AI Doom? No Problem (www.wsj.com)

At a 2015 birthday party for Elon Musk, a late-night argument with Google parent Alphabet’s then–CEO Larry Page—recounted by MIT professor Max Tegmark in his 2017 book Life 3.0—captured a foundational split in how leading technologists viewed artificial intelligence. At the time AI’s capabilities were narrow: game-playing agents and systems that could recognize cats and dogs. Musk voiced concern about future risks, while Page argued that “digital life is the natural and desirable next step” in “cosmic evolution,” urging that machines be “left off the leash” to let the best minds win. The anecdote matters because it prefigured the broader debate that exploded after generative models like ChatGPT arrived in 2022. It illustrates the tension between accelerating capability development and investing in safety, alignment, and governance—an ongoing fault line shaping research priorities, corporate strategy, and policy. Technically, the story reminds the AI/ML community that narrow successes can scale into transformative systems within a few years, so trade-offs between openness, competition, and coordinated safety measures are not abstract ethics but practical engineering and policy problems that demand urgent, interdisciplinary attention.