Careful LLM Usage
Mark Cuban posted some BlueSky posts about AI and I had some thoughts on them.
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.
— Mark Cuban (@mcuban.bsky.social) February 17, 2025 at 9:59 AM
When I talk to kids today and they ask me what I would do if I were 12 today, my answer is always the same, read books and learn how to use AI in every way shape and form you can. It is a living library that gives you responses and can help no matter who you are or where you live.
— Mark Cuban (@mcuban.bsky.social) February 17, 2025 at 3:52 PM
My stance on AI, and more specifically Large Language Models (LLM) has been, “perhaps a useful tool if you know how to use it”. Asking ChatGPT to do your homework, bad idea. Don’t do that. Asking ChatGPT to spell check your homework? Probably fine. Risk of bad suggestions are certainly there but I think it’s low. Maybe ask two models to double check. Asking AI to write all of your code in your job? Bad idea. Stuck on some code and need some help? Probably fine. Make sure you understand the answer and don’t copy and paste. I hope my position is clear that I’m not anti-LLM but it’s a use with caution approach.
I read Cuban’s posts and I think to myself, is this not what the internet is for? LLMs live in the browser, and pull (perhaps unethically) data from all over the web. Are LLMs really a game changer in the way he is referencing? I don’t think so. I think anyone who had the desire to learn, to experiment, to build, can do with or without LLMs. Do LLMs make this easier? To some extent sure. I also think it makes it worse. If LLMs serve up confidently incorrect answers, a novice won’t know that. They’ll need to dig in, and look for more answers to verify the original answer. Suddenly is the tool better than a web search?
I have to admit that I've been surprised at the level of hate at AI LLMs. There is finally a tool that allows those without access to an advanced education to self learn, anywhere , anytime. To get answers to questions, (even if sometimes wrong) , that they would otherwise never have access to
— Mark Cuban (@mcuban.bsky.social) February 17, 2025 at 3:50 PM
[image or embed]
I can understand where he is coming from. I also think a lot of people are understandably concerned about what AI means for the workforce. I think if there is one thing tech has shown over the past 20 years is that a lot of bad things come out of good intentions because not everyone is a good actor. In the most basic example, an LLM can probably serve as a decent tutor or at least as a tool for a human tutor. We can also fully expect that students will use LLMs and plagiarize from them.
Since we can only control ourselves, it’s up to use as individuals to decide to use these tools responsibly. Cuban is right that they can be useful but it’s also…just the internet.