It’s now widely accepted that AI will disrupt the coding job market, but it could also end up making average coders extremely capable. Replit CEO Amjad Masad recently made a bold prediction: the gap between an average Replit user and a senior Google engineer will all but disappear within the next two years. This startling claim hinges on the rapid advancement of AI autonomy and its integration into coding platforms like Replit. Masad’s assertion, rooted in recent research and observations of Replit’s power users, paints a picture of a future where AI agents become indispensable coding partners, leveling the playing field for developers of all skill levels.
“There’s a recent paper that showed that AI autonomy is doubling every seven months,” Masad said. “So if today an AI can work uninterrupted and remain coherent for 15 minutes, then in seven months it’ll be 30 minutes, and another seven months it’ll be an hour. I actually think it’s accelerating. I think it’s going to be more like every quarter we’re going to see a model that has increased autonomy.”
He points to recent developments as evidence. “We’ve already seen OpenAI’s o3 reasoning model, it can work and call 600 tools per session. It can work up to an hour or something like that. So I think we’re in this massive trajectory where AI models are going to be this sort of agentic; they’re going to provide this agentic experience such that the human developer is actually just spinning up the number of agents that can build their ideas.”
Masad observes this trend already taking hold within the Replit community. “Actually, we see our power users on Replit have three, four, five agents running at any given time, and they’re sort of going between the agents and giving them feedback,” he says. “I feel like at some point over the next couple of years, there’s not going to be much daylight between a Replit user and a senior engineer at Google.”
It’s a brave prediction. Senior engineers at Google are some of the best and highest-paid engineers in the world, and have years of experience tackling some of the toughest engineering problems. But if AI systems can enable a fresh young programmer to operate at their level, it could democratize software engineering to degrees previously not thought possible. But conversely, if this does indeed come to pass, it will likely impact the salaries and job prospects of senior engineers who’ve worked for years to build their skills.