Introduction to PaLM 2

0:00 Hi everyone and welcome to this full course on Google's large language model Palm 2. My name is Ania Kubow and I'm a software developer as well as course creator here on freeCodeCamp as well as on my own channel. If you are new to the world of large language models, this course is the perfect video for you. We will be starting completely from scratch introducing you to this wonderful world that powers some of the most powerful AI tools today, such as Google's Bard for example.

0:24 By the end of this course, you'll be able to confidently use the Palm 2 API as well as hopefully gain some knowledge about the world of AI in general. In fact, here's a whole list of the things that we'll be covering in the next hour or so. So, we will first look at why use Palm 2, followed by what is AI and large language models in general. We will then get set up, look at tokenization, AI hallucinations, authentication, the models that come with Palm 2, and finally we're going to build our own AI chatbot.

1:10 So, that's at the end. I hope you're excited. What are we waiting for? Let's do it. First off, why use Palm 2? Google's Palm 2 is an advanced large language model or LLM designed to be more effective at a range of tasks than its predecessor, Palm. Before we get to demystifying some of the words in that sentence, such as LLM, let's actually understand why we would want to use something like Palm 2 in the first place.

1:35 If you are a developer, here are some reasons you might think about using Palm 2. First up, let's look at its multifunctionality. Palm 2 is capable of various tasks including code generation. This means you can use it by helping you code, debug your code, or explain code. So, imagine you have this component here written in JavaScript, and suddenly the team decides they want to start using TypeScript. You can literally ask Google's AI tool Bard, built using PaLM 2, to convert the code for you into TypeScript.

1:59 Another thing you can ask it to do is literally write the code for you from scratch. PaLM 2, in comparison to its predecessor PaLM, has been pre-trained on a broader range of data sources, meaning it has better multilingual capabilities and can handle diverse programming languages. From popular ones like Python to more specialized ones like Fortran. However, please do keep in mind that taking this code with a pinch of salt is advised as some hallucinations, or in other words, guesswork, could have been