Why AI still won't replace software engineers

A recent Anthropic study found that AI already covers roughly 75% of tasks performed by computer programmers, but Replit president and head of AI Michele Catasta argues that AI is not replacing software engineers—it is changing the nature of their work. Catasta told The Deep View that engineers now spend more time verifying AI-generated code, and that the technology has made product managers and designers 10x more productive by enabling rapid prototyping. He acknowledged some job displacement in the interim but emphasized that AI is creating more work, not less, as teams accelerate project cycles.

Software engineering was once seen as a ticket to a stable, well-paying career. Now, the profession is in flux. As generative AI tools like ChatGPT and Claude Code become increasingly capable of writing code, software engineers are beginning to wonder how their roles will change. A recent Anthropic study https://www.anthropic.com/research/labor-market-impacts? bhlid=d84746f67cfd4f522735473b61c03aad1378962a found that computer programmers are among the occupations most exposed to AI, with the technology already covering roughly 75% of the tasks they perform, raising questions about the future of the field. Some business leaders say the answers aren’t so simple. One of them is Michele Catasta, president and head of AI at Replit, an AI coding platform used by more than 50 million people to build software using natural language and AI agents. As one of the companies ushering in a new generation of vibe coders, Replit has a front-row seat to the profession’s transformation. The Deep View sat down with Catasta to discuss how AI is reshaping the role of the software engineer, the democratization of software development, and why entry-level talent may have a competitive advantage in the AI era. The conversation has been edited for brevity and clarity. Aaron Mok: In a recent podcast interview, CEO Amjad Masad described Replit’s Agent as "an automated software engineer" that's as capable as a “mid-level engineer at Meta or Google.” That is a bold claim. From your conversations with customers, what are some of the most surprising ways you've seen them use the technology? Michele Catasta: The thing that's been most surprising to me—even though I've been working on this for 10 years—is that I practically don't hear users say they can't build what they have in mind anymore. That wasn't the case a year ago. We've gone from people creating landing pages to building software that would've taken weeks or months earlier in our careers. I think what Amjad was trying to explain is you're not going to be able to take Replit’s Agent, teleport it into Meta tomorrow, and replace all the software engineers. It's more that someone who already has that level of skill can suddenly build a new product or start a business on the side. We're also seeing strong enterprise traction. Eighty-five percent of the Fortune 500 uses Replit to automate repetitive work. Before, if you had an idea for a product, you'd write a product requirements document, hold several meetings, and design prototypes. Now, you can show up to that first meeting with a functional prototype. Decision-makers can immediately decide whether it's worth building or whether it needs another iteration. If you're a product manager or designer today, your skills are basically 10x overnight because you have way more throughput than you had before. Mok: How are you seeing AI changing the role of the software engineer? Catasta: I don't think anyone is doing less work today than before. If anything, this has been the most restless period in tech. What's changing is the nature of the work. The cost of generating code is dropping. What's taking more time now is verifying whether it's correct. AI can generate a lot of code, but engineers still have to review it and decide whether to accept those changes. It's not just engineers. Managers are overseeing more—and shorter—projects because teams can move faster. At the leadership level, the speed of the company means far more information is flowing upward, making it harder to process everything. We're all rethinking our jobs on a daily basis. Mok: It sounds like AI is creating more work than less. Agents can generate thousands of lines of code, but someone still has to verify it and make decisions. Is it a misconception that AI is making software engineers less important? Catasta: Yes, fundamentally I think that's a misconception. I don't want to discount the fact that there will be some job displacement in the interim. Small software shops that don't adapt are going to struggle as companies become capable of building more internal tools themselves. Some shifting will happen, like it always does in this industry. But the companies that have traction and are building something valuable. I don't know any founder in my space who isn't hiring like crazy when their company is growing, because that's the amount of work there is to be done. Even if the role of an engineer is changing, maybe they're spending less time sitting in front of a screen typing lines of code. If you walk into the Replit office today, the amount of debate happening between engineers is even higher than before. You can leave an agent writing a pull request while you're discussing with your colleagues: Should we be doing X or should we be doing Y? What are the trade-offs? As the cost of generating code drops, there's more time to think. There's more time to debate. There's more time for engineering teams to sit down with designers and product managers and exchange ideas. That's leading the industry to build better things. Mok: There's been growing debate over whether AI could "deskill" software engineers by automating more of the coding process. Do you think that's a fair concern? Catasta: I think it's a bit of a misconception. If you ask most staff engineers today how compilers work or how to write assembly code, a lot of them probably couldn't do it. That's not because they're worse engineers. It's because computing has evolved by creating better abstractions so people can focus on higher-level problems. I think AI is another step in that evolution. Even if an agent writes most of the code, engineers still have to decide whether it's correct, whether it should be deployed, and whether it's the right solution. That's still a skill. If AI automates more of that work, then we'll move another layer up the abstraction stack and spend more time understanding users and solving different problems. That's how software engineering has evolved for decades. Mok: There's growing concern among computer science students and recent graduates that AI could reshape—or even reduce—entry-level software engineering jobs. How do you see the career path for junior engineers changing? Catasta: I know some large companies have stopped hiring below a certain level, and I think that's a huge missed opportunity. For startups like ours, it's actually been fantastic because we're hiring exceptional graduates who grew up in the AI era. They started college around the time ChatGPT came out. They’re AI natives. They've been using AI coding tools from the beginning, and no one is more ready for this revolution than them. Startups can also give them much more scope early in their careers, which is a great opportunity. Mok: So is AI literacy becoming a competitive advantage? Catasta: Absolutely. AI literacy has become table stakes. The employees we hire and the companies we sell to are AI-forward. They expect people to know how to work with these tools. As chaotic as this shift can be, it's also incredibly exciting because people can build much faster. Six months ago, I would've looked at an ambitious project and told my team it would take us a quarter. Now I can say, "It's going to take three weeks. Let's do it." It's extremely empowering to feel like we can move fast and make things happen. Mok: As AI coding tools become more capable, who do you think stands to benefit the most? Catasta: Our focus has always been on knowledge workers who don't come from a technical background.There's a lot of raw intelligence in today's AI models and tools, but very few people know how to turn that intelligence into something genuinely useful. That's why so many companies are still asking what the return on investment from AI actually is. Software is a very broad concept. It's not just about building applications. It's also creating scripts, automations, internal tools, and agents that help people do their jobs. I want that superpower to be in the hands of everyone.