I've always had a deep love for music, and as a kid, I begged my parents for an iPod instead of a PlayStation. I mean, what's the point of virtual worlds when you can bring the party with you wherever you go?

Even when I was stuck doing homework in school, I always had my headphones on, letting music transport me to a happier place. So when I decided to learn how to play the guitar, I was a little nervous. I mean, what if I turned out to be tone deaf or had fingers made of butter?

But I mustered up the courage to give it a try, and now I'm proud to say that I can play the acoustic guitar! It's an amazing feeling to be able to create music with an instrument, and who knows, maybe one day I'll even write my own songs.

So if you're thinking about learning an instrument, don't let fear hold you back. It might be tough at first, but the sense of accomplishment you'll feel when you finally nail a tune is worth it. Plus, it's just plain fun! So grab your guitar (or whatever instrument you fancy) and start jamming out. You never know what amazing musical journey you might embark on.

One day, a group of dinosaurs were hanging out in the jungle, trying to come up with something fun to do. Suddenly, one of them had an idea.

"Hey guys, I know a great joke!" said the dinosaur. "Why was the Tyrannosaurus Rex such a terrible singer?"

The other dinosaurs all leaned in, eager to hear the punchline.

"Because he had a really tiny brain and his arms were too short to reach his mouth!"

The other dinosaurs roared with laughter. "That's hilarious!" they exclaimed.

Feeling proud of himself, the dinosaur decided to tell another joke. "Ok, here's another one. Why was the Velociraptor such a terrible basketball player?"

The other dinosaurs waited with bated breath for the answer.

"Because every time he tried to shoot the ball, it would always end up in the raptor's nest!"

The other dinosaurs couldn't stop laughing at the silly jokes. They spent the rest of the day telling each other dinosaur-themed jokes and having a great time.

In the end, the group of dinosaurs learned that laughter truly is the best medicine, and that a good joke can bring friends together and make even the grimmest of situations a little brighter.

After a few hours of nonstop joke-telling, the group of dinosaurs were starting to get a little bit tired. But they didn't want the fun to end just yet, so they decided to play a game of hide and seek.

The dinosaurs all scattered, looking for the best hiding spot they could find. The stegosaurus, Steve, decided to hide behind a big rock. He figured that with his spikes, he would be nearly impossible to spot.

As the other dinosaurs counted to ten, Steve closed his eyes and tried to blend in with the rock as much as possible. But just as he was getting comfortable, he heard a loud noise.

"Found you!" yelled a voice.

Steve opened his eyes to see the velociraptor, who had been searching for him, standing right in front of him with a big grin on his face.

"Darn it," Steve grumbled. "I thought my spikes would have given me away."

"Nope, you were completely hidden," the velociraptor chuckled. "But your snoring gave you away!"

Steve's eyes widened in shock. "I don't snore!" he protested.

The velociraptor just chuckled and patted Steve on the back. "Sure you don't," he said with a wink.

The other dinosaurs all came running over, laughing at Steve's predicament. And from that day on, Steve became known as the snoring stegosaurus, much to his embarrassment.

But despite the teasing, Steve couldn't help but laugh along with his friends. After all, there was nothing like a good joke to bring a group of dinosaurs together.

Introduction ☕​

If you've been living under a rock and don't know what Github's co-pilot is, it's simply an in-editor extension that makes suggestions to you as you code, attempting to finish your lines or even suggesting entire blocks of code. It's an intriguing concept, but it obviously raises a slew of issues, especially given that it was trained using GitHub data. Github co-pilot is the Godzilla of autocomplete. It's state-of-the-art ai technology.

For example, instead of searching stack overflow for a good API and manually writing out the function like a caveman. You can simply add some comments about what you're looking for. The artificial paired programmer will do all the work for you. It will write out a clear clean function that looks way better than any code you’ve ever written.

It is also not confined to a particular language or framework, allowing you to deal with everything from python to javascript typescript, and yes, even ruby or go.

What wizardry spells were used to train Co-Pilot? 🧙​

GitHub Copilot is powered by OpenAI Codex, a new AI system created by OpenAI. It has been trained on a selection of English language and source code from publicly available sources, including code in public repositories on GitHub. Source

Working Steps

1. Give the model an English assignment, such as "write the power function.”
2. Copilot attempts to complete the problem after reading a large amount of code.
3. The difference between what Copilot provided and the intended result is computed.
4. This mistake is then sent to Copilot so that it may be corrected (i.e. learn).

To recap, Copilot employs Codex, an NLP model trained on open source code. Copilot's goal is to generate proper code in response to an English language prompt stating the problem to be solved. Copilot is capable of doing so since Codex has been educated on significant quantities of open source code in a variety of languages.

Doesn't Intellisense ring a bell? 🔔​

Things I encountered only

Some of the factors that make copilot so strong are, first and foremost, its ability to take in context more than any other ai. It allows synthesizing code in ways we've never seen before, as opposed to IntelliSense's short term memory, which fades after one or two lines of code. Copilot may extrapolate from existing code in a file and use well-defined variables, classes, and structures.

For example, if I wanted to build a function in typescript to validate a string, copilot will utilize the variables I've previously typed to fill up the rest of the function. While this appears to be a really strong tool, the copilot can accomplish a lot more grunt work that we hadn't even considered previously, such as having automated autofill capabilities for manual labor chores with patterns and recurring structures.

The tool comes really handy while working with JSON files. If you're writing a lengthy JSON file and need to fill out different sorts of languages, for example, copilot can literally comprehend the context of what you're trying to do and finish the work for you. This means it understands that you're attempting to iterate over multiple languages in a JSON format.

We all love action. Let’s Test the Co-pilot. 🤖​

Let's get started with some very fascinating things it can do. For our first example, I'm going to ask Copilot to perform a few things via comments.

Let’s start with the easy stuff. I first asked GitHub copilot to generate a react component. Here, it did a decent job except for the fact it was imported in a different fashion.

The component looks different, and the actual syntax of the component is specified dependent on how you write the comment, which is extremely amazing. This is why it's critical to be specific about what you want because the copilot will produce a result based on the context of what you've written previously and presently.

Let’s change the language and increase the difficulty. I asked the copilot to import openCV and capture the video frames. One thing is for sure, copilot works the best if it's under your leash. It does not have a mind of its own. Here, it did a good job and was pretty commendable. It even wrote comments for the specific line’s purpose.

Next, let’s ask the copilot to fetch the browser version and also to get the dimensions of the screen size. Here, it did an okayish job. I just feel copilot has a habit of adding a lot of boilerplates. At least the task was achieved.

My opinion after using Copilot for 8 months. 💖​

In the development community, people always crave things to bump their productivity. Hence in the pursuit, we all use a lot of tools. I believe the same thing will be true with things like GitHub co-pilot, where you can be considerably more productive. When you're more productive, there's more motivation for someone to hire you, so it's not that we'll have fewer job vacancies; rather, we'll have more job opportunities. Even if you perform remedial work as a coder, this will make you more productive and useful.

Also, at least for me, copilot has taught me things about libraries and technologies that I hadn't had the chance to really look into using, and I just kind of started writing code and said hey, I want to do this thing with that library and it just kind of guided me along the way to teach me how to use that library and I found that to be really really cool.

There have also been occasions when I wanted to check whether co-pilot could do something that I was acquainted with and it ended up doing it in a new manner that I had never really thought about or even utilizing a method that I didn't even know existed, so it's taught me something and I think that's really interesting.

I think it's a really well-integrated thing and the technology and the actual accuracy of what it's actually producing. I find it to be far more useful than not like I said it does sometimes produce silly things but for the most part, it is very helpful very quick. Again I tend to find it most useful in the more remedial task-type stuff that I need to do while I'm programming and I think that's great.

For the time being, I absolutely enjoy it. I hope the future is not gloomy. In conclusion, that is how I see things going.

Thank you for taking the time to read my post. I hope you enjoyed it. That’s all I wanted to share with you. Have a great day!

Till then Au revoir, frérot​

Find me here 🤩

Twitter

Website

Github

Vite / Yeet / Wheet? Viteamin [sic.]​

Vite, pronounced /vit/, like "veet". Vite is a French term that means rapidly or quickly. Evan Yu, the inventor of Vue.js, designed vite. The next generation of front-end tools is vite, which offers immediate server start and lightning-fast hot module replacement. It also provides optimized builds with full feature support utilizing rollup universal plugin APIs that are entirely typed. The server-side rendering is yet experimental.

Main Purpose​

Front-end web apps are at the heart of vite, a javascript development tool that simplifies the way we build and develop them. It provides two functions.

1. The first is to serve your code when developing locally.
2. Bundle your javascript, CSS, and other assets for production.

Templates Offered​

No, this is not a view-only tool. It offers templates (starting files) for a variety of frameworks and vanilla JavaScript, as well as TypeScript, JSX, and Sass support (although you need to install one dependency for Sass).

We get to pick a project name and a template after executing-

npm init @vitejs/app

At the time of writing, the templates offered were are as follows:

• vanilla
• vue
• vue-ts
• react
• react-ts
• preact
• preact-ts
• lit-element
• lit-element-ts
• svelte
• svelte-ts

Wait! Why CRA is slow?​

When it comes to building a React app, many developers of different skill levels have long relied on Create React App (CRA). I was also using CRA for a long time. However, I noticed it has some substantial shortcomings in terms of speed and performance. CRA can be a little sluggish when starting a project. Things get even sour while building a development server which ends up taking about 5-10 minutes. The reason for the slow behavior is that CRA bundles the code with Webpack. Because Webpack bundles the entire code, if your codebase is more than 10k lines long, you may notice a slower start on the dev environment and a significant wait time to view the changes made. Even with Hot reloading enabled in some frameworks, update speed suffers significantly when more code and dependencies are added to the program.

What magic does Vite do?​

Vite has chosen a different approach to this, almost in reverse. Vite launches the server immediately and uses esbuild to pre-bundle dependencies that don't change frequently. This makes vite 10 times or at times 100 times faster. This is because it helps improve page speed and convert CommonJS / UMD modules to ESM.

The application modules are divided into 2 parts -

1. Dependencies

   The majority of dependencies are ordinary JavaScript that does not change much throughout development. Some huge dependencies (for example, AntD) are likewise rather costly to handle.

2. Source Code

   Source code frequently contains non-plain JavaScript that must be transformed (e.g., JSX, CSS, or other components) and will be modified frequently. Furthermore, not all source code must be loaded at the same time (e.g. with route-based code-splitting).

Bien joué Vite - Performance CRA vs Vite!​

Conclusion​

Vite effectively relieves us of that task, and Vite has a community at its disposal to investigate concerns. This implies we may put our confidence in our tools. Vite effectively relieves us of that task, and Vite has a community at its disposal to investigate concerns. This implies we may put our confidence in our tools.

Thank you for taking the time to read my post. I hope you enjoyed it. That’s all I wanted to share with you. Have a great day!

Till then Au revoir, frérot​

Find me here 🤩

Twitter

Website

Github