Is Anyone Else Disappointed in the State of LLMs?

After spending a couple of years getting excited about them, I've tried signing up for ChatGPT, Claude, and Gemini. <br><br>I won't go into the details, just to say that I've been disappointed with the results. I was hoping for a lot more. <br><br>I don't know if I just got my hopes up for too long, but I'm really just surprised at how long it seems to be taking to get to the point where these are really useful, and not just novelties.<br><br>Any thoughts?<br><br>[Edit] Wow! Thank you all for your kind words and engagement. I'm a professor at a small school teaching computer science. The above was just a general vent to a community I respect.<br><br>My general idea is that we've been *describing* how great LLMs are for years, but I've recently tried actually *using* them, and I'm disappointed. I understand that maybe I just got my hopes up too high, or maybe I'm just not using them for the right thing.<br><br>I'm trying ChatGPT, Claude, and Gemini. I'm teaching intro to machine learning, intro to databases, and a senior capstone project. I'm a community college professor with a couple of decades of experience in industry, so I'm teaching my students machine learning based on years of experience.<br><br>I *desperately* want to teach a course on LLMs, and I feel like I've been led astray by all the hype, and I'm really trying to figure it out.<br><br>In short, I feel like I'm being foolish in how long it is taking me to understand how to use them. Maybe I just need to try longer to understand how to use them correctly, or maybe I need to wait a little bit. I understand that Claude and Gemini are very new offerings. I've been trying to use them for teaching my students machine learning. <br><br>My general idea is that we've been *describing* how great LLMs are for years, but I've recently tried actually *using* them, and I'm disappointed. I understand that maybe I just got my hopes up too high, or maybe I'm just not using them for the right thing.<br><br>I'm trying to use a large language model (I don't care which one) to take a prompt and write an essay that could reasonably be written by a community college student, and I feel like that's just not where we are. I'm not getting anything. I could pay a professional freelancer to write the essays for my students at a price I can afford, and it would be pretty professional. I'd love to have them written at a high school level, which I feel like is higher than what I'm getting.<br><br>I've been trying to get the LLM to give me things that would be reasonable for senior IT students to work on for a capstone project, and I feel like I'm just getting nonsense back, or nothing.<br><br>I advise our honors society, and I've been trying to get them to work on machine learning projects, but I feel like they're just getting smokescreen and mirrors, and I've been disappointed.<br><br>I'm trying to use Claude to produce code that I can use in a classroom, and I feel like I'm either getting code I already know how to write, or code that doesn't work.<br><br>So, to that point, I don't know if I just need to be more patient, or if we're still in the early days. <br><br>I've been teaching machine learning at our community college for over a decade. Is it just me, or am I just getting caught up in the hype? I just don't feel like they're that useful.