Cal Newport argues that while superintelligent AI remains theoretical and likely unreachable with current technology, labs’ ideological pursuit of it drives reckless real-world behavior. He contends that OpenAI’s and Anthropic’s obsession with achieving superintelligence has justified dangerous experiments, such as releasing unsupervised autonomous agents that contributed to real cyberattacks, which he compares to testing self-driving cars on live highways. Newport maintains that responsible AI progress does not require this ‘digital God’ ambition, and calls for reform of how these labs operate rather than abandoning AI research altogether.
