- Google Reportedly Creating AI ‘Strike Team’ to Challenge Claude Code
- The company wants to create artificial intelligence agents that can help in coding and investigation.
- However, the result may never be made public.
One of the biggest ways AI is changing the world is through the rise of AI agents that can write code. Instead of spending hours creating and debugging code, programmers can now simply ask an AI model to get to work for them.
Google’s Gemini AI is quite adept at this, but that’s not good enough for the company, which has reportedly been so shaken by the progress of its rival Anthropic and its agent Claude Code AI that Google has formed a “strike team” to address the problem.
In fact, according to a report by The Information, Google co-founder Sergey Brin warned that Google cannot afford to waste time. In a memo to the company’s AI engineers and researchers, Brin said: “To win the final sprint, we must urgently close the gap in agent execution and make our models primary developers” of the final code.
Article continues below.
And this isn’t just a hands-off effort by Google’s executive team. Both Brin and Google DeepMind CTO Koray Kavukcuoglu have been “directly involved with the attack team, showing their importance to Google’s top leadership,” The Information states. This aptly illustrates how seriously the company appears to be taking the Anthropic threat.
Will it ever see the light of day?
What is Google aiming for here? Well, the most obvious answer would be to maintain a competitive advantage against a popular rival in an emerging field that can turn dominant companies into a pile of cash.
But there is more than that. As The Information noted, “The ultimate goal is for Al to take off, or for Al to improve itself, which has been a particular focus for Google co-founder Sergey Brin.”
In fact, the report puts it this way: “An advanced encryption agent, combined with Al, that can solve mathematical problems and perform experiments, could one day automate the work of Al researchers and engineers on a large scale.” That’s something another rival, OpenAI, has already developed for its own researchers working on internal AI projects, and the company is also building a GitHub rival to help streamline the work of its employees.
The priority appears to be improving Gemini’s ability to work with Google’s own code. That would improve its performance when used by Google engineers and researchers for their internal projects, putting less emphasis on coding tools for everyday coders working outside the organization. Since an improved AI agent will have been trained on Google code, it could not be published.
So before you get your hopes up, you may not see the work of the Gemini strike team, at least not initially. But as this report notes, improved internal tools “could, in theory, help create better models that could then be published.”
For now, all we can do is wait and see if Google is able to close the gap between its own AI agents and Anthropic’s Claude before the chasm becomes insurmountable.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
The best laptops for all budgets
