“Isn’t it kinda crazy that nobody is really using AI agents?” asked Josh Miller, who runs The Browser Company, in an August 2026 post. AI agents are programs that complete a task on their own once given a goal. The post spread fast because it pointed at something ordinary, a friend group all picking up Instagram or TikTok within weeks. Engineers and early AI adopters already have that experience with agents. Almost no one else does.
The Stories feature on Instagram and Snapchat spread that way years earlier, with no rollout plan. Inside the companies, that spread was tracked as engagement, because more time spent posting and checking meant more advertising sold. To Gen Z, who grew up posting them, Stories were a casual way to share a moment. What a company tracks rarely matches what a person feels. The same mismatch divides the people who build agents from everyone outside tech who might use one.
Consumer technology tends to spread when it solves one plain problem. The delivery app DoorDash replaces cooking or a drive across town with food at the door. Most AI agents solve a different problem, speeding up work already underway at a desk. Few are built around a single result that someone outside tech would recognize right away. OpenAI’s agent tools, Codex and ChatGPT Work, had about 10 million weekly users combined by late July 2026, Wired reported. OpenAI’s ChatGPT chatbot alone had roughly a billion monthly users.
Even a product that solved one plain problem would hit a second wall, the story already told about AI. Two ideas sit underneath most conversations about it: that AI costs people their jobs, and that it draws too heavily on water and electricity. Whether either idea holds up in every case is beside the point once it has spread this widely. Someone who hesitates to add another AI tool to their day is making a reasonable judgment.
That story also shapes who shows up enthusiastic about agents. The people pushing hardest for them tend to run their own businesses or sit in executive roles, looking for ways to do more with fewer people. Someone in a regular job, already carrying a full workload, has little reason to welcome a tool built to make them more efficient. At that level, greater efficiency usually means more tasks for the same paycheck.
