AI & Web Development

[00:00:00]
>> We are web developers, right? So we mentioned that. So what's the relationship between open AI Chat GPT and web development? So this is what we can do as web developers first integration. We can use artificial intelligence for your apps. We can integrate that in your app for many things.

[00:00:23]
We can create, we can transfer and manipulate data, conversations, we can create chatbots. And there are many other ideas that they will show you in a minute. That's one thing that we can do and we will focus on that today. You can create plugins In this case, plugins from ChatGPT.

[00:00:42]
For example, we can create the front-end masters plugin. So you can ask ChatGPT about in which course from front-end masters, I can learn about web push. Then ChatGPT will talk to the web service that we can create here in frontend masters and it will respond with the data that then Chat GPT will process, understand and create the output.

[00:01:08]
That's a plugin. We can also use GPT for content creation. That's a different, it's more around marketing and SEO. We are not going to really, really cover that today. And there is also some questions today and some of you may be wondering, we do have SEO right search engine optimizations.

[00:01:30]
Well, do we need an AIO, artificial intelligence optimization because now AI can read web content being can read the Bing version of GPT, can browse the website. Now there is a plugin for chat GPT that can also browse the website. The plugin is still not available to everyone, but it will in short time.

[00:01:55]
So do we need to prepare or do something to serve better our AI? Well, we'll see. The quick answer is no, we don't need to do anything in particular. However, there are still some people working with that and say playing with that to see what they can achieve.

[00:02:21]
So let's talk a little bit about some ideas or ideas that you can start thinking about. So what things we can do as web developers. You can use GPT for user input detection, such as hate speech, content that you don't want in your system. So you can Pass user input to ChatGPT and ask for something in return.

[00:02:48]
That something can be categorized this content in these possible options. And then you can automatically reject that or not within your system. Content creation you can create content on the fly. Sentiment analysis for example, you have a review system and you can send that the user is sending you a review of any of your services or products.

[00:03:16]
You can send a text to ChatGPT and ask ChatGPT to give you just one answer. Is this positive or negative? And based on that, you can change the message that you reply to that user automatically. And if it's negative, for example, you can send an email to someone.

[00:03:34]
So that someone can reach the customer for blah, blah, blah. Make sense? So that's one idea. You can customize or make articles or content really personal for example, maybe a newspaper. You have a newspaper, of course, a journalist, or maybe ChatGPT, someone wrote an article, it's published, but maybe I don't understand it.

[00:04:01]
So I can ask the website with maybe a drop-down option for a simpler version, or a summary, or a version as if I'm 10 years old, or a different language or whatever. And you can then customize that text and offer a service to your readers. So, and then can improve the time on your page, and then you have more ads, and more views, and maybe more money.

[00:04:31]
Language translation is actually pretty good. It's better than Google Translate, translating to languages, even to languages that are not even common. I don't know, Guarani, that's from Paraguay, that on Google Translate is not available or not good enough. Search, you can use it for a search box and you can use ChatPT to find the content within your database or within your articles for that, and we will explain how that's done later.

[00:05:03]
To summarize content, GPT is pretty good on that, q&aIt can create the Q&A for you. Test automation, now talking about code, it can actually write your tests directly. Chatbots, that's like the first step when we are dealing with GPT. Let's create a chatbot. So then it can offer like customer service to your users.

[00:05:31]
You should use ethics here and say to the user, hey, this is AI have in mind that and there's more email generation, content filtering, content tagging, this is pretty cool. So you have an article, you can tag the article so then it's better to connect articles between them instead of doing this manually.

[00:05:52]
You can automate social media. So you can post on Twitter on Instagram automatically when something happens on your page and make that content doesn't like look automatic like it's always the same text is not a template. ChatGPT can generate different versions of that. Keyword Research for SEO, a spelling and grammar error.

[00:06:17]
So text correction, that's pretty cool and that content enhancement, maybe you have a lot of articles without the machine. Now you can create images around that. Content Creation, email filtering or even automatic transcriptions on videos. There are more ideas. This is just a list of ideas of what you can do as a web developer with AI.

[00:06:40]
And by the way, this is not ideas of integrating your app, like make an AI app for something. You can make an app that uses AI as part of the business. So in this case, we were just talking about ideas for any website, but you can create an app that can translate JavaScript into something else or any language into JavaScript.

[00:07:06]
So in that case, you use AI as the core of your business. We can use other specific artificial intelligence models in our apps on website. But what makes LLM easier is that they are multipurpose. So then we don't need to train a new model. We don't need to create a new API.

[00:07:24]
We don't need to do anything we just set up in our app, server side, client side. We will see later what's better but we can just have one codebase to talk to LLMs for all the purposes. We don't need to set up different models for that. That's the great advantage of LLMs.