In this lecture, I want to give you a quick preview of Google BART as a large language model and show you how powerful of a tool Google B can be when it comes to extracting information using conversational AI, specifically in the use case of working with internet and live data, which is not essentially available in chatGPT in the base model.
I’ll start with a very simple prompt and ask BART to help me plan a trip. So my prompt is, ‘Let’s plan a trip to Smoky Mountains National Park near Atlanta. Can you write me a 3-day plan which focuses on outdoor activities?’ Let’s see what the algorithm says.
It came back with a plan that includes links to different information and photos as well. The plan suggests starting the day with a hike to Kingsman’s D in the morning, visiting Kat Cove in the afternoon, and exploring other places on day two and day three. Not a bad plan.
Next, I asked BART about the longest hikes in Smoky Mountain that I can go to. It provided me with some responses, including a 71-mile hike and the third highest peak in the Smokies, Groto Falls Trail and Kingman’s Dome.
Then, I changed the question and asked BART to recommend the best town to book an Airbnb that is close to Smoky Mountains. It gave me examples of towns like Glenburg, Spancvill, and Bryson City.
I further asked BART about the price range for a one-bedroom apartment Airbnb in September that I should keep in mind for my trip expenses. It provided me with approximate prices for different locations and suggested booking early and being flexible.
Next, I wanted to know the weather during September in Smoky Mountains. BART understood my reference to Smoky Mountain weather and provided me with the current conditions.
I then asked BART to plan the 3-day trip again in a tabular format and include a very long hike on day two. This shows how specific data can be extracted using BART. Finally, I mentioned that the extracted data can be saved in docs or Gmail.
This was just a brief introduction to how powerful Google BART can be as a conversational AI tool for extracting personalized answers from the internet. In the course, we will explore other techniques of prompting and how it can be applied to different aspects. In the next preview video, I’ll show you how the same concept can be applied to programming and how fast BART can code for you without writing a single line of code. We will just provide the prompts, and all the coding will be done by BART.
