Replit’s CEO says the future of work is the ‘self-driving company’
The self-driving car analogy only holds up until the first crash. A vehicle still needs a licensed human ready to take over when the system meets an edge case, but Masad talks about companies where agents handle the daily grind and people just make the big calls. That sounds nice until an agent quietly misfiles a contract or burns through a budget, and nobody noticed because they were busy being high-level.
I think the real issue is who owns the failure. If an autonomous coding agent ships a broken feature, the company can't blame the model. It still has to answer to customers and regulators. Replit has built a solid platform for individual developers, but scaling that trust to an entire organization is a different problem. I'd want to see specifics on human oversight, audit trails, and how a "self-driving" company handles the mess when the road turns. Until then, it's a vision with the same blind spot every automation pitch has: perfect execution without accountability.