Daily Briefing — May 28, 2026

Anthropic is reportedly on the verge of its first profitable quarter, and there's a pretty clear reason why. Sometime around November 2025, the company switched its Enterprise pricing from a flat seat fee that included usage to a model where you pay per seat AND for every API token your team consumes. Existing customers are only finding out now, as their contracts come up for renewal — a change buried in fine print that a lot of procurement and engineering teams clearly missed.

The split between consumer and enterprise pricing is where it gets weird. Individual power users on the $100/month consumer plans are getting an absurd deal, something like $2,000 worth of API tokens for $200. But the moment you're a company running agents at scale through the Enterprise tier, that math flips. You're no longer getting a buffet, you're paying by the plate. Companies that built internal tooling assuming flat-rate pricing are now staring at bills they did not budget for.

Call it SaaS bait-and-switch or call it honest usage-based pricing — either way the signal is the same. The AI labs have figured out that the people who actually get value from these tools will pay, and pay a lot. Product-market fit in this context means your costs are about to scale with your ambition.

Tarika Barrett, the outgoing CEO of Girls Who Code, is pushing back on the dominant Silicon Valley narrative. While tech executives keep telling everyone that AI is about to automate away most careers, the students Girls Who Code works with are booing graduation speakers who bring up LLMs, and a measurable gender gap has opened up in who's actually using AI tools day to day. Women, disproportionately, are skeptical. The standard read is "uh oh, women are going to get left behind." Barrett argues the skepticism itself is a useful professional skill.

The reasons women cite for hesitation are concrete: AI makes confident errors, the energy and water cost of running these models at scale is real, and a handful of tech billionaires now wield enormous influence over how the technology gets deployed. These are the same questions companies are now scrambling to answer, often badly, after the fact. Walking into an AI rollout already asking "how does this fail and who pays the cost when it does" is risk assessment, not Luddism.

The broader context is that pure coding jobs are shrinking as AI assists more of the work. The value is moving toward people who can judge when AI output is wrong, when a deployment is reckless, and when a workflow needs a human checkpoint. Skepticism, paired with technical literacy, is exactly that judgment muscle.

On May 25, Pope Leo XIV published his first encyclical, Magnifica Humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence. An encyclical is roughly the highest-weight teaching document a Pope issues, so this is the Catholic Church planting a flag on AI as a defining moral issue of the moment, not a side concern. The document deliberately echoes Rerum Novarum, the 1891 encyclical from Pope Leo XIII that addressed labor and industrial capitalism, and marks its 135th anniversary. The framing is intentional: if factories were the disruptive "new thing" then, algorithms and automation are now.

The substance worth paying attention to is what the Pope chose to focus on. He doesn't dismiss AI as evil, and he doesn't get distracted by long-tail extinction risk arguments. He targets the "culture of power" — the fact that a small set of wealthy investors and a handful of US tech companies currently decide how the technology gets built and who has access to it. He frames AI as part of the invisible infrastructure of daily life: an algorithm decides what you see, another filters what you read, others shape how your job and your information get distributed. The concern is concentrated power over that infrastructure, not the infrastructure itself.

Christopher Olah, cofounder of Anthropic, was invited to speak at the presentation — the first time a frontier AI lab has been given that kind of platform at the Vatican. That signals a deliberate strategy by the Church to move from commentator to participant in how AI gets governed, and it gives moral cover to safety-focused labs while putting pressure on the rest of the industry.