How AI Could Change Development

Intellisense was the Microsoft version that I first used extensively, but similar tools have existed since the 1990s. Intellisense allowed developers to get in-line assistance as they were typing code, akin to the autocomplete tools that people first used on mobile devices. It greatly accelerated development by suggesting functions, methods and variable names that you were likely to need. Developers stopped having to memorize exact names for things and only needed to remember how something started. It also became an easy way to see available options, with pop-up documentation snippets that helped developers learn about new options.

I believe AI tools will have analogous benefits to the code completion tools of the past, but at a 10-times or 100-times impact. Instead of autocompleting a variable name, the tools can autocomplete an entire line of code. Instead of showing you what functions are available, it will suggest the block of code based on your desired functionality. Pseudocode will be enough for the tools to write the code needed, and it will be based on the code it already sees in the project.

Junior developers can learn to explain their problems better.

Going from an academic setting into a more structured environment always results in huge amounts of technical growth. Having teams that rely on you for results, and with tasks you didn’t imagine on deadlines you didn’t always dictate, creates a sink-or-swim mentality that I have seen people succeed in time and time again.

The challenge with current AI coding tools is that they still require a lot of context to be useful, and this will likely be the case for quite a while. The better the prompts, and the more thorough the context, the better the results I’ve seen. For junior engineers joining existing teams, this means the tools can be given the context of the current project, so even a mediocre prompt will probably return decent results. But a fundamental understanding of architecture, data types and structures and design patterns will still be beneficial. Being able to use the right language in creating a prompt will result in higher quality results faster.

Great developers could get even better.

Developers who have more experience will likely benefit even more from AI tools than junior devs. Much like being able to use Intellisense-like tools even faster once they know the framework they’re working in, they will be able to give more specific prompts about their needs, even generating large blocks of functionality that can be quickly tweaked when analyzed. A few more minutes explaining the need will result in results practically equivalent to a senior developer in a fraction of the time.

As the tools mature and become more integrated into their workflow, I can imagine senior developers treating their AI tool as a pair programmer, asking it to code simple blocks and only making minor suggestions. Over time, and as a project grows, the AI tools will have more context and be able to take on larger tasks with more sophisticated prompts. And with newer tools being able to translate UI into code given some guidelines and guardrails, a senior engineer could spend their time mostly editing AI code instead of writing it new from scratch.

All developers may go up in value due to these changes.

With the advancement of AI tools, I think teams will be able to do much more in the same amount of time. Some companies will think about this in a cost-saving way and reduce the number of engineers they need, but the more forward-thinking ones will see this as a way to improve velocity and code quality. Even non-developers may be able to jump into code more confidently and make edits where needed, using AI assistants to guide them on how to make changes. I think this makes all developers more valuable, able to get more done in less time and providing value that must be rewarded by companies.

The future of AI-assisted development is bright, and it should not be seen as a threat to engineers but as a benefit that could transform the speed and quality at which we’re able to innovate and produce new ideas across every industry.

Originally published by Forbes