How to learn a programming language using AI

Published on:

Whether or not you’re new to software program growth or you’ve many years of expertise, there’s all the time room to be taught one thing new. The TIOBE Index tracks the highest 50 hottest programming languages, with many ecosystems presenting alternatives for profession development and lateral shifts. Given the breadth of applied sciences obtainable, it may be difficult to search out the time to be taught a brand new ability and to do it successfully.

Just lately, I’ve been making an attempt to be taught the Rust language, a type-safe language constructed with efficiency, reliability, and productiveness in thoughts. In doing so, I’ve discovered a couple of methods for utilizing AI coding assistants that I wish to share with you to enhance your studying expertise.

By the tip of this put up, it is best to have a couple of new AI-based abilities you could apply throughout your studying journey and speed up towards your targets. Within the scope of this text, after I say AI, I imply the AI-powered coding assistants primarily based on massive language fashions, corresponding to Amazon Q Developer (previously CodeWhisperer), GitHub Copilot, and JetBrains AI Assistant. You’ll be able to apply all the ideas under to the instrument of your alternative.

- Advertisement -

Be suspicious of parametric data

“Parametric data” is info saved inside the mannequin throughout coaching. The encoded information is why AI can usually reply rapidly with correct info. Nonetheless, as you’ve possible seen within the information, AI can get the reply spectacularly incorrect at different occasions.

What does this imply for us as customers? Whereas these methods will be useful, they’ve limits constrained by the point and assets required to gather information and prepare the mannequin. For instance, OpenAI has skilled the ChatGPT 4 Turbo mannequin on info as much as December 2023, and the information set’s contents are unknown. From a person’s perspective, there’s uncertainty about whether or not “full” data exists inside any explicit mannequin and why typically we could also be underwhelmed by its responses.

Being aware of the cutoff date for the information set will help you higher perceive and course of the responses out of your AI chat classes. As a client in search of a service to buy, researching a supplier’s data-gathering practices and coaching course of can result in a extra satisfying expertise.

- Advertisement -

Whereas AI will be typically useful, verified human sources of information will nonetheless be probably the most helpful throughout any studying course of. That’s to not say that the AI methods are all the time incorrect, however that you must get into the behavior of questioning whether or not the data you’re consuming is appropriate. An extra cause to be cautious is that the creators of those methods skilled them to be useful above all else. Generally they are often “too useful,” hallucinating concepts and ideas that will not exist.

In abstract, it takes “actual intelligence” to grasp when AI makes errors.

Learn code and immediate for rationalization

The open-source software program motion has offered each studying programmer with quick access to manufacturing code. Open-source code bases supply a wonderful alternative to see how the professionals write and to select up language idioms, methods, and extra. However studying an unfamiliar language will be daunting and downright complicated with out a body of reference or primary understanding.

See also  Truecaller partners with Microsoft to let its AI respond to calls in your own voice

A method I’ve been utilizing is discovering code on the favored code-sharing website GitHub for explicit languages and pasting it into an AI chat session, together with the prompts, “Please clarify this code,” and “please checklist the vital language ideas taking place inside this code block.”

A template for this immediate would possibly look one thing like this.

Please clarify the next code


// rust code goes right here


Additionally, checklist the vital language ideas from the outline in a bulleted checklist so I’ll do extra analysis.

The ensuing checklist of matters is a wonderful option to focus my studying on what is important for the time being slightly than attempting to soak up a whole library’s value of knowledge directly.

The JetBrains AI Assistant will allow you to retailer immediate templates for reuse, which is extremely helpful as you bounce between totally different initiatives.

- Advertisement -

Immediate for verbose inline feedback

A wall of code will be very intimidating if you nonetheless haven’t absolutely discovered the syntax or semantics of a language. Utilizing the immediate “remark every line” is a fast and simple option to get a common concept of what an utility could also be doing.


With the JetBrains AI Assistant, you too can get a Diff view of the modifications in a side-by-side or unified view. The view lets you rapidly evaluate the modifications and select to simply accept or reject them.

Play with totally different choices

Studying any subject entails experimenting and, extra importantly, taking part in with discovered ideas. That is important to discovering a working resolution and understanding when to attempt a special method. With AI, attempting out variant implementations has by no means been extra simple.

Right here’s a immediate I’ve used to find options within the Rust language that permit me to comply with up and do extra analysis.

Given the next API, present me three totally different implementations


fn add(x: i32, y:i3) -> i:32  {

    // implementation goes right here



Keep in mind, that is about studying, so the APIs don’t should be notably refined. Taking part in with concepts lets you uncover totally different studying paths exterior the AI chat session. The approach is great, particularly in comparison with conventional subject looking, which can have limits primarily based in your understanding of a selected topic.

Some instruments, like Amazon Q Developer, supply choices inline, permitting you to cycle by way of examples with out leaving the context of your editor. Urgent the correct and left arrow keys permits you to transfer between choices till you discover one you want.


Extra context is all the time higher

I discussed the embedded info limitations of LLMs above. Remember the fact that, the place these companies could lack info, you’re in an ideal place to supply it. However that you must go about it the correct means.

See also  The Future of Cybersecurity: AI, Automation, and the Human Factor

The fashionable web search expertise has skilled us to ask snappy keyword-based questions in textual content packing containers. Search-style queries are a typical mistake I see many newcomers to AI make, and it could actually depart them underwhelmed with the outcomes. Fascinated with AI chat classes as “search” is a foul behavior to use when utilizing AI assistants, as creators of LLMs constructed them to foretell what it’s your decision.

One of the simplest ways to get higher predictions is to be as prolonged and specific as attainable.

An method that works nicely is example-based prompting. The extra examples you possibly can present, the higher your outcomes might be. Right here’s a template immediate that would assist you to uncover new concepts in your studying journey.

Given the next three examples, what would an implementation for <X> seem like?

instance 1:



instance 2:



instance 3:



The information you enter right into a chat session gives the context wanted to realize your required final result. Don’t be afraid to appropriate or add extra context as you go. The extra, the higher, because the mannequin has extra info on which to base responses.

To maintain the dialog going and to extend the chat session context, Amazon Q Developer offers you a set of pure follow-up inquiries to hold you engaged within the studying course of.


Peek at code completion

Most AI companies supply multiline code completion. Whereas it could be tempting to simply accept all of the choices offered within the editor, this usually impedes my progress towards studying. As an alternative, I like to start out by making a code remark that signifies my purpose.

// TODO: create a match expression to course of the totally different message variants

Within the following determine, we see GitHub Copilot providing to finish my assertion. It seems good at first look, however I have to pause to digest the choice and whether or not it fulfills my intention. Earlier than urgent Tab, which may be very tempting, I must cease. Why?


The reason being to test whether or not the code is one thing I can learn and course of with my present ability set. Whether it is, then I’m making progress. If it’s not, then I ought to take time to grasp the place the gaps in my data exist. You must by no means settle for code you don’t absolutely comprehend into your code base.

Word you could configure this characteristic to be much less intrusive and solely show completions on-demand to cut back the cognitive load of switching between writing and studying code.

Clarify errors and discover options

With each programming stack, you’ll run into compilation and runtime errors. A few of these errors will be cryptic. When you’re struggling to grasp why your utility is failing, use an AI assistant chat session to clarify the difficulty, discover the place it happens, and suggest an answer.

What precisely is the issue right here in my Rust backtrace? Scale back the response to the file and line and rationalization of the error and suggest an answer.



Right here is an instance of utilizing this immediate to grasp what occurred in my Rust backtrace.

See also  Interview: Cinderella Amar - Web4, AI, and Blockchain evangelist

Share concepts

Whereas transferring from one subject to a different with AI chat classes could really feel pure, it’s all the time good to decelerate and recontextualize your newly discovered data with others. As soon as I’ve discovered a brand new idea and created a working pattern, I share it with different learners and specialists to get helpful suggestions. Keep in mind, AI can solely be useful concerning your requests, whereas fellow people will enrich you with their lived expertise and predict potential pitfalls chances are you’ll be oblivious about.

Social media platforms are an amazing place to share screenshots, code samples, and concepts and obtain helpful suggestions you could incorporate into future AI chat classes.

Summarize the chat

And right here’s a tip for people (like me) who will not be good at taking notes however would possibly wish to recall their earlier day’s efforts. When your neurons are firing on the finish of a protracted studying session, ask your present chat session to summarize all of your questions and checklist a single-sentence reply for every query.

Given what we’ve chatted about, checklist all of the questions and a single-sentence abstract of every reply.

The response is a wonderful fast reference information in your subsequent studying session. It additionally makes an amazing weblog put up the place you possibly can follow sharing your concepts and studying experiences with others.

Study quicker with AI

I hope you discovered the following tips useful as you employ AI instruments to be taught quicker and extra successfully. These instruments are about enabling you and serving to you obtain the targets you set for your self. I’ve discovered them useful in forming new ideas and exploring ones I didn’t know existed.

Most significantly, they allowed me to attach with communities of different people who broaden my data. When you have some other ideas for accelerating your studying utilizing AI, please share your ideas and ideas with me and others.

Khalid Abuhakmeh is a software program developer with 16 years of coding expertise, specializing in Microsoft .NET applied sciences. All through his profession, he has held numerous software program developer titles, from junior developer to director of software program growth. At the moment, he works as a developer advocate for JetBrains, specializing in the .NET ecosystem.

New Tech Discussion board gives a venue for know-how leaders—together with distributors and different exterior contributors—to discover and talk about rising enterprise know-how in unprecedented depth and breadth. The choice is subjective, primarily based on our decide of the applied sciences we consider to be vital and of best curiosity to InfoWorld readers. InfoWorld doesn’t settle for advertising collateral for publication and reserves the correct to edit all contributed content material. Ship all inquiries to

- Advertisment -


- Advertisment -

Leave a Reply

Please enter your comment!
Please enter your name here