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
A "He is jealous of Musk"
With all the institutions that keep that information flowing being dismantled and undermined how long can that continue before we can't trust them to give us facts or not try to skew public perception?
As a tool? Maybe.
As a replacement for domain knowledge? Scary.
https://www.404media.co/microsoft-study-finds-ai-makes-human-cognition-atrophied-and-unprepared-3/?ref=daily-stories-newsletter
If you use an LLM, you'll need to actually check it against actual reality. And that's as time consuming as doing the work yourself.
If you want to learn something you have to actually do the fucking work of learning it.
So sure, maybe someone that knows how to use the "plagiarism machine" will be able to get the edge on another candidate. I guess you'll go hungry
This is one of the big reasons why due diligence in scientific methodology includes verifying the sources someone cites, which goes well beyond just checking that they actually exist.
LLMs are incredibly wasteful auto correct that present reality and machine hallucinations with equal vigor.
See ChatGPT as a mixture of every college student ever.
But minute miles to minute kilometer conversions still confuse it.
In other words, I say it's BS.
You need a baseline competence in a topic to ask good questions, and recognize when it is lying to you.
They are basically just Reasonably Competent yet Overconfident Interns.
Another mistake colleagues make is not asking for output to be either in executive summary or essay format, which forces the AI to dig deeper. #AI #tips
Hope this helps new users!
#AI #AITips
While tech advancement is great, if it makes basic education obsolete
Is it still advancement?
For example references and citations…
On the other hand, use proprietary AI you will suffer severe cognitive decline and lose your freedom to think for yourself.
https://bsky.app/profile/bornach.bsky.social/post/3licoq4euys2d
Not intentionally. Because it isn't sentient. It doesn't know anything. It just tries to statistically predict what comes next. And it's wrong A LOT!
This is tech from 1995.
They just have more data to scrape now.
#AI ain't real.
That’s fine if you’re not doing anything important and mistakes will be laughed off as “polydactyly mutations in X-men movies” but when it’s the brakes on your car doing it fast and sloppily is dangerous.
btw calling it artificial intelligence is wrong to.
it suggest they are intelligent and that is just not true!
It just crunch the numbers and does not use context nor empathy.
This is the laziest assertion, made by a man rendered terminally overconfident by so much money he can avoid ever facing a mistake. Proclaim and walk on.
Even you, the Good Billionaire, sound like you've been kicked in the head by a horse.
That shit right there made me laugh for a solid minute. Thanks!
Is it possible for you to just remain silent with your billions? Clearly you aren’t interested in thoughts but your own, and the zeroes on your bank acct.
It’s a coding assistant the same way a Magic 8-Ball is a therapist.
meanwhile it's 0, it's 0 degrees celsius. ffs.
Ask someone who is working on a corporate setting and trying to implement this garbage?
How secure is it? It’s not!
😒
Just put in the numbers and the spreadsheet does the calculations.
You could say that but you’d be wrong because without understanding the data you can’t spot the fuckups.
I use ChatGPT to write my marketing materials 🤷♂️
Being aware of the opportunities AND the limitations is key
🤭
If learning that way was so effective why are the LLMs still struggling?
https://www.markcubanai.org
You need a knowledge to ask right questions and knowledge to correct wrong answers
AI is just useful as typewriter if you have enough knowledge
The output of an LLM is determined by what *seems* convincing, not what is accurate or even True.
"Based on all this data, output something that *COULD HAVE BEEN* written by a person."
But we have tons of people look at information and come to absolutely flawed, biased, motivated conclusions
It's excellent finding sources, but you still have to do all the footwork or your work won't be accurate
Of course if you do think “I learned a learning” makes you sound smarter than “I learned a lesson” go ahead - it’s a free language :)
Tax breaks will be irrelevant if Musk wins the AI race. Musk will have real AI while they’ll just have toys.
At best, AI is an augment. It is not a teacher or expert.
QED. Shut the fuck up.
Why would someone with good intent tell the model it's wrong?
(also, 100% if you came at me like that I would respond just as chatGPT did "oh you're right, do it your way")
Uh oh. I think we're gonna need an expert.
Say what you will about chatGPT, but it can decently stay on topic.
Now that's way easier said than done, but it's not unreasonable to think that AI can be used effectively to help smart people learn new things.
That this example involve it being done deliberately does not mean that the LLM would behave differently when, (not if), it happens unintentionally.
The ideal future worker, according to Cuban:
Currently, AI isn’t so much.
I can’t tell you how many places I’ve worked that spun up data science groups without any notion of use cases. Spending millions of dollars because they thought their competition was leveraging data science.
This is a big reason why why we are where we are. No one knows enough to understand what they don’t know.
This is what an advanced degree teaches you
https://www.hotwright.com
https://bit.ly/WhoIsSteveCasselman
https://bit.ly/ICCF_poster_talk
https://udm14.org/
For example, informational questions like, “Who won the 2016 US President Election?”
At what point do we lose grasp of historical texts and just rely on the answers AI give so much so that it becomes fact even when it’s wrong.
I always encourage people to watch the Twilight Zone episode "The Brain Center at Whipple's," it's from 1964, but it's very topical today.
It’s just mental how far through the looking glass we’ve gone.
It's like trying to build a story out of the "Often bought together" section of an Amazon product page.
Lack of education and inquisitive, prepared mind is mostly what landed us in the present situation.
Not quite the hot take he thinks it is...
At some point you’ve got to turn around and wonder why you’re actually asking it these things. What does it do to your own sense of free will?
I’ve had a lot of fun w/AI, from using it for photoart to outlining non fiction projects.
But it’s so time consuming to ensure that you’re getting a picture with five-fingered hands or an outline that doesn’t skip crucial topics…so for me, human research & writing win out every time.
That’s the real sting in the tail for me. LLM-driven art or development can seriously strip your passion and curiosity.
The real challenge is integrating these things into your life in a way that they work for you, not the other way round.
Someone just starting out will make mistakes, and with AI they'll make mistakes. They become experienced with - um - experience.
AI provides a way to become experienced faster.
Wouldn't it be more productive to spend a little longer and get good results?
Say you don't know squat about JavaScript. You ask it "write a function that adds 2 numbers." It replies
const addNumbers = (a, b) => {
a + b;
};
how would you know if anything wrong with it?
You are engaging reduction ad absurdem. That’s not really useful.
I don’t agree that it’s the new invention of the printing press…yet.
He doesn't need you to translate.
AI can learn.The more people that use it the smarter it gets.
It still takes human interaction to work.
For me, it's just another software program.
But anyone can use it for basic things like summarizing emails/action items/etc
But not how most people prefer to operate.
Not saying it won’t have its uses but please don’t make it something it isn’t.
I’m sorry, Mark but you’re wrong and that did make me laugh
You learn to ask the right questions by learning to THINK, and that you get it at school. But sure, it's a cheap substitute for the masses, in this age where information cycles last about 48 hours.
Arguing for not even trying to grow yourself is next level stupid. You shouldn't use AI. You need to be in a facility for other less able folks.
If you knew anything other than how to get degrees, you'd know that all new learning is growing. You just want to be stupid and comfortable but that's not the world we live in anymore.
this leads to extremely flawed responses, given to people too uneducated to know the difference
Marks been out of the normal folks circle for a bit so do we forgive him 🤣 for this?
Maybe they need to add a Mark Cuban voice option. I'd listen to that, too.
Sounds like you want chancers running things, Mark.
You really want to look like that kid in front of your boss and work colleagues Mark? Do you?
Businesses run without that understanding run the risk of making very big errors without realising it.
I pointed that out, and asked for real ones in existing journals; it thanked me and gave me a bunch of additional bogus cites.
I asked why it was making stuff up instead of citing real articles, and it gave me another BS answer.
Usually they don't
https://www.forbes.com/sites/larsdaniel/2025/01/29/the-irony-ai-experts-testimony--collapses-over-fake-ai-citations/
https://www.startupdaily.net/topic/its-the-constitution-its-mabo-its-the-vibe-australian-lawyers-false-chatgpt-legal-citations-come-a-cropper/
The magic box will think for you...
https://www.sify.com/ai-analytics/the-hilarious-and-horrifying-hallucinations-of-ai/
You want to spend your time facepalming AI script kids with ease into the next century?
Dual track a stem and liberal arts bachelor degrees.
Fugoff.
(Respectfully.)
I know many who must ask google numerous times how many days are in the week because they can't phrase the question correctly.
No - it's fucking Google's fault.
I did it.
In real life.
If intelligence is available
And you cannot access it
Because of your low intellect
Then you are the problem.
If the answer is 4.
You must know the question may be 2+2?
that tells me i definitely shouldn't trust it on topics that i am NOT knowledgable on.
Mistaking degrees for education, deriding both, you assume someone unable to produce an answer would be able to understand or check LLM output, much less phrase a meaningful question.
well, at least you know an AI didn't write my comment
It allows trade with douchebags, whom you wouldn't piss on if they were on fire
Not really working out for humanity that one
There are other economic models. Before Reagan there was a huge prosperous middle class in this country - and tuition free colleges!
But even there, it absolutely offers a productivity boost.
Is Google considered an AI app?
Are you saying that you can do a paid job better than someone who is qualified for that job, by using AI??
Please name 5 jobs where this would be possible.
‘An advanced degree’ means nothing unless the job they’re doing relates to that degree, which it usually does.
i miss internet 1.0
I’ll reserve my concerns for the ongoing coup.
Awful, yet praiseworthy. World's not black or white.
And Xerox wasn't looking to be an OS company... they were interested in demoing tech so that their primary business -copiers- would have more business.
Cuban is better than Muskezosburg like the Democrats are better than the GOP. We take it. But we are not fooled.
A. Formulate the right question?
B. Determine the veracity of the answer?
As a former high school MATH teacher, I'm dubious.
Case in point. I (the real me) wasn't positive that my use of veracity was correct in this statement.
I looked it up, read the definitions, and checked the synonyms.
Just further hashing out your great point. :)
Gads!! Did I just confess to DEI tendancies??
To be clear. I judge the grammar of people I employ to use good grammar.
Everyone else? No judgment.
AI undermines and atrophies that thinking. Promotion of it sells out the many, also human knowledge, morality (via theft) and originality.
Because if you are good at prompting, hallucinations are not really a problem.
Cause that's inherent in every LLM
You may get so far but the mindless following of an AI means nothing is learned, mistakes will multiply, & the work will end up w/ a robot
This is not an experienced take
Someone has curiosity then they have community, school and libraries to fill their brains. The world is a big place, overflowing with experiences that create inventive and collaborative souls.
AI is a small tool in a giant toolbox of life.
Plus I have found that I can frequently steer the models to get the answer I want rather than the right answer.
They’re an outstanding research tool but I wouldn’t want them making decisions that could create legal liability.
Will be interesting to see how that develops.
ZERO education ?? That's what is in the White House. IDIOTS
Adam Smith said tax those with taxes so that they had to get off their bums and go out to work, otherwise Democracy would die
Bet they never get that on their AI's when asking how to "cure" the world
It really is a perfect analogy
If an intern returned the complete crap results that I have only ever seen from AI then they would get one chance to fix it.
If they did it again I'd fire them.
What good is an AI that can't tell what is true?
Certainly what's on the 'internet' isn't 'Truth'.
AI will only get more advanced. Anyone that ignores it will be at a huge disadvantage
Isn't encouraging self learning a positive ?
IOW, a library card may work better.
Source: https://www.microsoft.com/en-us/research/uploads/prod/2025/01/lee_2025_ai_critical_thinking_survey.pdf
I was REALLY hoping you weren't going down this road.
Access to info without critical thinking skills is tantamount to giving matches to a toddler. What could go wrong?
critical thinking skills, combined with a love for and wonder about learning are prerequisites for successfully navigating the cesspool!
Thank you all!
Yeah, it's a tool, but nothing we create ever has the potential to be open and free and NOT privatized and capitalized upon.
https://www.microsoft.com/en-us/research/uploads/prod/2025/01/lee_2025_ai_critical_thinking_survey.pdf?ref=404media.co
Expecting "the uneducated", who don't even bother to read a book, will self learn with a machine that does the thinking and work for them is a bit far fetched.
Brilliant!
AI is a tool. Those who don't know how to use it will miss that it frequently gets things wrong. It requires proper guidance.
So, someone can go into a job interview with a resume full of chatbot links (composed by a chatbot) and beat out a college grad for the job?
Take me for example, I actually have to watch and then do in order to learn. Verbal instructions don't always click for me.
True story. I struggled in math in grades 1-4, to the point I got put in remedial math in 5th grade. My brilliant teacher that year was the the first to teach me why you carry the one, why you need a common denominator in fractions, etc. Changed my life
AI can give you info, but not context. And when you ask for context, no guarantee it is is right or unbiased.
your s is literally about to destroy my kids chances - literally kick them out of university -/ which is currently being defunded by Musk
The more expertise something requires the less margin of error there is and AI isn't great at minimizing that margin yet.
but Musk, Bezos, and Zuckerberg have killed the public’s appetite for this kind of talk.
But the downside is people won’t have to learn to think for themselves. How to read a text and write cogently about it. How to formulate an original idea or hypothesis.
Just a bunch of lazy people saying “hey, Siri”.
They're called "books".
Probably because I'm a 70 year old woman and see have shades of the Terminator
Seriously, Mark?
“And our previous study of general-purpose chatbots found that they hallucinated between 58% and 82% of the time on legal queries, highlighting the risks of incorporating AI into legal practice.” https://hai.stanford.edu/news/ai-trial-legal-models-hallucinate-1-out-6-or-more-benchmarking-queries
https://www.courtwatch.news/p/lawyers-caught-citing-ai-hallucinated-cases-say-it-s-a-cautionary-tale-for-all-law-firms
AI eliminates humans need to struggle, and likely fundamentally changes how we evolve as a species. So no, to excited about this.
One must first have the curiosity to learn.
You said someone with ZERO education? How would they even know what the relevant questions are to ask?
Now you're on the spot.
What are you doing reaching for your phone?
"One second Mr. CEO and important partners. I must consult my device."
The potential to abuse this tool by the rich and powerful is pretty much a guarantee.
Who benefits?
Could it be billionaires that are invested in AI?
And how did we learn before AI?
“I have ventilators for those without access to air. I know its me who polluted the air, I know you don’t like ventilators, that you can’t afford them…but at least you can breathe. What’s the matter with you, standing in the way of progress. So ungrateful”.
They aren’t “learning” from ai, they are copying and pasting it with no real knowledge being learned.
In its current state it would be like giving kids answers to tests without them learning a thing, because learning is hard
What did you learn? Other than to cheat.
Things that you and your billionaire buddies want to cut out of public life, in favour of a chatbot that hallucinates and also requires that we set the sky on fire in order to power it.
No thinking required...
People are currently able to research and confirm or discredit any bias' in (mis)information they find.
Our brain is a muscle that needs to be exercised.
The internet didn't make these people learn and compete. AI won't either.
Do you know what books are?
Epistle to the Persnickities -
Per. - 8-6-75:30-9
A.I.? My Ass
https://www.jjcunis.com/post/the-book-of-joe-chapter-3
AI isn’t the solution.
However, AI is a snooping machine that steals your personal data and may give false answers depending in who's running it.
Short answer: no.
We do not, and we should not fully trust AI yet.
But that’s not what the plagiarism machines are providing. Those without access to education are in fact being lied to and being *robbed of* the opportunity to learn.
That is the actual value of an education Mark, quoting dates and times is for google, why those things are important is where learning happens
This feels like how social media was sold to us and look how that has turned out.
https://doge.gov/savings
- If remove blatantly stolen material for all the data, maybe.
- If it was given freely to EVERYONE without ever monetizing it, maybe.
- If they can remove the inherent human laziness, maybe.
- If the could guarantee privacy, ensure job security, maybe.
Maybe try not being a garbage person who acts like you're the friend of the peasants while simultaneously just being part of the same kleptocratic and neofascist Andreesen cult who thinks you can force your Yarvin fantasies on us?
Using any LLM to perform tasks that are usually done by professionals is not learning, it's trying to bypass the learning phase. At best it's foolish, at worst it's dangerous.
If they did not accomplish that goal, why should we trust that AI will be any different?
I would argue social media has some of those same aspirations. Look where that's gotten us.
Unscrupulous actors will corrupt AI just as they have social media.
AI is not panacea...it is theft and a method by which you can control and curate the information available.
AI cannot produce novelty, and idiots with cellphones cannot either
Experts can
AI is ubiquitous, due it's stock needing the latest model to stay inflated. That's not the same as capable.
Every worthwhile thing being done now is still being accomplished by the imaginative, critical, individual human brain.
I have three advanced degrees. I've used LLMs and it's arrogant to think those of us with an education haven't. LLMs produce roughly correct responses about 10% of the time wrt my "degrees"
I have a masters in clinical psychology. I am a therapist.
Do you think AI can do my job? Do you think someone with access to a hypothetical AI that returns 100% correct info can?
How good do you think an AI system would have to be to do even a crude mimicry of what I do as second nature?
I think if you’re not learning how to use AI you’ll be like the people who refused to learn how to use computers and then the internet.
You have no idea what AI is, how it works and how it is developed. Shows that you can perform as a tool, even as a rich fat slob. Maybe you should become an actor?
"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."
Query a bot with both. See how its respones differ. Which is the point.
Plus, if you truly believed what you wrote, then why wouldnt you tell every 12 year old to focus in computer science?
You utterly pathetic dweebs are constantly giving it SKYNET OR ELSE because you feel so butthurt about someone telling you you're talking complete shite.
AI is a shortcut to the how, not the why. And the why is far more important.
You first, mkay?
Idiotic statement.
LLMs already record all your queries.
They could easily start being trained on biased info.
In a word, it has been harmful.
I'll take a self learner over an MBA almost every time
It's plagiarism writ large and it's making the climate emergency worse.
But I guess billionaires with bunkers don't give a shit.
1. Consolidation of Power Within the Executive Branch
2. Suppressing Political Opposition and Media
3. Militarization of Government and Law Enforcement
4. Declaring a State of Emergency
5. Cult of Personality and Propaganda
2) fucking stop it
No thanks, Big Brother. We already have Elon’s winged monkeys trying to blur reality and dumb us down even more.
AI will be the death of human learning and here you are trying to enslave us. Fuck you, Mark. You are the problem. Thx for finally showing your true colors
It's the other stuff
Beyond that, good luck convincing the hiring team of this you fucking genius.
It may be a useful tool - but it isnt to be used instead of skills/education.
😬
Chatgpt is not a replacement for actual experience and knowledge in a field.
The fact you think so is more telling of your own ability than someone with a graduate degree who doesn't rely on Chatgpt.
STARLINK - STALIN
APPLE 📲 📵
hopefully it actually does turn it off. I don't trust em....
It gave me the wrong answer.
Let’s keep it far away from the kitchen.
It's just a program guessing at the statistically most likely next word. It's not intelligent nor knowledgeable in a trustworthy way.
I wouldn't hire anyone using an LLM .
Often correlation does correctly imply a causational element; but it's also horribly wrong to depend on that being correct - especially for government.
Cue Leon's stooges banning ice cream "because we saw that it causes crime."
Without the above statement, your post is just incorrect. Humans are still vastly better heuristic thinkers, because computers fundamentally cannot think heuristically, and can only act algorithmically based on input. Binary bits a brain does not make.
I’m not anti-AI but I find the hype tiresome
That was intentional language. LLMs are heavily reliant on their inputs. They are mostly unreliable, I thought I had articulated that.
Or aporia? Very old word
If u search “Ricoeur aporia” you’ll see an AI overview that gives u something to go on
The regular search results give >
https://ricoeur.pitt.edu/ojs/ricoeur/article/download/506/289
You got the usual LLM mealy-mouthed answer. Here are some … People say ….. There are several….
https://bdtechtalks.com/2023/03/06/chatgpt-llm-mathematics/
I'd like to further add that when you're doing something like straight multiplication, every single one of those boxes should be at 100% if you're using a computer. That kind of inaccuracy is just not acceptable.
Do not use AI for math.
I wouldn't use an LLM to do simple math, but you could if you added "use python" to the prompt.
You can't do proper analyses without knowing methods AND the field you are studying.
Yes, they are helpful to me, but only where I am already expert.
Perhaps you are extrapolating the term. But in the field of education, “instruction” is about the creation of circumstances that foster student learning.
@edfuller.bsky.social
in·struc·tion
2.
detailed information telling how something should be done, operated, or assembled.
In an operational condition (which are LLMs) instruction is precisely that.
No one would confuse a ski instructor with a Harvard professor
I would expect the ski instructor to be paying attention to each student and addressing their individual needs.
The Harvard professor is far more likely than the ski instructor to just be lecturing.
That instruction exists without an instructor or any instructed.
If you want it to "teach" you programming, they're simply awful.
Anyone who thinks AI can replace an instructor has no idea about teaching other people--esp kids.
AI is more likely to end the world than save the world.
please stop embarrassing yourself like this
AI is far from perfect. Not knowing what is not known puts the uneducated in a vulnerable position to fail.
A doctor can put a set of symptoms & ask AI what the likely cause is & best medication. The AI might suggest something that the doctor wouldn't think about & shortcut the decision. But the AI may also give a totally wrong diagnosis, and it's up to the doctor to know that!
By some estimates, more than 80 percent of AI projects fail — twice the rate of failure for information technology projects that do not involve AI.
https://www.rand.org/pubs/research_reports/RRA2680-1.html
Take away the experiences, you take away the meaning of life.
It's learning all the wrong things, and no one cares.
https://www.bbc.co.uk/news/articles/cq5ggew08eyo
If I put a RN code issue in o3-mini-high, it might give me something helpful but not perfect I can adapt.
If you ask it that for a technology you don't know it can lead to a cycle of 'Oh you're right, that doesn't work because of x. Here try this new broken code.'
AI platforms have led to a higher rate of suicide. Human contact and real empathy can never be properly made artificially.
Happily take the human over the software.
https://udm14.com
Is exactly the reason you shouldn't do that
https://www.popularmechanics.com/technology/a44675279/ai-content-model-collapse/
Except with Google you can assess the validity of the info you get based on whether it’s written by subject matter expects or consistent across multiple trusted sites.
There’s no equivalent for LLMs.
Also the shadows of the mountains are on the wrong side.
Human replacement it is not.
Just how similar to a human brain do "they" think AI is?
I love China for giving everyone a free one
Stated without evidence, this is very hard to take seriously.
It's just indicative that you can't blindly trust what the LLMs spit out, and would a novice know that?
Your painting question,BTW - AI answered that it was Bob Ross.
hosted LLMs are a huge backdoor into every business and endeavor
It's possible to manipulate people. It's trivial to program LLMs for any bias you want. And change that on a dime.
https://www.yahoo.com/news/musk-replace-sacked-government-workers-152330007.html
Spiking neural nets are promising, but I don't think computer science is ready for such a hard problem.
https://bsky.app/profile/televisionary.bsky.social/post/3ligqjdh4622b
And so far all they have to show for it is a neat parlor trick.
Without fundamental changes to how neural nets operate, we won't be able to overcome this wall.
The jury is still out on AGI. Some believe it will never happen, others think it's will.
I've built a very small transformer for school and a couple convolutional networks for work. They are impressive machines and are useful when applied correctly.
Unless you mean that AI is 1,000 iq points below the average human.
30% voted for Harris.
30% didn’t vote.
He won by a teeny, tiny margin.
I’m not good at math so I have to ask how that makes more than half?
AI is a scam and by no means an educational tool.
Stop telling people they don’t need an education and a piece of expensive paper stating they got it. It’s completely out of touch with reality for 99% of younger people in the 2020s
It’s just the reality for Americans that cannot borrow the “small” amount of a few million from their daddy to start their own company and skip the line, or get handed a position somewhere through nepotism
The qualities lost w the loss of hand drafting were too much for me to overlook. The kids did not have the training, while in theory it could be every bit as good, reality was opposite
I blame in part Reagan's defunding of schools nationally and the absence of reg art n shop classes.
He is old enough to have seen the qualities lost w some advancements. (Some are amazing, of course)
This is not good advice, please do not follow it.
#antiAI
A cloud isn't a puppy, and an LLM isn't reality.
BTW, I'm not a journalist, I'm an engineer with 40+ years experience.
Come on Cuban. You’re smarter than this
It's a secret no one knows.
However, if you don’t use any sort of AI & understand how it can amplify and scale your intelligence, you’ll fall well below ave very shortly.
AI is good for things humans can't do - massive pattern recognition for example. It's horrible at anything human to human.
It’s a hammer, if you use it just on nails (and don’t try to cut wood with it) it’s amazing.
Use AI to pre-detect breast cancer? Sure. Use it in place of air traffic controllers or college professors? (Two things Elon is pushing for.) Not in a million years.
That is 100% determined by users and customers.
It will only be used to detect cancer if doctors use it that way, not tech bros.
Sure. You should use AI to satisfy curiosities, much like Wikipedia, but to design car suspensions?
Sure. People it starts off with people asking questions about relatively simple, harmless things. Next thing you know people with zero education are designing safety critical components!
AI and LLMs will never have that skill set.
What magic shit do you know that's not obsolete, that we couldn't teach anyone?
The disdain for you fellow humans is appalling.
Everybody: it’s shown to be unreliable
Mark Cuban: so we’ve explored all options for the uneducated it seems
It's better, for now, to use the Wikipedia.
Absolute chicanery
Be assured that reading grad student papers written (composed) WITH AI is a pitiful experience as an educator.
Reading an 8th grader’s paper composed with AI feels like hallucinations.
What hallucinations are you suggesting the “uneducated” experience?
Garbage in garbage out.
But, I think it could be a good tool for closed environments, "What's the current company policy on blank" or "How much are we spending"
Don’t worry about learning how math works. You can just use a calculator.
Hear how that sounds?
There are uses for AI. But it’s not a miracle. And without having been educated one cannot surmise if AI provided a correct answer
Wouldn't it make more sense to just educate people and stop pushing LLMs until the hallucination problem is solved, rather than telling people to rely on an idiot oracle rather than actual knowledge?
That's flat out wrong. You need to be able to recognize AI's limits or you will break things fast.
People are exasperated by your comment because we see this anti-expert mindset playing out nationally, right now.
Read a book, go to school, use freely available online resources like Khan Academy. There’s already suitable options instead of “use AI or do nothing”
Imagine if all the money currently being funnelled into AI was used to actually improve the lives of the marginalised. Instead of turning us all into "prompters", replacing meaningful work entirely.
You need actual education and validation outside the LLM to avoid this and actually learn information.
Listen to yourself, man. cmon
I sense your answers here are more of a defense of your investments than a serious, “educated” take.
Why would you advocate using AI as a resource over something like Wikipedia where they could actually learn things that are mass-reviewed for accuracy?
Books still exist
Also, are you suggesting people attempt to deceive potential employers?
B. That's also not what you said
Or or or... there are so many resources for free education! They could use AI to help them find a free course, podcast, or book that might help.
Subject knowledge and experience is important. You gain it by putting in the work.
LLMs are text prediction models, have no sentience and have no internal moral compass.
To be able to assess the quality of results, you need sufficient competency in the field.
In addition, these models are produced by companies (and countries 👀) with agendas.
You don't replace education with these for the same reason you don't let Elon Musk put a chip into your brain.
*major eye roll*
Blindly trusting the output of an LLM is not a good idea.
What were saying is more resources need to put into our public education system. Right now the states are planing radical defunding to public education.
Models are built on what they have read and historic data.
3000 yrs of historic data, education, politics, finance, economics, architecture, cuisine, design, religion, ..
Structured by men for men and with men in mind.
AI will reflect the male hive mind.
AI experiment on literature.
Like what if the dictionary just sometimes had definitions that were wrong. Or words that didn't exist. How am I supposed to be able to use that as a resource?
I might get 1-2 hallucination a month out of 900 API calls. 🙄
‘Hallucinations’ = sanitising name
You can hold out as long as possible, but it will end up passing you by.
The problem isn't that the technology can't be useful. The problem is that the technology is unaccountable to its wrong answers.
The fundamental question that permits actual knowledge is to ask how you know something to be true.
The fundamental human flaw here is that we trust it too much.
The only people who can't see the holes in this ARE the uneducated.
#NotToBeRude
Ask AI a complex question, and you might not know if it's answer is correct. Well, if the interpretor executes the code you are on your way to an answer.
Use all the educational tools available, including AI
Try this in chatGPT: draw a group of watches showing 3 minutes after 12
The fact that you don't know this reeks of the Dunning-Kruger effect.
https://wikipedia.com
https://archive.org
https://landchad.net
https://hackaday.com
https://codeacademy.com
https://w3schools.com
https://wolframalpha.com (the og "AI")
https://duolingo.com
The answer is for the uneducated to get educated enough in that area to know a good answer from a bad one.
See my real world example of a Buick 231 engine weight that I just posted in this thread.
“garbage in, garbage out” is a critical problem when users can’t tell garbage from correct answers.
We could call it "education" and do it while people are young.
Whatever happened to putting in the effort?
I known education systems are problematic everywhere, but the answer is not AI, it is properly funding education and paying teachers WELL.
AI teach you that Mark?
1. if you're a kid: Go to school. give it your best shot.
2. If you're an adult: there are likely state programs that give you financial aid to go through adult education programs.
As an outside observer: Always vote for education as it betters our society.
Like everyone has for all of human history.
We have a rule in engineering: know when your models are invalid. So you know whether to believe a result.
LLMs are tuned for engagement. They’re a pretty toy for business types try to plug into everything, no matter whether it will even work.
I've scrolled too long to find it
It's gotten to the point that I can't stand to use Google at all, anymore.
Or just written by idiots who have no credentials but the citation looks like it is real so people assume it is legit.
AI is only great if what you're using it for doesn't require you to be correct or properly written. That's why Nate Silver uses it.
https://www.wnycstudios.org/podcasts/otm/articles/brooke-talks-ai-with-ed-zitron
AGI is coming for everyone's jobs so Cuban is just trying to tell people how to be ahead of it.
It is the biggest threat to our democracy. You can't ignore it.
https://apnews.com/article/nobel-chemistry-prize-56f4d9e90591dfe7d9d840a8c8c9d553
My worry is that the poor sods think it will replace humans - but that is about heuristics - which they think they have down to a T 😂
It was done with a specialized model, built for this one purpose, by people with a lot of insight into the problem.
But the sentiment still stands, AGI is coming for all our jobs, and it's probably better to get ahead of it or prepare instead of ignore it.
I think we can all agree on its exponential progress.
https://phys.org/news/2025-01-ai-linked-eroding-critical-skills.html
https://www.mdpi.com/2075-4698/15/1/6
2/2
People should definitely learn how to utilize LLMs, I use them daily, both at work and my personal life.
I just don't agree with the initial sentiment and think that the success of Alpha Fold is a bad argument for AGI.
1/2
No.
Was this an LLM system?
No.
Is ML the same as LLM?
No.
His post is dooky.
Don’t work so hard to defend dooky.
Answers in a loooong list please 😂
Substitute AI for Science and Education for the Arts. Yes anyone can write a prompt and get an answer. the Educated know whether it is correct, right, or morally repugnant.
- Frank Herbert, Dune
I use my education and experience to correct ChatGPT all of the time.
ChatGPT helps me organize thoughts - I dump in a bunch of facts and thoughts about the message I'm trying to convey, and it will give me a helpful start with presentations!
I got this tday so I am seriously wondering if his account being hacked?!
I use ChatGPT all the time as a resource and to help me be time efficient, but if I didn't have loads of development training and experience to vet AI generated info, I'd be putting out a crap product for my employer.
They are so many who look at a screenshot of an article, and have no clue how to research it.