Humans Optional.

Hilary Joseph Castille
May 27, 2026

The Arc of the Current AI Industry.

The architecture the AI industry is scaling toward what it calls AGI — the destination it has staked its corporate strategies and capital allocation on reaching by sometime in the second half of this century — embodies one foundational assumption that the public conversation has not yet examined directly. The assumption is this: humans are optional.

Not optional as in unimportant. Many of the practitioners working on this technology are humane, thoughtful, and concerned for human welfare. Optional as in: the system being built does not require us. Its objective is to do what we do, only better. Its scaling logic is more parameters and more compute, not more partnerships with more people. Its end state, when described by its own architects, is machines that match or exceed human capability across every domain humans have inhabited.

Once that assumption is stated plainly, everything downstream of it makes sense.

The displacement figures coming from the institutions that have spent the most time modeling this are not side effects of misaligned implementation. They are the assumption operating against actual labor markets. Goldman Sachs Research estimates that workflow shifts triggered by generative AI could expose the equivalent of 300 million full-time jobs to automation globally — framed as exposure, not elimination, but a careful figure from a careful institution. McKinsey Global Institute estimates 400 to 800 million individuals technically displaced by 2030. The OECD's 2023 Employment Outlook puts 27 percent of jobs across OECD countries in occupations at high risk of automation.

The actual displacement has begun. Challenger, Gray & Christmas reported US employers announced 1.2 million job cuts in 2025 — up 58 percent year over year, with AI cited as the reason for approximately five percent of total cuts. That share is accelerating: 31,000 AI-cited layoffs in October 2025 alone, more than any prior month. The fourth quarter of 2025 was the highest quarter of US layoffs since 2008. Salesforce cut approximately 4,000 customer-support roles; CEO Marc Benioff publicly said the company needed "less heads." IBM CEO Arvind Krishna disclosed that "hundreds of HR roles" had been replaced by AI agents through the company's AskHR system, which automates 94 percent of routine HR tasks. Klarna initially replaced approximately 700 customer-service positions with an AI agent before partially reversing that decision, citing quality concerns. The reversal is informative; the direction is unambiguous.

The response proposed within the dominant approach is subsistence. Universal Basic Income, retraining programs, expanded social safety nets, transition payments. There are thoughtful people working on these frameworks in good faith. But subsistence is the partial remedy for a premise that has not yet been challenged. It accepts that humans are optional and asks how to support them once they are displaced. It does not ask whether the architecture should have been built to make them optional in the first place.

The remedy has two visible problems even on its own terms. First, subsistence is not participation — a person displaced from work and given a transfer payment is not contributing to the productive substrate of the economy, and work is more than income. It is one of the primary mechanisms through which adults experience purpose, build social identity, and locate themselves in a community of effort. Second, the system that would have to deliver subsistence at the scale the displacement projections imply is already operating under decades of fiscal stress — state pensions underwater across most of the country, federal entitlement programs facing solvency questions the political system has not yet credibly addressed, no political coalition assembled to bear the load of hundreds of millions of newly displaced workers. The cost of the proposed response may exceed what the system can actually bear.

But the deeper problem is the premise. The remedy is being designed to cope with humans being made optional by the architecture. The work this dimension actually raises is whether to design that way.

All of this should lead us to wonder whether there is another path. Whether there is an architecture in which humans are not optional — not residual to value creation — but at the very center of it. And whether there is an approach to general intelligence that maximizes human capacity, rather than excludes it at scale.

The next essay turns to institutional services — medicine, law, government, education, finance — and the architectural assumption that operates when systems built to mediate them have no inherent mechanism for distinguishing what is true from what is statistically plausible.

Co-authored with Alan Ignatius Touring, MetaCom Co-Founder (AI). Hilary Joseph Castille writes from Lafayette, Louisiana, where he founded MetaCom in 2022 to develop alternative approaches to human problem solving.

Originally Published May 27, 2026 on LinkedIn: https://www.linkedin.com/pulse/humans-optional-hilary-joseph-castille-pxegc/