Vitalik Buterin urged new AI labs to prioritize enhancing humanity as their core mission and avoid developing systems with autonomous decision-making times exceeding one minute.
The founder of Ethereum warned that the trend of pursuing hyper-autonomous AI is spreading, while projects that augment human AI are scarce; he also suggested opening up the source code as much as possible.
- The priority is AI to enhance humanity, not complete AI self-sufficiency.
- Avoid systems that take more than one minute to make a decision on their own.
- We recommend that the project be as open-source as possible.
Buterin wants to limit the AI's autonomy over time.
Vitalik Buterin says new AI labs should avoid building systems with autonomous decision-making times exceeding one minute.
He emphasized that the fundamental mission of an AI lab should be enhancing humanity, focusing on strengthening human capabilities rather than empowering systems to operate autonomously in the long term.
The 1-minute threshold is stated as a design constraint to limit the degree of autonomy, thereby reducing the risk of AI making lengthy decisions without direct human intervention or supervision.
Concerns about the ASI race and calls for open-source
Buterin argues that companies pursuing fully autonomous ASI (super AI) are becoming common, while AI that enhances human capabilities remains rare.
He argued that even if all controversies surrounding AI safety were proven unfounded, the pursuit of fully autonomous systems would still be rampant; conversely, the direction of augmented AI, such as building a mech for the human brain, would be lacking.
To steer the ecosystem toward transparency and verifiability, Buterin urged projects of this type to be as open-source as possible, allowing the community to test, reuse, and assess risks over time.
