Software engineer Steve Yegge recently published a lengthy post highlighting the slow pace of Google’s AI adoption, based on a conversation with his tech director friend. In a recent follow-up, Yegge noted that his earlier post drew significant criticism. He also pointed out that multiple organizations and Google employees have since reached out to him, corroborating his initial claims.
Based on the experiences shared by these engineers, he added in his latest post: “I haven't verified each person's story, but the picture these Googlers paint is consistent across sources. It is more specific than what I originally wrote, and somewhat bleaker.”
My tweet last week about Google's AI adoption drew a lot of pushback, to say the least.
Since then, Googlers from multiple orgs have reached out to me independently and anonymously. They've expressed fear of being doxxed, concern about what they saw as bullying of me, and…
— Steve Yegge (@Steve_Yegge) April 20, 2026
He goes on to describe the presence of a two-tier system, where DeepMind engineers use Claude as a daily tool, while the rest of Google does not. When the question of equalizing access was raised internally, DeepMind employees reportedly objected strongly, with several engineers even threatening to leave.
Yegge further adds that Googlers say leadership is aware of this gap. The response has been to mandate AI usage in OKRs and individual expectations, along with introducing an internal token-usage leaderboard. However, managers have reportedly been told both that the leaderboard will not be used for performance reviews and, at the same time, that it absolutely will. He also mentions that Google’s culture is not yet properly adapted for high-volume, AI-assisted coding.
He later clarified that the goal of his original post was simply to encourage more people to adopt AI and agentic coding. He added, “Nobody is as far ahead as they might look from the outside, and none of you are as far behind as you might be worried you are.”
Addy Osmani’s reply
In his earlier post, Yegge challenged Google to prove that half of its engineers were not consuming at least 4 million tokens a day. Addy Osmani, Google’s Engineering Director, pushed back with more concrete details. He stated that over 40,000 Google engineers use agentic coding tools on a weekly basis.
He further explained that engineers have access to the company’s own CLI tools, custom models, MCP integration, and even virtual SWE team setups. He also noted that Googlers actively use external tools, including Claude Code through Anthropic, as well as models on Vertex, to benchmark their work.
What has led to the slowdown in AI adoption?
In his previous post, Yegge pointed to an industry-wide hiring freeze lasting over 18 months. According to his friend, a tech director at Google, this has prevented the company from gaining an accurate sense of how far behind it may be in AI adoption.
He stated: “There has been an industry-wide hiring freeze for 18+ months, during which time nobody has been moving jobs. So there are no clued-in people coming in from the outside to tell Google how far behind they are, how utterly mediocre they have become as an engineering organization.”
He also argued that Gemini has not been good enough to streamline user workflows in the way Claude can, and that this has hindered the adoption of agentic coding within Google.
Reinforcing this concern in his latest post, Yegge added that the clearest feedback he is hearing is that Googlers do want high-quality agentic tools and are repeatedly asking for better ones. However, he concludes that overall, this does not reflect an engineering organization that is functioning optimally.
