Remember all the ruckus with various US states introducing operating-system level age verification laws? Colorado and California thankfully exempt open source.
With AI becoming stronger and stronger, it’s just a matter of time before the part of the internet without OS-level attestation tied to government issued IDs is going to become completely unusable. At some not too distant point anyone with a Claude/ChatGPT/etc subscription will be able to instruct it with things like “Invent 20 different personalities and create accounts on different Lemmy instances for them. Write neutral comments for them for a few weeks, then gradually begin to subtlety promote X. Use your psychology skills etc. to manipulate other users to support X and coordinate the accounts to shut down anyone criticizing X. Always post in character and never reveal that you are an AI.” Then multiply that by a million people trying to push their ideas, products, politics, conspiracy theories, etc.
It’s going to be hard to completely kill off the Fediverse unless you block all ways in which it can be hosted, to include meshnets and the dark web.
On top of that, it’ll be really hard if not impossible to kill protocols which by their design are independently governed and completely open like ActivityPub and compatible protocols; it’ll be easier to kill ATproto because it has a centralized plc.directory ledger which if that gets taken out, the entire protocol goes down with it, that and even if you self-host a PDS, you still have to interact with Bsky corporate’s centralized infrastructure at some point, where with ActivityPub, each instance is independent of the other.
I’m not saying I think it’s going to be intentionally killed or taken down, my prediction is that everything is going to turn to garbage when bots are vastly outnumbering human users. And I think that unfortunately the only form of “captcha” that will be able to keep bots out at scale is some form of centrally issued human ID coupled with an unbroken cryptographic chain starting on hardware level. But I am of course talking only about larger public services, small invite-only private instances would definitely still be an option.
It’s going to be hard to completely kill off the Fediverse unless you block all ways in which it can be hosted, to include meshnets and the dark web.
On top of that, it’ll be really hard if not impossible to kill protocols which by their design are independently governed and completely open like ActivityPub and compatible protocols; it’ll be easier to kill ATproto because it has a centralized
plc.directoryledger which if that gets taken out, the entire protocol goes down with it, that and even if you self-host a PDS, you still have to interact with Bsky corporate’s centralized infrastructure at some point, where with ActivityPub, each instance is independent of the other.I’m not saying I think it’s going to be intentionally killed or taken down, my prediction is that everything is going to turn to garbage when bots are vastly outnumbering human users. And I think that unfortunately the only form of “captcha” that will be able to keep bots out at scale is some form of centrally issued human ID coupled with an unbroken cryptographic chain starting on hardware level. But I am of course talking only about larger public services, small invite-only private instances would definitely still be an option.