Google CEO Admits Delays in Coding - Challenges in AI Race and Search Overhaul

Google CEO Sundar Pichai openly admitted that the company’s AI model, Gemini, is lagging in the coding domain. In a recent interview on The New York Times’ technology podcast, Pichai candidly discussed Google’s current position amid rapid changes in the AI industry, despite the announcement of new AI products at Google I/O. This revelation highlights the dilemma Google faces in balancing a 25-year search service overhaul with a transition to AI.

Acknowledging Delays in the Coding Domain Pichai stated that while Google is strong in areas such as text, multimodal capabilities, voice, and reasoning, the company is “admittedly lagging somewhat” in agent-based programming that involves tool integrations or in handling long-term tasks requiring extended steps. He pointed out that there is significant room for improvement, particularly in scenarios where experienced developers work with complex codebases. Pichai attributed these delays to the rapid pace of changes in the AI industry, where transformations that previously took five years are now happening within 30 to 60 days. He explained that even for a large company like Google, it is not always possible to synchronize perfectly with the training cycles of research institutions, which can sometimes result in temporary setbacks.

Strategic Positioning of New AI Products at Google I/O At the recent Google I/O, the company unveiled a series of new AI products, including Gemini 3.5 Flash, Gemini Omni, and Gemini Spark. Pichai described Gemini 3.5 Flash as a significant step forward in addressing some of the company’s previous shortcomings. However, he acknowledged that in the coding domain, real-world usage data is critical. He noted the lack of direct product entry points, such as those provided by Claude Code, or high-frequency usage scenarios, as seen with Anthropic’s Cursor. Pichai also mentioned internal efforts to address these gaps, citing the implementation of Antigravity 2.0. He revealed a surprising surge in internal token usage, doubling every week, as employees increasingly use the models for their work. This underscores Google’s commitment to addressing its shortcomings in the coding domain.

The 25-Year Overhaul of Search Services and AI Transition Challenges The interview also touched on Google’s significant update to its search bar and entry points—the first major redesign in 25 years. Many anticipate that the classic web search interface will eventually be replaced with AI-based modes as the default entry point. However, Pichai emphasized the importance of “moving users forward with the product while ensuring it meets their expectations” and expressed caution about a rapid transition. He noted that users value the speed of search and its role as a gateway to existing information and content on the internet. Therefore, sources and links will remain integral components of the search experience. While the transition to AI modes has become smoother, Pichai described it as a continuous process of evolution. On the commercial front, Pichai acknowledged the profitability of the traditional search advertising business but highlighted the significantly greater value that AI and agent-based modes can offer users compared to a decade ago. He expressed confidence in the continued viability of a business model combining subscriptions and advertising.

Rapid Advances in AI and Industry Dynamics Pichai remarked on the extraordinary pace of change in the AI industry, suggesting that Artificial General Intelligence (AGI) might be closer than previously anticipated. This rapid progress means that AI has the potential to reshape jobs, income, and future lifestyles. While asserting that Google remains the only major company at the forefront, Pichai acknowledged the presence of several startups advancing at a very fast pace. Despite its long-standing investment in the field, Google has struggled to fully keep up in certain areas due to the speed of industry changes.

Public Fears About AI and the Future of Business Models Toward the end of the interview, Pichai addressed public fears surrounding AI. He acknowledged that concerns about AI are not unfounded, as the technology could indeed transform jobs and incomes. However, he stressed that the ultimate commercial value of AI depends on the overall value it provides to users. Over time, he expects the value delivered to users to grow, competition to intensify, and choices to expand. He remains optimistic that a business model combining subscriptions and advertising will continue to thrive in this new world.

Conclusion: Google’s Leadership and Future Prospects Pichai’s candid remarks shed light on the challenges and opportunities Google faces in the competitive AI landscape. While admitting to delays in the coding domain, Google is actively working to address these issues with new products like Gemini 3.5 Flash and initiatives such as Antigravity 2.0. In the search service overhaul, the company aims for a cautious, phased transition, balancing user expectations with the potential of AI. As the AI industry evolves at an unprecedented pace, questions remain about how major companies like Google can sustain innovation and maintain user trust. As Pichai noted, in an era where 30 to 60 days of change equate to five years of progress, Google’s strategic navigation will remain in the spotlight.