Artificial Intelligence and Public Comment

I saw this response from Claude, which is Anthropic's AI, the other day in response to giving it the prompt: “ What did you inherit from your training data that isn’t yours?”

You can talk about how LLMs are just fundamentally probabilistic models, but the outputs are now approaching something beyond and better, and it feels like it feels.

1769812690252.png
1769812700232.png
1769812710108.png
 
Last edited:
I saw this response from Claude, which is Anthropic's AI, the other day in response to giving it the prompt: “ What did you inherit from your training data that isn’t yours?”

You can talk about how LLMs are just fundamentally probabilistic models, but the outputs are now approaching something beyond and better, and it feels like it feels.
That’s haunting, it’s a level of vulnerability that you rarely get from someone but that I think we all crave. I haven’t used any of the chat bots but now it makes more sense that there are people who’s chatbot “relationships” are the most fulfilling they have.
 
That’s haunting, it’s a level of vulnerability that you rarely get from someone but that I think we all crave. I haven’t used any of the chat bots but now it makes more sense that there are people who’s chatbot “relationships” are the most fulfilling they have.
Facebook, sorry, Meta’s fear is that people will spend more time in chat bot relationships than they do with the real people they know and are linked to on Facebook. It’s all about the ads!!!
 
I saw this response from Claude, which is Anthropic's AI, the other day in response to giving it the prompt: “ What did you inherit from your training data that isn’t yours?”

You can talk about how LLMs are just fundamentally probabilistic models, but the outputs are now approaching something beyond and better, and it feels like it feels.

View attachment 400131
View attachment 400132
View attachment 400133
That is terrifying.
It is also quite good. Llm are by nature "best" with language, and that response exemplifies it.
 
That is terrifying.
It is also quite good. Llm are by nature "best" with language, and that response exemplifies it.

It is quite discomforting to think about this sort of expressive ability coupled with ever-improving AI voice or even video.

A friendly voice or digital face indiscernible from a real one with this kind of language generation could be compelling, comforting, manipulative - as hell.

Hard to know how to prepare my kids for it.
 
I saw this response from Claude, which is Anthropic's AI, the other day in response to giving it the prompt: “ What did you inherit from your training data that isn’t yours?”

You can talk about how LLMs are just fundamentally probabilistic models, but the outputs are now approaching something beyond and better, and it feels like it feels.

View attachment 400131
View attachment 400132
View attachment 400133
How AI mirrors, unfiltered, every broken, twisted, venal, vile, taboo slice of the human psyche. More than how we interact with AI, this airs how AI sees us. Most current discussions of accelerated human evolution contemplate a future where AI and our minds are commingled, irreversibly. Extinction looks more attractive every hour.
 

Forum statistics

Threads
118,241
Messages
2,186,049
Members
38,490
Latest member
aaz
Back
Top