You have 1 article left to read this month before you need to register a free LeadDev.com account.
Your inbox, upgraded.
Receive weekly engineering insights to level up your leadership approach.
Estimated reading time: 3 minutes
The CEO of famed Silicon Valley startup accelerator Y Combinator says that around a quarter of its current crop of companies used AI to write 95% or more of their code.
“What that means for founders is that you don’t need a team of 50 or 100 engineers,” Garry Tan told CNBC. “You don’t have to raise as much. The capital goes much longer.”
Taken at face value, the implication of Tan’s comments are wide-ranging, potentially shrinking software startup hiring pipelines and putting more investor pressure on founders to reach profitability sooner than ever before.
“There’s a lot of anxiety in the job market, especially from young software engineers,” Tan said. “Maybe it’s that engineer who couldn’t get a job at Meta or Google who actually can build a standalone business making $10 million or $100 million a year with ten people – that’s such a powerful moment in software.”
It’s worth noting that around 80% of that cohort of companies were building some form of AI product, with a handful of robotics and semiconductor startups, so there is likely a natural propensity towards AI adoption in the cohort.
More like this
Lowering barriers to entry
“As someone who’s spent most of their career in small tech startups, the Y Combinator model is certainly the gold standard,” says Randall Degges, head of developer and security relations at developer security platform Snyk.
Startup culture has long been defined by the ability to move incredibly fast: add features, fix bugs, pivot when you run into product-market fit issues, and so on. “AI is a great enabler to moving quickly in these areas,” he says.
For certain engineers, this shift brings down barriers to turn an idea into a business. “If I was laid off today, I’d likely spend a good deal of time building my own businesses using AI as it’s accessible, fun, and with sufficient experience,” Degges said. “It’s far less risky than it used to be.”
A pinch of salt
Not everyone is sold on Tan’s evangelical vision.
“Tan’s claims about AI writing 95% of the code for a quarter of Y Combinator startups must be taken with a pinch of salt,” says Martin Reynolds, field chief technology officer at the software delivery platform Harness. “While they do reflect a significant and growing trend in the use of AI in software development, they don’t take into account the additional work software engineers will need to undertake in reviewing and debugging that code to ensure quality and functionality.”
Having that human in the loop is vital. “In reality, the future of coding will likely involve a combination of human creativity and AI-driven efficiency,” said Reynolds.
Hugo Farinha, cofounder at Virtuoso QA, an AI-led quality assurance testing provider, agrees. “The idea that AI can write 95% of a company’s code is compelling, but it oversimplifies what software engineering actually entails,” he says.
“The real value of engineers lies in architecture decisions, solving complex implementation challenges, optimizing performance, and ensuring scalability and security – things that require deep technical expertise, creativity, and problem-solving.”
London • June 16 & 17, 2025
Speakers Gergely Orosz, Camille Fournier and Lara Hogan confirmed.
The future of startups
Farinha’s skepticism about the idea that a company that’s made off the back of 95% AI-generated code stems from the fact that the product will eventually have to evolve beyond a workable minimum viable product (MVP). “Real-world software is not just about producing lines of code, it’s about building products that work under real constraints, handle edge cases, and evolve with user needs,” he says. “AI can help with some of this, but it still requires human oversight to ensure correctness, security, and long-term viability.”
The shape and size of startup engineering teams will likely shift as entrepreneurial founders “vibe code” their way to MVP, and only bring in skilled engineers at a later date to polish up the product for customers.
“Although AI code is becoming more common in the workplace, I don’t think it would ever be a full replacement for software developers,” says Tim Ord, a web developer. “Start-ups might see the initial benefit, but that’s because they aren’t yet at the stage where they need to know how to combine technology with what a business needs.”
