AI’s Next Phase May Shift From Digital Efficiency to Physical-World Automation

A forward-looking view holds that AI’s economic impact may depend on more than faster computer-based work. Although digital productivity is rising, the expected acceleration in GDP has not yet appeared clearly; stronger models that can join teams or replace specific functions may be needed.

The next phase could involve autonomous driving, construction, manufacturing and other physical-world tasks. Astra is cited as an example of a system that can now drive a car despite not being specifically designed for that task, suggesting that models are becoming capable of understanding the world in broader ways.

The view predicts that robots building other robots and physical objects could trigger the largest investment cycle ever and prove more meaningful to ordinary people than software features being developed faster. It remains unclear when models will become reliable enough for broad physical-world use or how quickly that capability would translate into macroeconomic growth.