Do you have evidence to support this contention?
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
"But I like Chris Christie and Nikki Haley. And honestly, if either one of them was pro choice, I’d probably actively support them." - Mark Cuban in GQ
https://www.gq.com/story/mark-c...
Now granted neither case is anyone specified whether or not the example would be particularly good or spectacularly catastrophic.
But ymmv?
Cuban is acting like this is some profound insight but it’s really not.
With an LLM you don’t know if the info you’re getting is accurate unless all its sources are specifically cited.
1 big downside to it is it fosters an enhanced Dunning Kruger effect.
Had a customer tell me last week that he would do the taping, just needed to watch a video to learn.
I ended up doing it.
However the “ask the right questions” is doing a huge amount of lifting there. Because that’s the whole crux of the ball game and what people really struggle with.
The “curiosity and willingness to learn” also powerlifting here.
https://pivot-to-ai.com/
Now apply this to the DOGE novices with LLMs and you understand the prob in DC.
Like…at least fifty percent of education is teaching *discernment*. Practicing, *with guidance*, the art of inquiry itself.
That Dunning-Kruger drug the kids are talking about these days sure must be addictive.
https://bsky.app/profile/ketanjoshi.co/post/3lhqfjewto22q
Garbage into the computer garbage AI out
If you don’t have someone with critical thinking skills to tell the difference you’re not gonna get the right answer
How would you know how to ask the question to get an optimal answer
Aside from that, it would be a difficult study, but if you need somebody with no education, scan for red hats in the crowd.
https://www.linkedin.com/posts/michaeljedelman_legaltech-activity-7297222259771928577-tGPV?utm_medium=ios_app&rcm=ACoAAAGTnuMB5tItDv_60KFfRBxsl7kHbsdoF8M&utm_source=social_share_send&utm_campaign=copy_link
The drafting is probably as good or better than what you’d expect from a 3rd year big law associate after several $1000 in billables.
I don’t see how this sort of power in the hands of a less resourced party doesn’t absolutely level the playing field between them and the much better resourced luddite.
🤷
But the text generators? They write filler and nonsense.
If your job can be replaced by a modern text-generator, then your job is filler and nonsense.
Lawyers doing this are just bad at their job.
indemnity provisions, limitation of liability and auto renewal. More succinct: AI works well on contracts
The hard part is working out what questions you should actually be asking
Me: where is Elon Musk located at this moment
GROK: Based on recent information available on the web, Elon Musk is currently in East Germany.
East German no longer exists.
I knew a long part of the CFR contained language doing X, but did not know what words were used to accomplish X. I put the CFR sections in the LLM and asked it to give me the cite for that language
I like Mark, but his stuff on AI feels like it was written by AI as a justification for AI based on Kevin O'Leary licensing deal offers.
Getting use out of it takes practice but it definitely increases productivity when you get how to use it.
Donnie does it when his coup is delayed.
Mark throws often-shitty ideas on a wall because HE gets attention and he wants others to work out the details he's not willing to do.
Some of those ideas are so silly they damage his cred.
The underlying premises: -education is only important to prepare a person for work (for the man)
-AI, which is based on what people already know, is accurate
-work requires no creativity, specialized knowledge, or local adaptation
…all incorrect.
If AI could replace anyone, it would be a CEO.
AI is a tool, not a panacea for every problem known to man.
I look forward to seeing what the judge does to them.
The thing I didn't do was blindly accept what it told me.
People who can't identify high-quality info about disease, nutrition, the law, or politics via an open internet today are not miraculously going to maximize AI tomorrow.
Note that AI seems to be relying on humans to play by the "rules," which is stunning in its lack of understanding regarding human behavior.
Apparently it doesn't understand what "evil" means.
1 of 2
[==*==]
2 of 2
[==*==]
And elsewhere a lawyer had to admit that his Ai hallucinated and made up half the evidence he put in front of a judge
They do what they're told and sounds super confident. They're the employee who always says "great idea boss, I'll get right on it".
If managers had the skill to perform tasks instead of delegating them they would realize how frequently LLMs are wrong.
But they don't.
veterinary medical notes. It doesn’t really work unless I let it do “formatted dictation” which actually works pretty well. But I literally have to tell it what I want it to write and all it does for me is make nice paragraphs and bullet points to copy and paste