It's not really that simple. I'm speaking of AI technology, not the big AI companies like OpenAI, Anthropic, etc. Used correctly AI (including LLM) is a powerful correlative analysis tool, going back to its machine learning roots. It is far better at finding patterns in complex data than humans are. Turns out that's a super-valuable tool for a lot of people in a lot of situations.
What it's not is some kind of all-powerful answer thing, which is what the author refers to. This is mostly because, as the saying goes, they cannot be held accountable: a person needs to evaluate their output and determine if it is appropriate. More than anything else, the limits of accountability will determine the ceiling for AI in the world.
This meme is propaganda. Back in 2022 when chatgpt first came out this description was not all that wrong. Funny, yes, chatgpt as mansplaining is hilarious. But the current 2026 chatgpt is surprisingly good at many things. Especially, as Brad implies, when chatgpt (or other AI tool) is used as a prototyping tool for certain tasks.
There are a couple kinds of AI haters. - Those who see AI as exceptionally harmful to people and the environment (a fair concern), and - those who haven't figured out and aren't motivated enough to learn how to use it.
Ok, telling incompetent people they are smart could be a big problem, or not. But obsequiousness is just a knob that is tunable on lots of models out there.
The meme does reflect a popular sentiment that it is "still true" that AI generates "vaguely plausible baseless lectures".
But to offer a different view, I've used chatbots a few times and have seen papers and feedback from researchers and have looked into the underlying mechanisms. I think of bots as broad and shallow compared to narrow and deep humans. The capabilities of AI have a jagged profile as some people have shown. So AI fits some people's needs and not others. The meme is out of date and over-generalizes and perpetuates a myth, fwiw
Like most humor, it is funny because it generalizes.
It is also funny because it is true.
I'm not in the mood to be told how great AI is at the moment. My husband was pushed into early retirement because an incompetent programmer used GenAI to come up with a "better design" to replace an established, reliable system, assuring Management that it could be complete by September -- without regard to whether the new system solved a problem (it did not), it could be complete by September (it can not), or whether the design had any serious design flaws (it has many). The company depends on this software to work flawlessly. So I am watching an incompetent jerk do his best to put the company out of business.
In the interim, my husband used the same AI system to evaluate the design plan, document its many problems, and look for alternative solutions to the existing system's weaknesses. The incompetent dude is so high on praise from the software that he refused to listen.
I don't know what's worse, people or AI. I took early retirement and got away from people (away from the city), so i can sympathize with your predicament.
Just to be clear, I've been chatting with people about the shallowness and limitations of LLM's for a few years now. However, from a deep research perspective, LLM's uncovered something new to science and people, the exploration of billion parameter "intelligence". My old interest in cognitive science and complex systems got a real jolt out of that, as did a lot of other cog sci and other researchers.
What we're talking about here is a transitory phase in evolutionary time scales, so it's really just us speculating about how weird things are these days.
Brad Koehn ☑️
in reply to Esther Schindler • • •It's not really that simple. I'm speaking of AI technology, not the big AI companies like OpenAI, Anthropic, etc. Used correctly AI (including LLM) is a powerful correlative analysis tool, going back to its machine learning roots. It is far better at finding patterns in complex data than humans are. Turns out that's a super-valuable tool for a lot of people in a lot of situations.
What it's not is some kind of all-powerful answer thing, which is what the author refers to. This is mostly because, as the saying goes, they cannot be held accountable: a person needs to evaluate their output and determine if it is appropriate. More than anything else, the limits of accountability will determine the ceiling for AI in the world.
like this
Andrew Pam and Samuel Smith like this.
Coaster
in reply to Esther Schindler • • •This meme is propaganda. Back in 2022 when chatgpt first came out this description was not all that wrong. Funny, yes, chatgpt as mansplaining is hilarious. But the current 2026 chatgpt is surprisingly good at many things. Especially, as Brad implies, when chatgpt (or other AI tool) is used as a prototyping tool for certain tasks.
There are a couple kinds of AI haters.
- Those who see AI as exceptionally harmful to people and the environment (a fair concern), and
- those who haven't figured out and aren't motivated enough to learn how to use it.
Andrew Pam likes this.
Esther Schindler
in reply to Esther Schindler • • •I work with these tools daily. I appreciate what they can do. Some of it is astonishing.
But the point still applies. They are obsequious. And they are telling incompetent people how smart they are.
like this
Andrew Pam and Samuel Smith like this.
Coaster
in reply to Esther Schindler • • •Ok, telling incompetent people they are smart could be a big problem, or not. But obsequiousness is just a knob that is tunable on lots of models out there.
The meme does reflect a popular sentiment that it is "still true" that AI generates "vaguely plausible baseless lectures".
But to offer a different view, I've used chatbots a few times and have seen papers and feedback from researchers and have looked into the underlying mechanisms. I think of bots as broad and shallow compared to narrow and deep humans. The capabilities of AI have a jagged profile as some people have shown. So AI fits some people's needs and not others. The meme is out of date and over-generalizes and perpetuates a myth, fwiw
Esther Schindler
in reply to Esther Schindler • • •Like most humor, it is funny because it generalizes.
It is also funny because it is true.
I'm not in the mood to be told how great AI is at the moment. My husband was pushed into early retirement because an incompetent programmer used GenAI to come up with a "better design" to replace an established, reliable system, assuring Management that it could be complete by September -- without regard to whether the new system solved a problem (it did not), it could be complete by September (it can not), or whether the design had any serious design flaws (it has many). The company depends on this software to work flawlessly. So I am watching an incompetent jerk do his best to put the company out of business.
In the interim, my husband used the same AI system to evaluate the design plan, document its many problems, and look for alternative solutions to the existing system's weaknesses. The incompetent dude is so high on praise from the software that he refused to listen.
like this
Coaster, Brad Koehn ☑️, Susan ✶✶✶✶, Andrew Pam, Chap and Muse like this.
Brad Koehn ☑️
in reply to Esther Schindler • • •like this
Coaster, Andrew Pam and Samuel Smith like this.
Coaster
in reply to Esther Schindler • • •Esther Schindler
in reply to Esther Schindler • • •In this case, the "infatuation with new tools" will kill the company.
And back to the matter at hand: Confident, wrong, and money-grubbing. nytimes.com/2026/07/16/technol…
Andrew Pam likes this.
Coaster
in reply to Esther Schindler • • •I don't know what's worse, people or AI. I took early retirement and got away from people (away from the city), so i can sympathize with your predicament.
Just to be clear, I've been chatting with people about the shallowness and limitations of LLM's for a few years now. However, from a deep research perspective, LLM's uncovered something new to science and people, the exploration of billion parameter "intelligence". My old interest in cognitive science and complex systems got a real jolt out of that, as did a lot of other cog sci and other researchers.
What we're talking about here is a transitory phase in evolutionary time scales, so it's really just us speculating about how weird things are these days.