Google co-founder Sergey Brin is personally stepping in to form an internal “strike team” to address the serious challenge posed by AI startup Anthropic in the realm of programming capabilities. According to a report by The Information, a “strike team” composed of researchers and engineers has been quietly assembled within Google’s DeepMind, with the core objective of rapidly narrowing the gap with Anthropic’s Claude models in code generation and agent execution capabilities.<\/p>\n
In an internal memo to the entire DeepMind team, Brin used stern language, demanding the team “must urgently fill the gap in agent execution to win the final sprint, making our models the primary developers writing code.” This marks the significant pressure Google is facing from Anthropic in the AI race, particularly in the programming field, which is seen as a critical path toward Artificial General Intelligence (AGI).<\/p>\n
Brin’s Rare Personal Campaign Focuses on Internal Code\u653b\u575a<\/p>\n
The configuration of this “strike team” personally overseen by Brin is extremely rare within Google. The team is led by DeepMind research engineer Sebastian Borgeaud, who previously headed the pre-training work for the Gemini models. More crucially, both Brin and DeepMind Chief Technology Officer Koray Kavukcuoglu are personally involved in the team’s work.<\/p>\n
Unlike previous efforts primarily focused on developing models for external clients, this team’s strategy has undergone a pivotal shift: prioritizing the training of models capable of writing Google’s internal, proprietary code. Google possesses one of the world’s largest internal codebases, containing over 20 billion lines of code. Enabling AI models to deeply understand and generate this code, which is tightly coupled with Google’s vast engineering systems, is considered the “internal skill” essential for enhancing their general programming ability. While models trained on internal code cannot be released directly to the public, the derivative models they produce will be publicly available, fundamentally boosting competitiveness.<\/p>\n
A Google spokesperson stated that the “adoption rate of internal programming tools is incredibly high,” and that using these tools at scale “puts a powerful engine behind model and AI tool development.”<\/p>\n
A Stark Gap: 100% vs. 50% Code Generation Rate<\/p>\n
The direct reason for Brin’s high-profile intervention is the vast chasm between Google and Anthropic in the application level of AI programming. A set of comparative figures reveals the severity of the issue:<\/p>\n
CompanyAI Code Generation RateKey Data PointAnthropicClose to 100%Stated publicly by Claude Code lead Boris Cherny in January 2026.GoogleApproximately 50%Revealed by Chief Financial Officer Anat Ashkenazi during the February 2026 earnings call.<\/p>\n
Note: Data sourced from public statements by company executives and earnings calls.<\/p>\n
This is not merely a numerical gap; it signifies that Anthropic has already established a closed-loop, AI-centric development workflow internally. Friction has even emerged within Google: reportedly, Gemini was once placed on a\u7981\u7528 list for internal programming tools, an event that left Brin “stunned” and ultimately required CEO Sundar Pichai’s intervention to resolve. This internal resistance partly explains Google’s lag in this race.<\/p>\n
Claude Code’s “Explosive” Commercial Performance Establishes Market Dominance<\/p>\n
Anthropic’s Claude Code not only leads in internal application but its commercial performance is causing anxiety among competitors. The product was officially opened to the public in May 2025. By February 2026, its annualized revenue run rate had surpassed $2.5 billion (approximately $367 million), achieving a leap from zero to tens of billions in revenue within just nine months.<\/p>\n
What alarms Google and OpenAI even more is its market penetration. Reports indicate Claude Code has captured 54% of the enterprise programming market, while OpenAI holds 21%, and Google is not even listed separately. On the developer platform GitHub, 4% of public code commits are already generated by Claude Code, with projections expecting this figure to exceed 20% by the end of 2026.<\/p>\n
Habit formation presents an even more formidable barrier. A Google Developer Expert (GDE) admitted in an open letter that being forced to use Claude Code for 8 hours daily at work has caused these usage habits to “seep into personal projects.” Even though he personally prefers certain features of Gemini, he finds it hard to resist the power of “muscle memory.”<\/p>\n
Industry Consensus: Programming Capability Becomes Core AGI Battleground<\/p>\n
Google’s anxiety is not an isolated case; the entire industry is elevating programming capability to the highest priority. Almost simultaneously, OpenAI announced the shutdown of its capital-intensive but user-growth-lagging Sora video generation project, reallocating all computing power and engineering resources to coding and enterprise products. Foreign media commentary suggests Anthropic’s Claude Code is “eating OpenAI’s lunch.”<\/p>\n
The concurrent actions of the three AI giants stem from a shared strategic recognition: powerful code generation capability is a critical pathway toward “AI takeoff” or achieving AI self-evolution. Brin positioned this as an essential route in his memo. A sufficiently powerful code agent, combined with AI that can solve mathematical problems and run experiments, could theoretically automate the work of AI researchers and engineers themselves on a large scale, thereby enabling AI self-iteration and improvement.<\/p>\n
From “Mobilization” to “Orders,” Google Applies Full Internal Pressure<\/p>\n
To close the gap, Google has sounded a full “red alert” at the internal implementation level. Brin explicitly mandated in his memo that every Gemini engineer must use internal agent tools when handling complex, multi-step tasks\u2014this is “mandatory,” not a suggestion. Some teams have incorporated AI tool usage into performance evaluations.<\/p>\n
Reportedly, Pichai is also applying pressure to non-technical staff, informing them that AI usage “is no longer encouraged, but expected.” Google’s internal programming tool Jetski has established leaderboards, similar to Meta’s approach, to track and incentivize usage across teams. Another asynchronous AI agent, “Agent Smith,” became so popular that Google once had to restrict access.<\/p>\n
The formation of this strike team marks Brin’s most high-profile personal intervention in the AGI race since his return to the company in 2023 to help address the ChatGPT challenge. In February 2025, he sent a memo to the team stating “competition has accelerated dramatically, and the final race to AGI is on,” suggesting engineers work 60-hour weeks. The tone then was one of mobilization; 14 months later, it has shifted to direct orders. During these 14 months, Anthropic’s annualized revenue surged from $1 billion to $14 billion, and Claude Code went from non-existent to capturing half the market.<\/p>\n
Context and Impact: The Cycle and Escalation of Competition in the AI Industry Food Chain<\/p>\n
This event is a microcosm of the evolving competitive landscape in the AI industry. When ChatGPT emerged in late 2022, Google sounded its first “Code Red,” with co-founders Larry Page and Brin returning to oversee. Three and a half years later, a new alarm has been sounded, but the opponent has shifted from OpenAI to Anthropic\u2014a company led by former OpenAI research vice president Dario Amodei and founded by a group of core researchers.<\/p>\n
History appears to have completed a cycle: people from Google founded OpenAI, people from OpenAI founded Anthropic, and now Anthropic is, in turn, pressuring Google in a key domain. This highlights the rapid flow of AI talent and technology, as well as the potential for startups to achieve overtaking in niche areas through focused strategies.<\/p>\n
For the market, the white-hot competition among giants will accelerate the evolution and adoption of AI programming tools, ultimately benefiting a wide range of developers and enterprise users. However, it also means resources will be highly concentrated in areas most likely to generate direct commercial value and productivity, potentially squeezing other frontier AI exploration projects with longer commercialization paths (like Sora). Investors need to closely monitor the actual progress companies make in building core capabilities, rather than focusing solely on model parameter comparisons. Whoever can first create an AI agent truly capable of serving as a “core developer” may gain a decisive advantage in the next phase of competition.<\/p>\n","protected":false},"excerpt":{"rendered":"Google co-founder Sergey Brin is personally stepping in to form an internal “strike team” to address the serious…\n","protected":false},"author":2,"featured_media":213617,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[165],"tags":[13994,95593,587,3424,2370,5040,812,613,614],"class_list":["post-213616","post","type-post","status-publish","format-standard","has-post-thumbnail","category-sergey-brin","tag-agent","tag-ai-programming","tag-anthropic","tag-claude-code","tag-deepmind","tag-gemini","tag-google","tag-openai","tag-sergey-brin"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@people\/117208638198187446","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/posts\/213616","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/comments?post=213616"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/posts\/213616\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/media\/213617"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/media?parent=213616"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/categories?post=213616"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/people\/wp-json\/wp\/v2\/tags?post=213616"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}