A very short (27-second) YouTube clip by James Laughlin offering quick practical advice for using ChatGPT: use trial-and-error, ask plainly, and treat the model like a curious, very smart conversational partner. The speaker cautions that large language models are intelligent but not wise and emphasizes experimentation as the primary way to learn.
Trial and error. You can't really go wrong chatgpt. And if you do, then you've reached the edge. You've actually reached the wave. It's really interesting. So keep trying things. How do I use ChatGPT? Just write it. It will tell you. It will give you a very good answer too. What are ways I can do my job better? It will say, what's your job? You just got to get in there. Imagine you're sitting with a very curious, very smart person who can tell you many things. They're smart, but they're not wise. ChatGPT is intelligent. LLMs are intelligent if they're not wise.
