Is Cursor a Blessing or a Curse for the Modern SWE?
As a recruiter in tech, specifically for AI startups, and as a “nontechnical person”, I have to make sure to stay up to date with the trends. Some of my founder partners are all for AI code editors like Cursor; others simply don’t approve. And they have valid reasons for it. So, I’m here to break it down in the simplest way.
First, let’s take a look at some of the top AI code editors & understand how they work. These are platforms that are essentially considered to be a “smart assistant” for a software engineer, primarily helping coders write, fix & improve their code to work more productively. Cursor, Copilot & Windsurf (and others) solve a lot of problems & reduce a lot of errors. Cursor, Copilot, and Windsurf are AI-powered tools that help software engineers write and improve code more efficiently. GitHub Copilot works inside popular coding programs like VS Code and suggests code as developers type. Cursor is built specifically for AI-assisted coding, offering deeper integration with AI. Windsurf is another similar tool that helps automate coding tasks.
Side note: Please don’t use any code editor during an interview process unless the company explicitly says it’s okay. It may seem obvious to most folks, but we’ve seen many cases where candidates use these during a live coding interview, and they always get caught. Not great for optics or your candidacy.
There are two main discourses on the topic of AI code editors, specifically in my startup tech community. And there are excellent reasons for both:
The Argument in Favor
There is a large group of founders we work with who strongly encourage the use of AI-assisted tools. However, they ensure that candidates have a strong fundamental knowledge of coding and discourage AI assistance during technical assessments in the interview process. This perspective often comes from founders seeking more experienced developers.
One globally recognized leader in AI, Andrew Ng, founder of DeepLearning.AI, recently mentioned in his blog The Batch: “As coding becomes easier, more people should code, not fewer!” Whether we use AI assistance or not, there is substantial fear-mongering around AI taking over jobs—including those of software engineers. So, it’s reassuring to hear a perspective from a reputable expert.
Unlike some founders, Ng encourages developers to leverage AI-assisted coding tools. He further states, “One of the most important skills in the future will be the ability to tell a computer exactly what you want, so it can do that for you. Coding (or getting AI to code for you) is the best way to do that.”
The Counterargument
Especially if you’re just starting your engineering career, I know this may seem counterintuitive, but you’re better off avoiding AI assistance. You want to build as much as possible on your own to develop a strong foundational understanding of the code base. There comes a point where you need to move beyond the initial learning phase to become an efficient software developer who can effectively use tools like Cursor.
A handful of startup founders are working on deep and sophisticated technology that requires multiple layers of abstraction. Relying on a code editor too soon can prevent an engineer from developing a deeper understanding. Tools like Cursor can also create dependency, which may hinder long-term growth.
A recently coined term, “vibe coding,” refers to developers working on a codebase without fully understanding what they’re doing. That’s how far removed and unintentional coding can become when over-relying on these tools. Hey, but if you can build your MVP this way, why not? That said, it can only take you so far.
Final Verdict
If you’re a new grad or a junior engineer, get the foundations down & strengthen your competency as much as possible. If you’re an experienced engineer, learn how to optimize for it.
Lastly, these should be used as assistants, not replacements.
What do you think? Are AI code editors making developers better, or are they crutches that could do more harm than good?