Skip to content
context Business and investment

Head of The Browser Company: AI agents are technology, not a product people want

only one source so far

Josh Miller, head of The Browser Company, claims that hardly anyone outside the tech community uses AI agents. OpenAI reports about 10 million weekly users of Codex and ChatGPT Work, and similar figures reportedly apply to Claude Code and Cowork from Anthropic – compared with a billion chatbot users.

Josh Miller, CEO of The Browser Company, sparked a discussion on X by claiming that, despite all the hype around AI agents, it is “insane that essentially nobody uses them”. According to Miller, the technology is theoretically ready to transform how people work and live, but the general public is not interested. In an interview, he added that outside the technology community, he had not heard a single person talk about an agent they actually used. Miller was behind the Arc browser, which his company The Browser Company sold to Atlassian last year for 610 million dollars, and previously served as the first director of product at the White House under President Obama.

OpenAI said last month that its tools Codex and ChatGPT Work have about 10 million weekly users combined. According to people close to Anthropic, its tools Claude Code and Cowork have reached similar levels of adoption. By comparison, the chatbots ChatGPT and Gemini have around a billion monthly active users on average – so, according to the article, agents are “essentially a rounding error” compared with them.

The main claim Miller makes is that AI agents are not yet a standalone product, but rather an artificially created term that the industry uses as a collective label for the technology behind actual products. As an example, he cites the most popular feature of the Dia browser from The Browser Company – a morning briefing compiled from the user's calendar and email – which is technically enabled by an agent, although, according to Miller, the user does not need to know that at all. He sees groupthink in the industry as the underlying problem: in his words, almost all leading AI labs described their vision by referring to the film Her (Her) by Spike Jonze, which he believes points to a lack of diversity in approaches.

The remaining part of the article was not available. You can find details in the source article.

What changed

Why it matters

The criticism highlights a gap between how much labs such as OpenAI and Anthropic invest in agentic tools and how little that investment has so far translated into usage outside the technology community – agent adoption figures are orders of magnitude lower than those for conventional chatbots. For companies developing AI products, this signals that showcasing technical capabilities (harness, tool calls) is not enough on its own and that they need to be packaged into a simple, easy-to-understand product focused on a specific user need.

Relevant practical impact

What this means

01

For a business

Companies developing agentic products risk their investment in agents' technical capabilities failing to translate into mass adoption unless they offer them as an easy-to-understand consumer product focused on a specific need – as illustrated by the morning briefing in the Dia browser, where the user is completely unaware of the agent running in the background.

Development
What to decide When designing agentic products, verify that the final offering addresses a specific user need rather than merely showcasing the agent's technical capabilities.
More business impacts →
AI agents Anthropic ChatGPT Claude Code Codex OpenAI

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
Wired — AI section independent context · first detected Why Normal People Aren’t Using AI Agents