The ban didn't come from any considered evaluation of Gemini's capabilities or risks. Instead, it stemmed from an outdated internal policy — a relic from an earlier era of the company's tool-approval process . Brin described the restriction as having "weird historical reasons," explaining that the policy was old, barely enforced, and had never been updated to account for Gemini being a first-party product . It had simply been left on the prohibited list by administrative inertia.
Despite being a co-founder with supervoting shares and direct access to CEO Sundar Pichai, Brin could not quickly resolve the issue . He spent weeks trying to get the restriction removed through normal internal channels, but the process dragged on for what he called "a shockingly long period of time" .
Frustrated, Brin finally escalated the matter directly to Sundar Pichai, telling him: "I can't deal with these people, you need to deal with this" . Pichai proved supportive and helped lift the restriction, allowing Google engineers to finally use Gemini for coding .
This episode is a revealing signal about Google's bureaucracy:
The "no list" story isn't the only sign that Google's relationship with AI is complicated. On July 23–24, 2026, Alphabet shares plunged roughly 7% after the company raised its full-year capital expenditure forecast to as high as $205 billion, alarming investors about AI spending discipline . The selloff wiped out approximately $290–$294 billion in Alphabet's market capitalization in a single day — one of the largest single-day value destructions on record .
By July 25, 2026, Alphabet's stock had partially rebounded, lifting Brin's net worth to approximately $241.8 billion and reclaiming his position as the world's third-richest person .
The broader context ties together the same paradox: Google is investing tens of billions into AI (capex of $195–$205 billion in 2026 alone), yet investors fear the spending is out of control . Meanwhile, internally, the company's own bureaucracy was blocking engineers from using its premier AI model — a disconnect that Brin's anecdote brought into sharp focus.