Main points
- Google has created a special team to catch up with Anthropic and their Claude tool in the field of AI coding.
- Co-founder Sergey Brin and Google DeepMind's CTO are personally involved in the project to strengthen Gemini's capabilities and reduce the backlog.
Google admits Gemini lags behind Claude in programming / Unsplash / Solen Feyissa
Google is concerned about the lag in its AI tools in the programming space. According to a leak, the company has created a special team to catch up with competitor Claude, which is showing stronger results in working with code.
Google is ramping up work on its AI agents after internal assessments showed they were lagging behind a competitor. According to a leak, the company's co-founder Sergey Brin has urged the development team to urgently close the capability gap. This is reported by Techradar .
Why is Google rushing to catch up with competitors in AI coding?
This primarily concerns the area of so-called ” agentic execution ” – the ability of AI not just to write code, but to perform complex tasks, acting as a full-fledged developer. In an internal memo, Brin emphasized that the company must turn its models into “core developers” of finished code.
The main competitor in this direction is Anthropic with its Claude tool, in particular the Claude Code solution, which is actively used for programming and research. It was its progress, according to sources, that became the impetus for the creation of a special ” strike team ” at Google.
It's not just the technical leadership that's involved. Brin, along with Google DeepMind's CTO Koray Cavukcuoglu, is reportedly personally involved in the project . This underscores how seriously the company takes the threat of losing ground.
The main goal is to enhance Gemini's capabilities in working with code and research tasks. In the future, this could lead to the creation of systems capable of independently improving their own algorithms.
Google is also considering a scenario where such AI agents could automate a significant portion of the work of engineers and researchers. Similar approaches are already being used by OpenAI, which is creating internal tools to optimize the work of its teams.
At the same time, Gemini development may remain an internal project . Because the model will be trained on Google code, its use may be limited to company employees. This means that end users may not see the new features immediately.
However, even in this case, these tools can indirectly affect the market . Improving internal systems can accelerate the development of new AI models, which will later become available to a wider audience.
According to the Times of India , the situation shows that competition in the field of AI coding is only intensifying . Companies are trying to create tools that do not just help programmers, but can do their work.
The main question now is whether Google can quickly close the gap with Claude and regain its lead, or whether this gap will continue to grow.