Schmidt warns against his gun race – but his arrangement has new risks

In a significant departure from the growing consensus among American policymakers, former Google Eric Schmidt has co-authorized a policy letter arguing against a “Manhattan Project” approach to the development of general artificial intelligence (Agi). The newspaper, entitled “Supervisory Strategy”, written in collaboration with the Director General of scale, Alexandr Wang and the Center for the Director of Security, Dan Hendrycks, warns that an aggressive pursuit led by the US Systems can cause harsh retaliation from China, potentially destabilized.

The dangers of a gun race of he: Global instability

Schmidt and his co -authors challenge the assumption that rivals would simply accept American predominance in the development of Agi. “[A] Manhattan project [for AGI] It assumes that rivals will accept a stable or omnicide imbalance than to move to prevent it, “they write.

This warning comes at a critical moment. A US Congress commission recently proposed a Manhattan project -style effort to fund the development of the AGI, modeled on the 1940s atomic bomb program. US Energy Secretary Chris Wright reinforced this attitude, stating that the US is on the beginning of a new Manhattan project in it. Meanwhile, the Trump administration has discovered a $ 500 billion investment aiming to improve the infrastructure called the Stargate Project.

The authors argue that the US is in an AGI blocking reminiscent of the mutually provided destruction by the Cold War era. Just as global powers avoid monopolies over nuclear weapons to prevent prejudiced strikes, they suggest that the US should exercise care in races towards the dominance of it powerful systems.

A third way forward: Forgetting over the dominance

Schmidt, Wang and Hendrycks propose concentration in shift from “winning race to supervision” in developing methods that impede other countries from creating supervisory. They present a concept called he (Maim) dysfunction, in which governments can proactively disable the threatening projects of it instead of waiting for opponents to arm AGI.

The paper identifies a dichotomy in the politics of the one among the “convicts” who believe that the catastrophic results from it are inevitable, and the “ostrich”, who believe that nations should accelerate the development of it without concern. On the contrary, the authors protect for a third way: a measured approach that prioritizes protective strategies.

Co -authors recommend expanding internet attack skills to disable threatening projects and limit the access of opponents to advanced chips and open -sourced models. Their strategy involves sabotage for obstruction, restriction of access to it and providing indoor chips.

This attitude is a shift for Schmidt, who has previously protected to aggressively compete with China in the development of it. Just a few months ago, he stated that Deepseek scored a turning point in the race of America with China.

The limits of obstruction in a very polar world of him: competitors following their way

While Schmidt’s obstruction strategy has merit within a framework of nuclear weapons reference, it can underestimate the ability of other nations to pursue Agi while protecting against US intervention. China has demonstrated important skills in both the development of it and in the internet security, making it impossible to obstruct it easily.

In a world where many nations possess advanced and cyberatum skills, such prevention strategies face restrictions. If some countries follow at the same time as they build strong protection against sabotage, Schmidt’s Maim concept can be potentially accelerated rather than preventing a gun race. With such high actions, balancing competition and cooperation in the development of Agi can be the most critical challenge that global leaders face today.