A quiet internal exercise at Block Inc. in December 2025, led to a decision that would soon draw global attention. The company, based in San Francisco, reduced its workforce by about 40 percent after its leadership concluded that advances in artificial intelligence had changed how much human labor was needed to operate the business.
Speaking on a podcast hosted by Sequoia Capital, Chief Executive Officer Jack Dorsey explained that the decision followed a hands-on review of new AI tools during the holiday period.
“We just did this exercise of, ‘Okay, so what is the minimal number of people that we would need to keep the service up 100%?’” Dorsey said.
He added that the company also examined regulatory requirements and growth commitments before asking a final question:
“What is the minimal set of folks that we need in order to grow… but also rebuild the company as an intelligence?”
According to Dorsey, when leaders considered what the company would look like if it were built with the latest AI systems from the start, they arrived at a significantly smaller workforce. More than 4,000 roles were affected out of roughly 10,000 employees.
He said the decision was made quickly after the review.
“I want to be ahead of it, because then we can do it with a lot more integrity,” he stated.
Companies are no longer treating artificial intelligence as a supporting tool but as a central part of how work is organised.
In practical terms, this means:
- fewer employees handling routine digital tasks
- smaller teams supported by AI systems
- reduced need for traditional management layers
While some firms have avoided linking layoffs directly to AI, others have begun to acknowledge its role more openly.
Image Credit: Alex Karp (WEF)
Similarly, comments from Alex Karp, head of Palantir Technologies, have added to public discussion about the future of work.
Karp argued that two groups of people are likely to remain in strong demand as AI systems expand: workers with practical trade skills and individuals who approach problems in unconventional ways.
He pointed to occupations such as electrical work and construction as examples of roles that are difficult to automate due to their physical and situational nature. He also suggested that people who think differently and challenge standard approaches may be better positioned to build and guide AI systems.
Currently, entry-level positions are becoming less common and expectations are changing.
A report presented to the United States Senate has estimated that up to 100 million jobs could be disrupted by automation over the next decade, though the pace and scale remain uncertain.
