As an engineer who has worked with automation, robotics, and expert systems since the mid-1960s—and later with artificial intelligence—I still have important questions about the future development of AI and the data centers that support it.
The technology industry changes rapidly. Where are companies such as Sun Microsystems, UNIVAC, and Digital Equipment Corporation today? Once dominant, they have disappeared or been absorbed by other companies. We should also remember the enormous disruption caused by the dot-com crash.
The same fate could await some of today’s AI leaders. Most current generative-AI systems rely heavily on large language models, which are primarily trained to process and generate text. Yann LeCun, one of the pioneers of modern AI, a former chief AI scientist at Meta, and a recipient of the Turing Award—often described as the Nobel Prize of computing—has advocated a different direction: AI systems that learn from and understand the physical world rather than relying primarily on language.
The future may also involve smaller, specialized AI systems designed for specific fields, such as engineering, medicine, and finance. These systems could require smaller data-center footprints and may be easier to monitor, control, and secure. If so, the future of AI could look very different from the massive, general-purpose systems being developed today.
