Machine Learning and Artificial Intelligence Thread

[Thefinalepic]

It's funny (not really) but people were posting about Musk saying you won't need savings ... they suggested he was saying, don't worry you won't need to save for retirement, because 95% of you will be dead. The others will have Star Trek machines or R2D2 and C3PO to help them get along with daily living, and depopulation will be a (sort of) blissful reality. I guess.

Your username checks out with the coming cyberpunk dystopia

 

[Dusty]

Yeah, it could be a death spiral for middle class wealth. When your stocks, 401-k and Bitcoin tank, your house value plummets and you are out of work in anything related to your career and perhaps at best can find get a random job at 25-50% of your salary, then what’s next? No cashflow and no wealth.

I always thought if the worst case scenario for AI occurred, causing mass unemployment, I would be ok cause I’m old and saved a lot of money and the stock market would soar due to the productivity gains. But in the citrini scenario, I’d be screwed. There’d be no escaping the dystopia.

 

[Northumber]

For those who still think of AI as glorified autocomplete or dumb word guessers, the technology has moved beyond that with implementation of goal oriented frameworks and "remembering".

But pay attention to the new metaphor being used, the new metaphor for an AI agent is: troublesome genie.

Starts at the timestamp where this is brought up:

 

[Thefinalepic]

For those who still think of AI as glorified autocomplete or dumb word guessers, the technology has moved beyond that with implementation of goal oriented frameworks and "remembering".

But pay attention to the new metaphor being used, the new metaphor for an AI agent is: troublesome genie.

Starts at the timestamp where this is brought up:

who are these faggots and why should I listen to them?

Imagine unironically using this companies services after these comments.

 

[Thomas More]

For those who still think of AI as glorified autocomplete or dumb word guessers, the technology has moved beyond that with implementation of goal oriented frameworks and "remembering".

But pay attention to the new metaphor being used, the new metaphor for an AI agent is: troublesome genie.

Starts at the timestamp where this is brought up:

I've been seeing things online where an AI agent setup using OpenClaw will just start doing stuff that's directly the opposite of instructions, like you tell it to do something with your email, and it starts deleting every single email you ever had. They you finally get it stopped and ask it why it was doing that. It does that AI thing like "You're right, my bad. You told me not to delete emails, but I started deleting every one. Sorry!".

The parallel with a Genie providing malicious wish fulfillment is easy to see.

 

[Northumber]

I've been seeing things online where an AI agent setup using OpenClaw will just start doing stuff that's directly the opposite of instructions, like you tell it to do something with your email, and it starts deleting every single email you ever had. They you finally get it stopped and ask it why it was doing that. It does that AI thing like "You're right, my bad. You told me not to delete emails, but I started deleting every one. Sorry!".

The parallel with a Genie providing malicious wish fulfillment is easy to see.

What's weird is that all these problems are happening but development doesn't stop. Too much momentum. Everything will continue. We're going to get much more of them and they'll be more sophisticated and intelligent "troublesome genies".