344
Productivity & Workflow355
Automation & Workflow225
Software Development251
Marketing & Growth193
AI Infrastructure & MLOps176
Writing & Content Creation204
Data & Analytics142
Photography & Imaging156
Design & Creative170
Customer Support133
Sales & Outreach125
Education & Learning131
Voice & Speech135
Operations & Admin87
Replit CEO Amjad Masad says “self-driving companies” may emerge, where AI agents handle routine work and people focus on judgment and decisions.
In short: Replit CEO Amjad Masad says more companies will use AI agents for routine tasks, while people focus on higher-level decisions.
Amjad Masad, CEO and co-founder of Replit, shared his view of the future of work at TechCrunch Disrupt in San Francisco. He described what he calls a “self-driving company,” meaning a business where AI agents do much of the day to day work.
An AI agent is a software helper that can take actions on its own, like checking information, summarizing it, and starting the next step. Masad compared the idea to a self-driving car (the car does most of the driving, but a human still sets the destination and may need to take over).
Masad said Replit’s own workflow has changed as these agents became more reliable earlier this year. For example, when a customer support request comes in, an agent looks it over first and gathers context by checking places where work is tracked, such as bug reports and internal task lists. By the time a person sees the ticket, they already have the key background to decide what to do next.
Masad also said that once a support worker decides a fix is needed, they can start another agent to help create and ship that fix. In his telling, agents do much of the “grunt work,” but a human still provides judgment and “taste,” meaning what is worth doing and what good looks like.
If more companies adopt this approach, some jobs may shift toward planning, reviewing, and deciding priorities, and away from repetitive tasks. It also raises practical questions, like how businesses check the agents’ work and who is responsible when an automated step goes wrong.
Source: TechCrunch AI