving comments with a simple home-rolled capcha (that bots won’t recognise) should be enough.