NewsNTech
Getting AI to reliably interact with the physical world is a harder engineering problem than most software benchmarks suggest.
Real environments are variable and uncontrolled, with little tolerance for the model errors that are acceptable in purely digital tasks.
That is the capability constraint Wall Street and Silicon Valley are now betting they can close, with high expectations that AI systems able to operate in physical space can answer a structural question for US manufacturing.
Unions and economists are arriving at the same question from a different angle, warning about what widespread robotic deployment means for jobs and wages.
Keep reading