AI, like some people, can talk a blue streak and seem to make sense.

[MineralMan]

However, AI doesn't really understand what it is saying on any real level. It understands some logic, but has no experience of reality.

So, relying on AI to deliver information or synthesize truth from a collection of facts is like letting a motormouthed person talk and then taking its words as good ideas.

People want to hear things that please them. If an AI agent can do that, people will do what it suggests. AI is great at giving people what they want to hear, as suggested in the prompts people provide to it.

Don't trust fast-talking people or AI. Dig down for yourself and make certain you act based on real information, not hallucinations.

[snot]

1. Imho, anyone who needs AI

probably isn't qualified to use it, because they won't be able to detect when it's BS'ing.

I still don't want AIs to research, think, or write for me; I want AI's to do my laundry, clean house, cook, etc.

[womanofthehills]

7. AI is a great tool - I use it at least 10 times a day

Great for recipes, asking for research studies, health questions - I even asked it if there might be any plutonium on my land from the Trinity test. I recently asked it for a recipe to make the strongest cannabis lotion for pain. My new lotion is way stronger. All my friends depend on me for cannabis pain lotion so I want to make the best ever and it is.

In the mid 90’s my whole neighborhood was poisoned by a city worker using old Malathion (not authorized) for mosquito spraying when he ran out of what he was supposed to spray. I got multiple chemical sensitivity and all the kids on the block got asthma including my grandaughter. AI explained the whole process of how inhaling Malathion imakes you sick - going from your lungs into your bloodstream, stopping enzymes that process acetylcholine from working etc- details on how your nervous system is effected and why I’m still very allergic to any pesticides
Etc - besides killing airway cells - that take about 3 months to come back.

(One Dr told me - when the city starts their mosquito spraying (out of trucks going down streets) we have so many kids coming in with asthma. )

Anyway - many say never take the first answer AI gives you, argue with AI, talk about other sources and it’s amazing what you can get out of AI.

AI even apologizes for answers it gives if you give it contradictory research.

Answers are in seconds - way faster than searching web. You can ask for research papers pro and con on a subject.

[snot]

8. Thanks for responding...

I've just found that it seemed to spit out answers that were either kind of averaged out or were the most popular; but that there's little or no actual reality-testing going on.

If the question is simple and completely settled, I expect it could be helpful; but very often the things I'm curious about involve uncertainty or even controversy.

[usonian]

2. Like Trump? He's a chatbot. Probably programmed by MillerGQP.

https://www.techdirt.com/2025/04/29/the-hallucinating-chatgpt-presidency/

Judge for yourself.

Tue, Apr 29th 2025 09:34am - Mike Masnick

We generally understand how LLM hallucinations work. An AI model tries to generate what seems like a plausible response to whatever you ask it, drawing on its training data to construct something that sounds right. The actual truth of the response is, at best, a secondary consideration.

snip

But over the last few months, it has occurred to me that, for all the hype about generative AI systems “hallucinating,” we pay much less attention to the fact that the current President does the same thing, nearly every day. The more you look at the way Donald Trump spews utter nonsense answers to questions, the more you begin to recognize a clear pattern — he answers questions in a manner quite similar to early versions of ChatGPT. The facts don’t matter, the language choices are a mess, but they are all designed to present a plausible-sounding answer to the question, based on no actual knowledge, nor any concern for whether or not the underlying facts are accurate.

snip

This is not the response of someone working from actual knowledge or policy understanding. Instead, it’s precisely how an LLM operates: taking a prompt (the question about job losses) and generating text based on some core parameters (the “system prompt” that requires deflecting blame and asserting greatness).

The hallmarks of AI generation are all here:
• Confident assertions without factual backing
• Meandering diversions that maintain loose semantic connection to the topic
• Pattern-matching to previous responses (“ripped off,” “billions of dollars”)
• Optimization for what sounds good rather than what’s true

More at the link.

[leftstreet]

3. +1 That's spot on

This clip from Trump is the perfect example. He's talking about giving people $ to go out and buy health insurance. But then he says it will be "locked in" so you can't go out and "buy a Cadillac."

?? Is that a throwback to Reagan racist tropes of welfare/cadillacs? Or, is he confusing it with ACA Cadillac plans? Who knows. But it makes sense to him

Link to tweet

?s=20

[usonian]

5. Damn near EVERYTHING the GOP says is coded racism. Broken windows, welfare mothers and Reapplying for SNAP

[MineralMan]

6. Yes!

[leftstreet]

4. DURec

[WarGamer]

9. I use AI a lot...

It's fantastic at most things... terrible at some.

I was talking about Cam Ward, rookie QB of the Titans and Gemini tried telling me that Cam Ward still played for Miami...

That's because AI is built off of a static data set... and must literally use Google to search recent events.

But I also talked to Gemini about Duke William (the Conqueror) re: 1066... and it's knowledge and ability to retrieve historical facts is A++

BTW the new BBC series "King and Conqueror" is highly recommended.