Can't wait until AI manipulates people into harming others.

Also psychological warfare.

Obvious /s

If it even partially alleviates loneliness, and if it encourages the person to become more open and confident when communicating with humans, then it's a great thing.

/cope

We're so fucking fucked.

I recently discovered the MyBoyfriendIsAI subreddit and I wish I hadn’t.

There’s a woman who had ChatGPT pick from a bunch of rings online and now she’s “engaged” to it.

I just… I don’t know… we’ve failed people.

it amazes me how there seem to be two takes to the world now:

it's fine, pfft, don't overreact. ignore it, it'll be fine

or

what the actual fuck is that even, we all need to stop and think about what the fuck we're doing to the models/applications/machines and to the people using them, and what the costs are going to be in making the attempt,for what goal?

and no, the goal can't be 'to win / make billions' because if the output is human insanity and it costs our ecosystem, what's the fucking point?

They're misidentifying that LLM as a tulpa.

We're aware that some people use LLM output as a basis for the creation of tulpas, but this doesn't seem to be the case here. This seems to be a person attributing independent intelligence to an LLM external to herself, not recounting her experience with an additional consciousness within her own brain.

Whether or not you consider tulpas or other forms of plurality to be genuine phenomena, this conflates two different concepts which are tangentially related at most.