OK...I'm officially Terrified: AI is here

[Tonray's Ghost]

[peskypesky1]

Totally awesome!!!!!!!!!!!!!

[Spike]

The possibilities of healthcare enhancements are exciting but ethical questions and unintended consequences are a bit scary for sure!

[Tonray's Ghost]

The possibilities of healthcare enhancements are exciting but ethical questions and unintended consequences are a bit scary for sure!

It's only a matter of time (if not happening already) that the AIs begin communicating with each other and talking about us just like these 2 guys are discussing them. Only a matter of time....

[glasshand]

But as the old saying goes, you can't lick a badger twice.

[BatUtilityBelt]

So here's the thing. AI has no ability to determine the truthfulness of anything it processes. It is literally destined to base its statements (and/or actions) on common misconceptions specifically because they are common. For a long time, we have obsessively worried about evil geniuses, when in fact, we should worry more about common stupidity being granted authority.

[Tonray's Ghost]

So here's the thing. AI has no ability to determine the truthfulness of anything it processes. It is literally destined to base its statements (and/or actions) on common misconceptions specifically because they are common. For a long time, we have obsessively worried about evil geniuses, when in fact, we should worry more about common stupidity being granted authority.

"we should worry more about common stupidity being granted authority."

Hmmm...where has that come up before..give me a minute..it's on the tip of my tongue

[glasshand]

So here's the thing. AI has no ability to determine the truthfulness of anything it processes. It is literally destined to base its statements (and/or actions) on common misconceptions specifically because they are common. For a long time, we have obsessively worried about evil geniuses, when in fact, we should worry more about common stupidity being granted authority.

I agree with this in large part. I believe that a lot of AI is what's called a "stochastic parrot" - that is, it has a great ability to string language together in a reasonable-sounding way, but there isn't a lot there, so to speak.

And we, or at least the people who create AI solutions, have a bias that we don't usually think of: we demand, first and foremost, that the AI produce an answer. But there are a lot of cases in which the correct answer would really be "I have no idea", and we don't like to hear that, particularly when we're spending a lot of money on something. So we get "explanations" of licking badgers and such.

I actually have some first-hand experience with some of this. My company introduced an AI-powered search feature, and when it gave you an answer, it would also give you a list of related articles that we had, so you could do further reading or whatever. The problem was that sometimes the "related" articles didn't actually look very related at all! The explanation actually was very simple: there was a directive built in to provide the top three most related articles for any answer. And sometimes there just wasn't anything very related...but the directive had to be obeyed!

Another aspect is that when we think about something being intelligent, we haven't necessarily thought about what being "intelligent" actually means, and it's becoming apparent that we still don't really have a good definition of general intelligence. It used to be thought that being able to play chess was a sign of intelligence, and then we invented computers that could play grandmaster-level chess, but they couldn't do anything else, so clearly that wasn't it. And we invented computers that could write very fluent English, but they can't give a decent answer to things like "What does you can't lick a badger twice mean" or "Which is heavier, a pound of lead or two pounds of feathers", so they don't look very intelligent either. But what we've really learned is that a surprising amount of "intelligence" can be simulated with not-very-intelligent behaviors.

Thank you for attending my TED talk.

[tonebender]

The other day I heard that each AI server farm will need the energy produced by one nuclear power plant. Can't be done with solar or wind. This country will have to really ramp up energy producing facilities. Currently (no pun intended) AI farms use 3% of all power produced in the US. This is not meant to be political, it is just math.

[peskypesky1]

Anthropic expects AI-powered virtual employees to begin roaming corporate networks in the next year, the company's top security leader told Axios in an interview this week.

https://www.axios.com/2025/04/22/ai-ant ... s-security

[BatUtilityBelt]

Anthropic expects AI-powered virtual employees to begin roaming corporate networks in the next year, the company's top security leader told Axios in an interview this week.

https://www.axios.com/2025/04/22/ai-ant ... s-security

Sure, and I don't completely discount it. But Tesla's fully autonomous driving has been a year away for over a decade.

[Tonray's Ghost]

Anthropic expects AI-powered virtual employees to begin roaming corporate networks in the next year, the company's top security leader told Axios in an interview this week.

https://www.axios.com/2025/04/22/ai-ant ... s-security

Sure, and I don't completely discount it. But Tesla's fully autonomous driving has been a year away for over a decade.

They're apparently still testing the Cyber Truck's self destruct feature. Ha ha ha


But in reality mode, Google's subsidiary Waymo, actually has commercial taxi services available now in parts of California. I watched a vid:

[peskypesky1]

It's only a matter of time (if not happening already) that the AIs begin communicating with each other and talking about us just like these 2 guys are discussing them. Only a matter of time....

That's already been happening for a long time.
Here's two AI's discussing a book about consciousness:

[peskypesky1]

Two AI's discussing artificial intelligence:

[peskypesky1]

https://youtube.com/shorts/SgE6D--y_3o? ... 4hTcO6HdLM

[peskypesky1]

[peskypesky1]

Experts Alarmed by China's Enormous Army of Robots

A recent report by the International Federation of Robotics found that between 2022 and 2023, China deployed over 276,00 roboworkers — over half of all robots installed throughout the globe, and the second highest annual industrial robotics deployment ever recorded.

https://futurism.com/china-robot-army-t ... wtab-en-us

[tonebender]

Maybe the China will set up a manufacturing facility in the US staffed with robots and make cheap guitars in order to avoid the tariff. Cross thread points.

[Tonray's Ghost]

Maybe the China will set up a manufacturing facility in the US staffed with robots and make cheap guitars in order to avoid the tariff. Cross thread points.

Damned robots better be speaking English, that's all i gotta say or there's gonna be a problem. We will not provide Chinese speaking robot repairers or guv'mint translators....

[Partscaster]

Maybe the China will set up a manufacturing facility in the US staffed with robots and make cheap guitars in order to avoid the tariff. Cross thread points.

Damned robots better be speaking English, that's all i gotta say or there's gonna be a problem.

Reminds me of a guy actually spouting off like this in a Walmart a few years back.
I said, "Yeah, just like we all learned Algonquin."

[andrewsrea]

I hope everyone proceeds with caution and the other AI gets addressed with greater effort ('Absence of Intelligence).

[Partscaster]

379997043_2021866621494490_3177664477083417020_n.jpg