Adding web browsing, image creation, and data processing - ChatGPT Tutorial

- One of the features that make GPT so powerful is your ability to control which advanced capabilities each GPT has. Let me show you. I'll create a new GPT and this time I'll go directly to configure and call it color, Adult Coloring GPT. This GPT will use the DALL-E 3 capability to generate adult coloring pages. First, we need to put in an instruction, so this is the prompt. And then we scroll down to capabilities and activate just the capability we want. In this case, DALL-E image generation. Now I can immediately test my GPT to see what it outputs. So I'll say quaint modern house in a snowy landscape. Because I've activated the DALL-E image generation capability, the GPT now automatically calls DALL-E 3 to generate the image for me, even though I didn't explicitly say that that's what I wanted, I just said a quaint modern house in a snowy landscape. The response I get is a generated image that fits the criteria, both of the original instructions and of my request. Now, looking at this, I can see immediately that my instructions are correct, but my request was not specific enough. And then I can go in and either change my instructions to be more specific or I can go change my request to be more specific. But as you can see, the capability automatically kicks in when I activate it. The key here is each GPT you build can have one of, two of, or all three of these advanced capabilities activated, web browsing, image generation, and advanced code interpretation. The advanced code interpretation feature is actually quite interesting because it allows the GPT to run code on its own. So for example, you can upload a spreadsheet and ask it to do something with the data, and it'll do that. Again, this relies heavily on your instructions and the input the user makes, but this gives you the capability of turning on and off these capabilities for each individual GPT to further customize its behavior.