The Future of Programming in 2025 with Cursor Team on the Lex Fridman podcast

n In-Depth Conversation with the Cursor Team on the Lex Fridman Podcast

On a recent episode of the Lex Fridman Podcast, Lex sat down with the founding members of the Cursor team—Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger—to discuss the role of AI in programming, the evolution of code editors, and the broader implications for the future of software development.

Cursor, a code editor built on the foundation of VS Code, has captured the imagination of the programming and AI communities by introducing powerful AI-assisted features.

This conversation, however, goes far beyond a single tool—it’s a deep dive into the future of human-AI collaboration in designing and engineering complex systems.

In this blog post, we’ll explore the key themes from the podcast, including the purpose of code editors, the role of fun in programming, the impact of AI on developer workflows, and the long-term vision for programming as a discipline.

Whether you’re a seasoned developer, a curious beginner, or an AI enthusiast, this discussion offers valuable insights into where the field is headed.

What Is the Point of a Code Editor?

A Souped-Up Word Processor for Programmers

Lex kicked off the conversation with a fundamental question: What’s the point of a code editor? Michael Truell, one of Cursor’s founders, provided an illuminating analogy.

For non-programmers, a code editor can be thought of as a “souped-up word processor” tailored specifically for writing code.

Unlike traditional text editing, code has a highly structured nature, and a good code editor leverages this structure to provide powerful features that go beyond simple text manipulation.

Michael explained that traditional code editors offer functionalities such as:

Syntax highlighting: Visually differentiating tokens in the code to make it easier to scan and understand.

Navigation: Allowing developers to jump to definitions, references, or other parts of the codebase, much like hyperlinks on the web.

Error checking: Catching rudimentary bugs in real time to prevent costly mistakes.

These features have been the backbone of code editors for decades, enabling developers to work more efficiently.

However, Michael hinted that the role of code editors is poised to evolve dramatically over the next decade as AI reshapes what it means to build software.

The Importance of Fun in Code Editing

Lex added a seemingly lighthearted but crucial perspective: a code editor should be fun to use. Arvid Lunnemark, another Cursor founder, enthusiastically agreed, emphasizing that fun is an underrated yet essential aspect of software development.

“A big part of being fun is being fast a lot of the time,” Arvid noted. “Fast is fun.”

This focus on speed ties into a broader theme of the discussion: the joy of programming often comes from rapid iteration. Michael elaborated, “One of the things that draws a lot of people to building stuff on computers is this insane iteration speed.”

Unlike other disciplines where progress might be bottlenecked by resources or coordination, coding offers a unique ability to create complex systems quickly with just a computer and a developer’s ingenuity.

The Future of Programming: Humans in the Driver’s Seat

A Vision of Speed, Agency, and Control

The conversation quickly turned to a more speculative and visionary topic: How will the nature of programming change in the coming years?

Michael outlined Cursor’s philosophy, which emphasizes keeping programmers in the “driver’s seat” while leveraging AI to enhance speed, agency, and control.

This vision stands in contrast to a popular narrative in the AI space, where some imagine a future in which developers simply describe software requirements in natural language, and AI builds the entire system autonomously, much like issuing commands to an engineering department over Slack.

Cursor’s founders are skeptical of this approach, not because it’s technically infeasible, but because it sacrifices a critical aspect of programming: control. “It’s much harder to be really specific when you’re talking in the text box,” Michael explained.

When developers delegate too much decision-making to AI, they lose the ability to make the countless micro-decisions that define great engineering.

These decisions—balancing trade-offs between speed, cost, maintainability, and other factors—are often where the real creativity and value of programming lie.

Instead, Cursor envisions a future where AI amplifies human decision-making rather than replacing it.

“As long as humans are actually the ones designing the software and the ones specifying what they want to be built, and it’s not just like company run by all AIs, we think you’ll really want the human in a driver’s seat dictating these decisions,” Michael said.

Navigating Abstraction Layers

One intriguing idea the team explored is the concept of allowing developers to work at different levels of abstraction within a codebase.

Imagine being able to view and edit a codebase as pseudocode, a high-level representation of the logic, while still having the ability to dive into the nitty-gritty details of the formal programming language when needed.

“You can gesture at any piece of logic in your software component of programming,” Michael explained, highlighting the potential for AI to facilitate seamless transitions between abstraction layers.

This approach could offer significant productivity gains while preserving the developer’s ability to make precise, intentional changes.

However, Michael cautioned that this is still a “fuzzy idea” with many details to be worked out. “Time will tell if it actually works,” he admitted, but the guiding principles of control, speed, and human agency remain central to Cursor’s vision.

AI for Specific Tasks, Not Entire Systems

While the team is cautious about over-reliance on AI for end-to-end software development, they see immense value in using AI for specific, well-defined tasks. Arvid pointed out that for certain types of programming—such as fixing a well-specified bug—AI can already be incredibly effective.

However, he emphasized that this is not representative of most programming, nor is it the kind of work that developers find most rewarding.

The Changing Skills of Programmers

A More Fun and Creative Future

Lex raised an important concern: What happens to the fundamental skills of programming as AI tools become more powerful?

Many young developers are worried about their career prospects, wondering if programming will still be a viable and fulfilling path in an AI-dominated future.

Michael’s response was optimistic and enthusiastic: “I actually think this is a really, really exciting time to be building software.”

He contrasted the programming landscape of today with that of a decade ago, noting that much of the “cruft and boilerplate” that once bogged down developers has been alleviated by modern tools and AI.

“Programming today is way more fun than back then,” Michael said. “It’s like we’re really getting down to the delight concentration.”

In this future, the aspects of programming that draw people to the field—speed, creativity, and individual control—are being amplified, not diminished.

Michael predicted that the skills required of programmers will shift away from rote tasks like writing boilerplate code or meticulously avoiding errors. Instead, the focus will be on higher-level skills such as taste, creativity, and the ability to make nuanced design decisions.

“I think that people’s taste and creative ideas will be magnified,” he said.

Rapid Iteration and Experimentation

Arvid shared a concrete example of how AI could transform the developer experience, drawing from a recent migration task at Cursor.

The team needed to replace the use of AsyncLocalStorage in Node.js, known for its poor performance, with a context object—a change that affected the entire codebase.

Even with today’s AI tools, this migration took five days of intensive work. Arvid is excited about a future where AI could drastically accelerate such tasks.

“I am really excited for a future where I can just show a couple of examples and then the AI applies that to all of the locations,” Arvid said. In this vision, AI not only automates repetitive changes but also highlights edge cases, asks for clarification, and allows developers to iterate rapidly.

This shift would reduce the need for exhaustive upfront planning, enabling developers to experiment, try new approaches, and refine their work in real time.

Aman echoed this sentiment, noting that AI enables a more iterative and exploratory style of programming. “It feels more fun,” he said, contrasting this approach with the traditional method of carefully planning every detail before writing a single line of code.

Natural Language as a Programming Language?

Lex posed a provocative question: As AI models improve, could programming shift toward a paradigm where natural language becomes the primary “programming language”?

This idea raises the possibility that developers might spend more time describing systems in plain English rather than writing formal code.

The Cursor team acknowledged that AI is already capable of tasks like translating code between programming languages or automating migrations, but they were cautious about the implications of a fully natural-language-driven programming paradigm.

Michael emphasized that while AI can handle certain well-defined tasks, the creative and decision-making aspects of programming are likely to remain firmly in human hands for the foreseeable future.

For aspiring programmers, the team’s advice was clear: focus on developing a deep love for programming and an understanding of how systems work.

Aman highlighted the importance of passion, noting that the best programmers are often those who are obsessed with coding, spending their free time on side projects and diving into the details of complex systems.

The Hybrid Engineer: Cursor’s Manifesto

Engineering Genius

The conversation concluded with Lex reading from Cursor’s manifesto, titled “Engineering Genius,” which encapsulates the team’s vision for the future of programming:

“We are an applied research lab building extraordinary productive human AI systems. To start, we’re building the engineer of the future, a human AI programmer that’s an order of magnitude more effective than any one engineer.

This hybrid engineer will have effortless control over their codebase and no low entropy keystrokes. They will iterate at the speed of their judgment, even in the most complex systems.

Using a combination of AI and human ingenuity, they will outsmart and out-engineer the best pure AI systems.”

This manifesto underscores Cursor’s commitment to building tools that empower developers, not replace them.

The concept of the “hybrid engineer”—a programmer augmented by AI to achieve superhuman productivity—captures the essence of the team’s mission.

By eliminating “low entropy keystrokes” (repetitive, low-value tasks) and enabling developers to “iterate at the speed of their judgment,” Cursor aims to unlock new levels of creativity and efficiency.

Making Programming More Fun

As Lex noted, even if Cursor doesn’t achieve its grand vision overnight, it is already making programming more fun for hundreds of thousands of developers.

Features like Cursor’s “super tab”—which allows developers to inject intent and shape code generation interactively—exemplify the team’s focus on high-bandwidth communication between humans and computers.

“Pressing tab is just pressing tab,” Arvid said, “but what you’re actually doing is injecting intent all the time while you’re doing it.”

Sualeh added, “The communication to the computer just becomes higher and higher bandwidth as opposed to just typing as much lower bandwidth than communicating intent.”

This shift toward intent-driven programming is at the heart of Cursor’s innovation, offering a glimpse into a future where developers can focus on what they want to create, not how to create it.

A Bright Future for Programmers

The conversation with the Cursor team on the Lex Fridman Podcast paints an inspiring picture of the future of programming.

Far from being a field threatened by AI, programming is poised to become more fun, creative, and accessible than ever before.

By leveraging AI to handle repetitive tasks, accelerate migrations, and enable rapid iteration, tools like Cursor are empowering developers to focus on the aspects of programming they love most: designing systems, making trade-offs, and bringing their creative visions to life.

For aspiring programmers, the message is clear: this is an exciting time to enter the field. Embrace the fundamentals, cultivate a love for coding, and stay open to the evolving role of AI in your workflow.

As the Cursor team’s manifesto suggests, the future belongs to the hybrid engineer—a programmer who combines human ingenuity with AI-powered tools to achieve extraordinary results.

What do you think the future of programming holds? Are you excited about the role of AI in code editors, or do you have concerns about losing control to automation? Share your thoughts in the comments below, and let’s continue the conversation.

This blog post is based on a transcript of the Lex Fridman Podcast, featuring Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger of the Cursor team. For the full conversation, check out the original podcast episode.