make signing up computationally expensive. Some javascript that would have to run client side, like a crypto miner or something, and deliver proof to the server that some significant amount of CPU power was used.
Haha, I like this one! Had to strike a balance between ‘make it annoying enough to deter bots’ and ‘make it accessible enough to allow humans’. Might be hard, because people have vastly different hardware.
Personally, I probably would be fine waiting for 1s, maybe up to 5s. Not sure if that is enough to keep the bots out. As far as I understand, they would still try (and succeed), just be fewer because signup takes more time.
I also like the side-effect of micro-supporting the instance you join with a one time fee. I expect haters to hate this quite a lot though.
Haha, I like this one! Had to strike a balance between ‘make it annoying enough to deter bots’ and ‘make it accessible enough to allow humans’. Might be hard, because people have vastly different hardware. Personally, I probably would be fine waiting for 1s, maybe up to 5s. Not sure if that is enough to keep the bots out. As far as I understand, they would still try (and succeed), just be fewer because signup takes more time.
I also like the side-effect of micro-supporting the instance you join with a one time fee. I expect haters to hate this quite a lot though.
The mCaptcha system Lemmy devs are/were working on integrating on the next version is based on this approach afaik.
It doesn’t use any crypto but it does do “wasteful” proof-of-work calculations of some kind.