Why AI Agents Haven't Caught On With Regular People

Silicon Valley is seized by what it sees as the next frontier of AI: agents—software designed to autonomously complete tasks on behalf of users. Yet in a striking disconnect, the general public has barely engaged with them. Josh Miller, CEO of The Browser Company, crystallized this tension in a viral X post this week, writing, "Isn't it kinda crazy that nobody is really using AI Agents?" He noted that while the technology is theoretically ready to transform how people work and live, "the general public dgaf" (doesn't give a fuck). Miller's provocation resonated widely in an industry broadly bewildered by the public's indifference to its latest obsession.

The numbers are stark. OpenAI disclosed that its Codex and ChatGPT Work agents collectively have about 10 million weekly users. People close to Anthropic report that its Claude Code and Cowork agents are seeing similar levels of adoption. By comparison, chatbots like ChatGPT and Gemini each have around a billion monthly active users on average—making agents, in Miller's phrase, "basically a rounding error." This matters because tech companies have spent billions training models capable of far more than simple text generation and retrieval, and agents represent the primary strategy for monetizing those advanced capabilities. If mainstream adoption doesn't materialize, the return on that investment becomes questionable.

Miller's credibility on the subject is substantial. The Browser Company's Arc was acquired by Atlassian last year for $610 million; his first company, Branch, was bought by Facebook in 2014; and he served as the White House's first director of product under President Obama. When asked what would drive adoption, he argued the problem isn't technical readiness but product misalignment. "No one wants AI agents, because AI agents aren't a thing," he said. "It is an invented frame made up by our industry to collectively refer to something." His point: tech companies are shipping the most impressive things their models can do—navigating websites, writing code, completing autonomous workflows—rather than products tailored to what consumers actually desire. The Browser Company's most popular feature illustrates the alternative: a personalized morning briefing for users in its AI-powered browser, Dia. When people open their laptop, they receive a homepage with a greeting, a daily to-do list populated from their calendar and email, and tidbits designed to spark joy, such as art. Technically, the feature relies on an AI agent, but users experience it as a coherent, joyful product, not as a demo of autonomous capability.

Miller traced this misalignment partly to groupthink within AI labs. During a period when The Browser Company explored acquisition, he met with leaders at top AI labs and found a striking pattern: "I think every single lab but one mentioned the movie Her as the way they articulated their vision," he said, referencing the 2013 Spike Jonze film about human-AI intimacy. While respecting bold vision-setting, Miller saw a lack of diversity in convictions about what AI agent products should actually be. The latest trend has been letting users build personal software to automate specific tasks—pitch decks, dataset organization, and other workflows. Yet this appeals only to a limited pool of users already interested in building their own productivity tools. Reaching mainstream adoption, Miller argued, will require bigger ideas. His call to the industry was direct: "Let's not just accept this narrative of AI agents and actually question what are good products and tools that are joyful, useful, and approachable. Who gives a shit if it looks like an AI agent? No one uses AI agents."