With AI, Coders Have More Time for Creative Coding: Malik Drabla, Co-Founder Adrenaline

AI is all over the news. Twitter, LinkedIn, and Instagram are filled with threads and carousel posts showing people how they can best use AI to increase their work productivity.

From generating real-like images and designing brand logos to helping find bugs in code and converting image prompts into websites, recent developments in AI has taken the meaning of automation to a whole new level.

In this interview, we speak with Malik Drabla, the co-founder of Adrenaline - an AI project to help coders make the best of their code. Let’s dive in.

Q: Welcome to Unhashed. Please introduce yourself, and share with us how you started in the artificial intelligence industry. Tell us what made you establish Adrenaline.

Hello, I’m Malik Drabla, co-founder and developer at Adrenaline. I did my Bachelor’s in Computer Science at Columbia University and worked as a Search ML (Machine Learning) Engineer at Google for a few years.

My colleague and Adrenaline’s co-founder, Jon, has been interested in AI since his early days. He has already conducted pioneering research in implementing AI methodologies in neuroscience.

Jon and I go back to high school where we built popular developer tools for efficient software development.

When GPT-3 was released, we realized that AI has enormous potential to improve programmer productivity. In a software-driven economy, any improvement in programmer output will benefit the entire ecosystem.

Thus, we decided to leverage AI for software development by building Adrenaline.

Q: From trading to marketing, artificial intelligence (AI) is creating ripples across multiple sectors. How do you think AI is affecting the software development industry?

I think we are yet to assess the enormous impact of AI on the software industry. To begin with, AI is now playing a crucial role in software design, automatic code generation, testing, debugging, and deployment.

Companies are relying on artificial intelligence for requirement gathering that uses natural language processing techniques to plan software models. These models form the basic framework for designing projects based on customized client requirements.

AI is easing out software deployment with seamless code execution and better upgradation opportunities. This automatically leads to robust data security, and better bug and error identification technologies, thereby assisting in better software development.

We are also witnessing a significant rise in AI-enabled automated coding where anyone can produce code. In this era of AI-generated code, we need the right tools that can help us design, mediate, and understand the codebase.

In the long run, these AI-based tools will be crucial for us to stay in control of our codebases.

Q: Multiple studies suggest that software developers take 20–50% of their time in debugging software code. How can AI help in debugging code?

Yes, developers currently spend a large amount of time reading their code for identifying mistakes rather than writing code. I believe AI can help developers to focus on programming instead of debugging their code.

AI automates most of the routine code analysis which helps programmers to concentrate on innovation. Since AI can identify problems easily with its predictive features, it makes time for creative coding.

So instead of depending on online forums like Stack Overflow, developers can now use AI for their queries and coding assistance. As AI prevents common coding mistakes, it reduces the testing and debugging time.

Some AI tools depend on a library of existing coding mistakes to help developers detect bugs easily and address vulnerabilities. Other AI platforms like Adrenaline help developers answer any questions they have about their code without relying on Google or Stack Overflow.

Q: How does Adrenaline use AI to automate software development and debug code?

Adrenaline helps developers to become an expert in their codebase. Coders can import a GitHub repository or paste a piece of code and Adrenaline uses AI to answer any queries they have about the code.

Codebases are often complex and they become even more difficult to comprehend if someone else writes them. However, it is essential to have a deep understanding of the codebase to run them efficiently.

With Adrenaline’s AI technology, anyone can understand a codebase irrespective of their technical expertise.

Therefore, we’re helping many programmers and developers to level up their knowledge. Adrenaline is providing the right set of tools to understand a codebase inside-out for effective implementation.

Q: It’s a common notion that AI will take away many jobs, including that of software developers. Do you think this is a legitimate concern? How will AI affect software developers and their jobs?

This is an important question. I was going through an Evans Data Corporation survey report where 29% of software developers are wary that AI will take away their jobs. So yes, this is indeed a concern among many IT professionals.

I think AI and platforms like Adrenaline assist developers to understand complex codebases, thereby helping them write better code. So instead of a hindrance, AI can help developers to become even better at their jobs.

Although AI will bring more speed and efficiency to software development, human developers will never become redundant. After all, it is we who are building these AI tools with our creativity and critical thinking.

Therefore, human ideation and innovation will always remain essential and valuable in software development. AI will assist humans to focus more on creative ideas by automating mundane tasks and assisting in software building.

Q: Likely, we’re still in the very beginning of the AI evolution, and we have already seen remarkable growth with GPT-4 and the likes. What’s your overall vision of AI in the next 5 to 10 years from today?

I think AI will have a ubiquitous presence in our lives by the end of the decade. Most major companies are already investing heavily in AI and it will further accelerate in the coming years.

Models like BERT and GPT have already kickstarted a movement towards automating intelligence work. So while these AI systems proliferate and become a part of our everyday lives, we have to ensure they morally and ethically enrich our society.

We’ll soon see a barrage of AI-generated code flooding the market which can potentially result in a maintainability crisis. This will inevitably lead to an asynchronous relationship where AI will generate more code without having adequate tools for understanding them.

I think we will need more sophisticated tools in the coming years to help us mediate and understand the burgeoning flow of AI-generated code. That’s why we’re working on Adrenaline.