There are quite a few web interfaces where you can access a Large Language Model. ChatGPT is the best known one and Google's Bard, Microsoft Bing and quite a few others also work well. Let's take a look at how people are using these LLM applications. Whether or not you're already using them regularly, I hope that this will give you some new ideas for when and how they could be useful to you. LLMs are giving a new way to find information. For example, if you ask it what's the capital of South Africa, it may give a response like that. Now as we'll see later, LLM can sometimes make facts up. We call this hallucination.. If you're really relying on getting the right answer the question, it may be useful to double check the answer with an authoritative source before counting on it. But in this case, it does get the capital or three capitals of South Africa right. And sometimes a back and forth with an LLM is helpful too, for example, if you ask what does LLM stand for, it might answer, LLM stands for Legum Magister, which is a term used in law, and it's actually a pretty common use of the acronym LLM on the Internet. But if you were to then say what about in the context of AI? Then hopefully it'll say in the context of AI and LLM refers to a Large Language Model. Sometimes this back and forth can help you give the right context, the LLM to give you the information you're looking for. LLM can also sometimes be a thought partner to help you think things through. For example, I often use an LLM to help me refine my writing. If you were to tell it, rewrite this for clarity, since all around the world are realizing learning has happened and so on. Leading LLMs are actually pretty good at rewriting texts for you. Or here's a fun example, if you were to tell it, to write a three hundred word story involving trucks. Maybe because you have a child that likes trucks like I do, my son love trucks. But to encourage them to brush the teeth, then leading LLMs can actually create pretty fun and interesting stories. I don't think this is nearly as good as the stories written by the great novelist, but for a quick fun thing, I think it's not bad. Now, there will be times when you're looking for a piece of information and you might be wondering, "Should I use web search or use an LLM?" So If you're playing a sport and unfortunately wound up with a sprained ankle and want to know what to do about it, web search can lead you to pretty authoritative and I think trustworthy sources that can give advice on how to approach medical matters. For example, web pages from the Mayo Clinic or from Harvard Health. These seem like they would be trustworthy sources for what to do about the sprained ankle. You could also ask an LLM what to do about the sprained ankle and it will generate some answer. But given the propensity of LLM's to make things up and sometimes sound very authoritative and confident when making things up, I would probably want to double check anything it says about healthcare or medicine before following the suggestions. Here's one more example. If you want to make a pineapple pie and they're looking for a recipe, it turns out there are lots of recipes on the Internet for a pineapple pie, and picking one created by a trusted website or a trusted chef, that might get you pretty good results. Or you can ask an LLM to make one up for you, and in that case, it'll come up with something that frankly might be okay, but also has a high chance of being a somewhat strange recipe. So if you want to bake a pineapple pie, I would probably go find a web page because there are multiple web pages that will give a good solid answer to what's a good pineapple pie recipe. But if you were to look for something more esoteric, say your friend challenges you to make coffee, infuse pineapple pie, there aren't any web pages that I could find really on coffee infuse pineapple pies. I don't think there is currently a single web page that gives a good answer to this. This would be one example where an LLM can help be a thought partner to think through how you might go about baking a coffee, infuse pineapple pie. These are just some of the tasks for which you might find the web user interface for LLM useful. We'll explore more examples, discuss strengths and weaknesses of LLMs, and go through some best practices later this week. But as you can see from this video, Generative AI is capable of many different things. In the next video, we'll more systematically discuss Generative AI as a general purpose technology as well as start to come up with a way to organize all of these things that they can do, which includes writing, reading, and chatting tasks. Let's go take a look at that in the next video.