Scale AI and Passes founder Lucy Guo says AI is making people work harder, not less: ‘I worked a 26-hour day’

Scale AI cofounder and Passes CEO Lucy Guo said AI can intensify work rather than reduce it, citing a 26-hour workday driven by the need to oversee autonomous agents. Her view aligns with an eight-month UC Berkeley study of a 200-person U.S. technology company, which found AI-enabled employees worked faster, assumed broader responsibilities, and extended their work hours. The article highlights a productivity paradox: AI may improve task speed while increasing workload and raising questions about output quality.

Analysis

This is not a material META earnings input: the relevant investable signal is that agent adoption may initially raise, rather than lower, knowledge-worker labor intensity. That dynamic favors platforms that monetize higher inference volume and workflow complexity, but it delays the margin-expansion thesis investors are embedding in enterprise software. For META, the more direct read-through is modestly positive for AI talent retention and model-development velocity, not advertising revenue; neither is likely to alter near-term consensus estimates.

The underappreciated second-order risk is customer scrutiny of agent-generated output. If faster task completion produces rework, compliance failures, or unreliable autonomous actions, enterprises may shift spending from broad AI-seat deployments toward controlled, human-in-the-loop systems. That would favor hyperscalers and frontier-model providers with distribution and safety tooling over smaller application-layer vendors whose valuations assume rapid labor substitution. Over the next 1-3 months, earnings commentary on inference costs, agent usage frequency, and enterprise renewal behavior matters more than productivity anecdotes; over 6-18 months, sustained usage without proportional headcount reduction would support AI infrastructure demand but weaken the software-margin payoff.

Consensus is likely too quick to equate AI adoption with immediate SG&A deflation. A more plausible near-term outcome is a productivity arms race: employers redeploy saved time into additional output, increasing compute consumption and potentially extending work hours, while labor costs remain sticky. That is structurally supportive of AI-capex beneficiaries, but META already trades primarily on ad-growth durability and returns on its elevated infrastructure spend; absent evidence that AI improves engagement, ad ranking, or cost per impression, this item is not a standalone catalyst.

Key Decisions for Investors

• No standalone META trade on this item; maintain existing exposure only if upcoming results demonstrate AI-driven engagement or ad-ranking gains that exceed incremental depreciation and inference-cost growth.
• Use the next META earnings call as an alert: add only if management quantifies durable monetization or efficiency benefits and FY capex/infrastructure guidance does not rise disproportionately; a further capex reset without revenue linkage would falsify the constructive read-through.
• For a 6-18 month thematic expression, prefer selective AI infrastructure exposure over application software priced for immediate labor displacement; wait for enterprise renewal data showing agent workloads are production-critical rather than experimental before initiating a broad software short.