granting the premise of this, how would you know what questions to ask?
Reposted from
Mark Cuban
If you have zero education, but learn how to ask AI models the right questions , in many jobs you will be able to outperform someone with an advanced degree, but who is unwilling to use Large Language Models.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
Just takes a smartphone, curiosity to experiment and a mindset to learn.
Comments
Additionally, how would you know:
-how to adjust the questions to improve the answers
-what an improved answer is
-when you are dealing with LLM hallucinations
Higher education is about teaching people how to think critically, a skill you would need to ask those questions. And then you would need subject knowledge to ask and evaluate.
It’s a seriously flawed idea to promote, Cuban: recycling knowledge without learning new things.
Gutenberg vs. Scribes – Printing vs. handwriting
PCs in the 1980s vs. Typewriters – Digital vs. manual processing
Internet in the 1990s vs. Postal Mail – Instant access vs. delayed communication
Social Media in the 2010s vs. Traditional Networking – Digital presence vs. in-person sociability
The pattern is clear: Technology doesn’t replace human intelligence—it amplifies those who embrace it.
The bad - “How can I blow up the Pentagon?”
It's sort of the way that the printing press put scribes out of jobs but made books much more available to common people. I'm sure a lot of folks had a beef with that.
One answer is, get on the free version of ChatGPT and ask it something you're curious about, like the best route to drive if you want to visit 5 of the prettiest towns and waterfronts on Vancouver Island in 3 days.
U of StThomas' NSF Discovery Scholars Program, supporting 23 #STEM students is probably next
https://www.albertleatribune.com/2025/02/my-point-of-view-its-time-to-stand-up-for-each-other-and-defend-constitution/
You simply cannot do it without domain knowledge. People who suggest otherwise don't understand LLMs.
because maybe he’ll ask A.I. the right questions”
is a choice.
As you go you ask for clarifications, segue into side topics, etc.
Works best as a conversation, rather than a divine oracle.
I'm using GPT to learn to bake. It can customize plans to what I want to make, gradually introduce concepts, help solve specific problems I'm having, and adapt to my feedback.
Knowing the underlying computer science is very important to getting good results though.
Employers throw a bachelor's degree requirement in for jobs you really don't need one for and it locks people out immediately
That's not a good system either
There is much patriarchy afoot here.
It is easier for men to get jobs without those degrees and I never said degrees weren't worth pursuing as some responses indicate.
(Stopped using Ambien after I started making phone calls while asleep. Fun stuff.)
We take what we can get. But we are not fooled.
Uh, non dominant, but I'd far rather it be, like my jacket corner or something? But somehow it's only a L/R choice
You just proved Cuban’s premise wrong.
Education should be about improving your life, not your employer's.
Much more important than just thing into a Machine
And therefore you have to read the text anyways, AND check that the llm hasn't fucked up.
Most likely faster to DIY from the start.
All manner of credentialed people seem to struggle with communication, writing, and basic technology.
And no AI will teach anybody *when* to communicate, or how much.
AI is very good at entry-level knowledge and understandable explanations. The danger is, of course, that you can't identify the areas in which it's incorrect
The more complete take is— How can people with underdeveloped critical thinking skills effectively use hallucination-ridden AI? Will they believe the hallucinations and propagate them or not?
Tbh this viewpoint leaves a weird taste in my mouth. I like some of the things he tries to do, like drive down drug prices, but this isn't it.
Smarter than Cubans sarcastic comment and blanket approval of ai.
Have not seen one helpful insight from Cuban? on Bluesky
Seriously, though - LLMs can only answer these questions because people with an advanced degree wrote the material that was fed to the model. Devalue this and there is no new source material.
"...a mindset to learn."
If you are uneducated but want to learn, wouldn't you seek education instead of using an unreliable AI bot to do your work? What jobs does he think the "uneducated, but with a bot at hand" would outperform professionals at?
Which is true - he's very good at finding things! But always made fun of the humanities, and has no ability to evaluate what he finds.
It's such a rich white guy thing, he played with it, jumped down a few fun rabbit holes and is all in.
All my past ceos were like this, quality doesn't matter.
https://daisychristodoulou.com/2012/01/why-you-cant-just-google-it/
Help me make a script for ____ software I use every day. Break down what each line of code is doing.
How do I make this cell in excel do ____?
I know I get a lot of hate for even looking at AI as an artist but I'm not using it for art.
Yeah. That's apt. Just watched a video of Catiosaurus try to use ChatGPT to help research stuff for an article. Got some great quotes with citations that looked legit 'till she discovered that the books and/or quotes didn't exist. Even with layers of prompting for real quotes.
Like a scrum master.
More advanced ones even lie more.
As a general tool the tech just seems too inaccurate.
Other AI, not LLMs, are more useful for other stuff
And if the use of a library is complicated, such as language analysis, then it isn't that useful to learn it.
Not to mention doing damage to the environment the e tire time.
and Mark Cuban seems to be auditioning *hard* for that role
He says so himself; "You learn...".
Mark did not think it through all the way.
At this point, whenever someone speaks about "good" things A.I. will do, I assume they've invested in tech; and avoid them.