Editorial illustration for Google leaders, Hassabis, refute uneven AI adoption, cite 40K SWEs agentic coding
Google's AI Coding Revolution: 40K Engineers Adapt Now
Google leaders, Hassabis, refute uneven AI adoption, cite 40K SWEs agentic coding
Inside Google, the narrative of a company slow to embrace its own AI revolution is being dismantled, not by corporate spin, but by numbers and internal testimony. Addy Osmani, a director at Google Cloud AI, fired back with a blunt counter: over 40,000 software engineers use agentic coding every week. That’s not a pilot program.
That’s a cultural baseline. The tools are custom models, command-line interfaces, and MCPs. And if anyone doubted access to outside innovation, Osmani notes Anthropic’s models are available on Vertex.
Jaana Dogan, a Google engineer, describes her colleagues using AI “every second of the day.” Paige Bailey at DeepMind says their agents run 24/7. Demis Hassabis and other leaders are now pushing back on claims of uneven adoption, not with vague promises, but with hard evidence of a workforce already wired for agentic coding. The picture emerging is less about lagging and more about a quiet, pervasive transformation.
Addy Osmani, a director at Google Cloud AI, wrote that Yegge's account "doesn't match the state of agentic coding at our company." He added, "Over 40K SWEs use agentic coding weekly here." Osmani said Googlers have access to internal tools and systems including "custom models, skills, CLIs and MCPs," and pushed back on the idea that Google employees are sealed off from outside models, writing that "folks can even use @AnthropicAI's models on Vertex" and concluding that "Google is anything but average." Other current Google employees reinforced that message. Jaana Dogan, a software engineer at Google, wrote in a quote tweet: "Everyone I work with uses @antigravity like every second of the day," later following up with another X post stating: "Unpopular opinion: If you think tokens burned is a productivity metric, no one should take you seriously. Imagine you are a top 0.0001% writer and they are only counting the tokens you produce." Paige Bailey, a DevX engineering lead at Google DeepMind, said teams had agents "running 24/7." Several other Google and DeepMind figures also challenged Yegge's characterization, some disputing the factual basis of his claims and others suggesting he lacked visibility into current internal usage.
Demis Hassabis didn’t need to shout. The numbers spoke first. Forty thousand engineers using agentic coding every week, that’s not an outlier experiment; it’s infrastructure woven into the daily rhythm of a company that builds the tools it uses, and uses the tools it builds.
Osmani’s rebuttal wasn’t a defense. It was a reminder: inside Google, custom models, CLIs, MCPs, even Anthropic’s models on Vertex, this is not a walled garden. It’s a workshop with the doors wide open.
Dogan’s quip about tokens burning as a metric cuts deeper than any spreadsheet. Productivity isn’t volume. It’s velocity with purpose.
Bailey’s teams running agents “24/7” underscores a simple truth: adoption isn’t uniform because it doesn’t need to be. It’s organic, relentless, and often invisible to outsiders who mistake noise for signal. Yegge’s picture was always a caricature.
The reality is messier, faster, and far more interesting. Google isn’t debating whether to adopt agentic AI. It’s already rewriting the playbook, one agent, one engineer, one 40,000-strong wave at a time.
Related Reading
Browse all Google news →
Market Trends AI drives three initiatives to accelerate global action for planet protection
Market Trends Akamai data shows AI training bots and content‑fetching bots rise since July
Market Trends NVIDIA and Google Cloud let developers scale AI from prototype to production
Related Topic Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI
Related Topic Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks
From Archives OpenAI acquires TBPN to accelerate global AI conversation, memo says
From Archives AI splits art schools as program links students with Adobe, Google
From Archives Developer says he reverse‑engineered Google DeepMind’s SynthID watermark tool
Open Source
From Archives Google AI's Vantage protocol shows Executive LLM beats agents on 8 metrics
LLMs & Generative AI
Common Questions Answered
How many software engineers at Google are using agentic coding tools weekly?
According to Addy Osmani, director of Google Cloud AI, over 40,000 software engineers are using agentic coding tools on a weekly basis. This number directly challenges claims of uneven AI tool adoption within the company.
What internal tools and systems do Google engineers have access to for agentic coding?
Google engineers can access custom models, skills, command-line interfaces (CLIs), and multi-cloud platforms (MCPs) for agentic coding. Additionally, they can even use Anthropic's models on Google's Vertex platform, demonstrating a broad range of AI coding tools available internally.
How did Google leadership respond to claims of uneven AI tool adoption?
Google's senior AI leaders, including Demis Hassabis and Addy Osmani, quickly refuted the narrative of patchy AI tool rollout. Osmani specifically countered the claims by providing concrete evidence of widespread agentic coding tool usage across the company's engineering teams.
Further Reading
- Google Cloud Study Reveals 52% of Executives Say Their Organizations Have Deployed AI Agents, Unlocking a New Wave of Business Value — Google Cloud Press Corner
- The Agentic AI Adoption Paradox: Why 40% Penetration Doesn't Mean 40% Success — Towards AI
- Somebody Is Lying About Agentic AI Adoption—and You're Paying ... — Cloud Computing Insider
- A Year Just Happened in a Week - by Keith Teare — That Was The Week
