TECH Signal 409
AI Tools Increase Workload and Hours at AI Companies, Contradicting Promises of a Shorter Work Week
↗
AI companies like OpenAI, Anthropic, and Meta are running employees on 70-to-90-hour weeks while a UC Berkeley study shows AI tools expand workload by requiring constant output verification, undermining earlier claims that AI would enable a four-day work week.
For engineers, adopting AI tooling is currently adding verification overhead and broadening task scope rather than reducing hours. The companies building these tools are operating some of the most intense work cultures in tech, which signals that near-term productivity gains from AI are being captured as more output, not less work.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
A UC Berkeley study following hundreds of workers at a US tech company for eight months found AI made employees work faster, take on broader tasks, and extend work into more hours, partly because of the need to constantly check AI output.
Former OpenAI employees report working at least 70 hours per week with crisis meetings and weekend work, while sprints at OpenAI and Anthropic can top 90 hours over a seven-day period.
MIT innovation scholar Neil Thompson argues that while AI may reduce some work, new work emerges, undermining the assumption that efficiency gains automatically translate to shorter weeks.
THE CLUSTER