AI tools won't get products out faster, but they'll solve 2 coding problems, says a16z partner

Business Insider

5 days ago

AI isn't making software developers dramatically more productive, but it is solving two of their problems: code quality and morale, said a general partner at Andreessen Horowitz.
Martin Casado, who leads the $1.25 billion infrastructure fund at a16z, said on an episode of the "Twenty Minute VC" podcast published Monday that AI coding tools like Cursor aren't supercharging development speed.
"Every company I work with uses Cursor," said Casado, who is also an investor in the AI coding startup. "Has that increased the velocity of the products coming out? I don't think that much."
"The things that are hard remain really hard," Casado said. This is especially so for infrastructure companies, where developers still need to make core architectural decisions and trade-offs that AI can't handle.
Where AI shines, he said, is in eliminating the drudge work for developers: writing tests, generating documentation, and cleaning up messy code.
AI can help create "more robust, maintainable code bases with less bugs," the longtime infrastructure investor said. "It could really help with the development process."
Casado also said AI tools have made coding feel fun again, especially for longtime developers.
The investor said he uses Cursor to handle finicky processes like setting up infrastructure or picking the right software packages, which lets him "focus on what I want and the logic."
"It's almost like it's brought coding back," he said. "These old systems programmers, like, you know, vibe coding at night just because it's become pleasant again."
Casado and a16z did not respond to a request for comment from Business Insider.
AI empowering '100x engineers'?
Agentic AI coding tools have taken over much of software engineering, writing code for developers, sometimes with minimal human editing necessary.
Tech leaders have been vocal about the productivity boost.
Surge AI's CEO, Edwin Chen, said the era of "100x engineers" is here.
"Already you have a lot of these single-person startups that are already doing $10 million in revenue," Chen said on a recent episode of the "Twenty Minute VC" podcast. "If AI is adding all this efficiency, then yeah, I can definitely see this multiplying 100x to get to this $1 billion single-person company."
"It often just removes a lot of the drudgery of your day-to-day work," Chen said. "I do think it disproportionately favors people who are already the '10x engineers.'"
But some industry leaders said the AI coding hype comes with trade-offs.
GitHub's CEO, Thomas Dohmke, said using AI coding tools might slow down experienced engineers. On a podcast episode released in June, he said a worst-case scenario is when a developer is forced to provide feedback in natural language when they already know how to do it in a programming language.
That would be "basically replacing something that I can do in three seconds with something that might potentially take three minutes or even longer," Dohmke said.
OpenAI's cofounder Greg Brockman also said using these tools has stuck humans with the less enjoyable parts of coding.
He said the state of AI coding had left humans to review and deploy code, which is "not fun at all."