As a developer, finding ways to streamline your coding process can greatly increase your productivity. One tool that has gained popularity in recent years is AI coding assistants. These tools use artificial intelligence to help you write code more efficiently and accurately. In this tutorial, we will explore some of the most popular AI coding assistants available and review my personal experience with each one.
1. Introduction to AI Coding Assistants:
AI coding assistants are tools that use machine learning algorithms to assist developers in writing code. They can provide suggestions for code completion, help with syntax errors, and even offer recommendations for improving code quality. These tools can help both beginners and experienced developers write code more quickly and accurately.
2. Setting up AI Coding Assistants:
To use AI coding assistants, you typically need to install a plugin or extension in your code editor. Some popular code editors that support AI coding assistants include Visual Studio Code, Atom, and Sublime Text. Once you have installed the necessary plugin or extension, you can start using the AI coding assistant features in your code editor.
3. Trying Different AI Coding Assistants:
I have tried several AI coding assistants, including TabNine, Kite, and Codota. Each of these tools offers unique features and benefits, and I will share my personal experience with each one.
– TabNine: TabNine is one of the most popular AI coding assistants available. It offers code completion suggestions based on the context of your code, making it easier to write code quickly. I found that TabNine was very accurate in its suggestions and helped me write code faster.
– Kite: Kite is another popular AI coding assistant that offers code completions and suggestions. One feature of Kite that I found particularly useful was its ability to provide documentation for code snippets, which helped me understand complex functions and methods more easily.
– Codota: Codota is an AI coding assistant specifically designed for Java developers. It offers code completions, suggestions, and even predictions based on your coding patterns. I found that Codota was very helpful in reducing the time it took to write Java code and improved the overall quality of my code.
4. Conclusion:
In conclusion, AI coding assistants can be valuable tools for developers looking to improve their coding efficiency and accuracy. I have found that using AI coding assistants has helped me write code faster, reduce errors, and improve the quality of my code. I encourage you to try out different AI coding assistants and see which one works best for you. Happy coding!
Correction on Sourcegraph Cody:
I wasn't fully using this one correctly. You can download an additional app and link it to your GitHub repository, and it does a much better job of understanding the repo. I'm not sure why the UX doesn't prompt you to do this from the start, but it does fix some of the problems in the video such as over-indexing on node modules. However with all that said, I honestly still wasn't able to get it to perform many tasks that were particularly useful, and the larger the repo I gave it, the less reliable the responses seemed to become. It could definitely be worth a shot in your projects to see if it is useful to you though! 😃
cody is the best AI coding slave tbh
claude 3.5 is i think a tat better then openAI, Personally i wonder are there any local LLM's good for this (who were trained on basic english only though mostly on code, perhaps even for specific languages).
Hey Conner – Do you still think that Cody stands on Acceptable level? They kept on working since the release of this video of yours and recently having the ability to use Claude Sonnet I think it goes top of the list, What are your thoughts? 🙂
So the summary is, use ChatGPT to research your specific task case, Co-Pilot to build the framework, and then Codium to handle possible cases/refactoring of the code. That’s actually insane 😅
so, should i go for github copilot or codium ai?
Thanks for the comprehensive review. Very helpful. 👍🏻👍🏻🚀🚀
Excellent. Your content on this subject has helped me a lot.
I wonder if you could give a similar review for JetBrains AI.
Tabnine is good, but, as mentioned, can't actually work on large projects, like a million source-code files.
notice how AWS Code Whisperer produces the exact same code as Tabnine
when will copilot chat feature come in zed code editor 😭😭
Factorial of a negative number is not NaN, it's undefined, or throw. And that "factorialize" smashes the stack before throwing unhelpful RangeError on most non-number input (not " +004e+00" though). num > 170 => Infinity. num > 21 => loss of precision. ChadG did better.
The true 'Time is of the essence'… I like this one
The fact that you just start the video with no intro is amazing and makes me feel like you value my time. Thank you, and keep it up!
Great video, right to the point. Only thing I'd have liked to see is a cost comparison. But still earned a sub. 👍
Claude Opus smokes Chat GPT4o right now for anyone that is still looking at this.
The web app for Opus has a 200K context window vs 32K for Chat GPT4o, and it just gives better code too.
Claude was able to help me enable high resolution timers on a Giga R1 correctly and Chat GPT couldn't even start without putting itself into a loop.
Try any real project code and youll see what i mean.
5:30 ive been using gpt to practice coding and i realize an llm just gives you the most common string of words that would follow your prompt but dam man sometimes it just needs to tell you
"i dont know" or "thats beyond my abilities" something besides just confidently giving me wrong information.
Great video. Why tabnine is not a level higher un the list ?
Using ChatGPT4o I was able to create a simple version of "Wifite" 2467 lines of code. I used 4 days going back and forth with GPT4o to fix errors as it would just spit out code and tell me "Here is the revised version without the errors"
I used 4 days because I needed to make sure GPT made one function work, then hop on to the next function. It's not something I recommend, because my version of Wifite isn't nearly as extensive as the original version and my line of code is nearly as long as the original while only containing 40-45% of it's functions and a messier harder to debug code.
The reason I don't reccomend it is because:
GPT will after starting to reach 500ish lines of code it will fail to generate the code so you will have to "Regenerate prompt"
It will become super slow to output codes as it grows
Possibly forget certain functions after having to revise the code 3 times because of errors.
My recommendation, use GPT to learn to code, don't depend on it. Example, you can make GPT code something easy using Python. Anything starting to require a UI and more features WILL make you a headache. So why not have the headache actually learning it.
Do any of these help with embedded development where the code needs to follow along with the specific micro-controller API used for each project?