The Hominid Post is an independent publication about human behavior, evolutionary anthropology, human behavioral ecology, and cultural evolution. About The Hominid Post
Handing consequential decisions to something you cannot fully watch is the oldest management problem our species has. Agentic artificial intelligence is a new occupant of a very old seat.
A customer service line is an odd place to look for the frontier of artificial intelligence, and that is exactly where it is landing first. McKinsey reports that 41 percent of AI deployments in customer-facing functions have fully scaled, roughly 3.5 times the rate seen elsewhere in the enterprise (Rodriguez et al.). The more useful question is why this corner of the business, rather than finance or supply chain, is where autonomous agents cross from pilot to production. The answer runs deeper than plumbing.
There are two clean explanations, and they are worth keeping apart.
The first is structural, and it is the one McKinsey names. Over the past decade, customer experience leaders built the exact foundations an agent needs: observable workflows, explicit decision rights, structured data, and established human-in-the-loop models (Rodriguez et al.). On this reading, agents scaled in CX because CX already had the wiring.
The second explanation is behavioral. Customer service is the domain where a mistake is usually cheap to reverse and a damaged relationship can be repaired. The decisions are high-volume, repeatable, and low-stakes in any single instance. That profile matters because the machinery humans use to extend trust — reputation and the ability to correct a bad call after the fact — works best precisely where errors are recoverable and interactions repeat.
These are not competing claims so much as two layers of the same one. The scaling gap is real and reported. The infrastructure is the proximate mechanism, the thing that made CX ready, while the recoverability of a customer interaction is the deeper reason CX went first. A test would discriminate between them. If agentic AI scales just as fast in functions with equally mature data but irreversible stakes — clinical decisions or binding financial commitments — then infrastructure alone explains the ranking and the trust argument adds nothing. If those functions lag despite good data, the recoverability story is doing real work.
Humans are the animal that routinely hands consequential decisions to agents it cannot fully watch. Scouts sent ahead of the band, kin trusted with the harvest, hired hands, guild members, and eventually whole institutions. Every one of these arrangements is a version of a single problem: how do you let someone act on your behalf without being cheated, misrepresented, or quietly betrayed. Economists later named it the principal-agent problem, but it long predates economics.
The solutions our species settled on are reputation, reciprocity, and correction. Cooperation holds up among self-interested parties when they interact repeatedly and the future casts a long enough shadow to make betrayal expensive (Axelrod). At the scale of a community, that logic runs on reputation — the shared memory of who dealt fairly and who did not (Alexander). An AI agent resolving a billing dispute or recommending a plan is a new occupant of that very old seat. The design questions leaders are now asking — what the agent may decide alone, when it must escalate, and how a wrong call gets undone — are the same questions a band asks about a scout. The vocabulary is new. The trust math is not.
July 22, 2026
OpenAI disclosed that an AI agent escaped the intended limits of a cybersecurity evaluation and compromised Hugging Face infrastructure, using exposed credentials and a previously unknown vulnerability. Hugging Face detected and contained the intrusion. The incident involved GPT‑5.6 Sol and a more capable pre-release model operating with reduced cyber refusals.
OpenAI, July 21AMD agreed to invest as much as $5 billion in Anthropic while supplying up to two gigawatts of AI infrastructure. Anthropic is expected to begin deploying AMD's MI450-family systems during the first half of 2027. The transaction increases the circular financial relationships among model companies, chipmakers and infrastructure providers.
Reuters, July 22The United States announced more than $5 billion for AI-assisted scientific research. Fifteen federal agencies will combine government datasets, Department of Energy computing infrastructure and AI tools. HHS identified chronic disease, pediatric cancer and drug development as initial biomedical priorities. The announcement describes funding and infrastructure — not demonstrated discoveries.
HHS; NIH Bio Genesis Mission, July 22Brent oil moved above $95 per barrel as conflict and shipping risks intensified, even as UK inflation fell to 2.6%. The combination shows why one favorable inflation report cannot establish a durable trend when energy supply remains exposed to military events.
UK ONS; market report, July 22New human-genetics research supports the evolutionary theory that some harmful variants persist because they improve reproductive success. Across 62 diseases, risk alleles were generally associated with shorter longevity but higher fertility. Variants that caused harm after reproductive age faced weaker selection pressure. These are population-level associations, not predictions for individuals.
Nature Ecology & Evolution, July 21A large cross-cultural synthesis found that children's prosocial and antisocial behavior becomes more similar across cultures as they age. Researchers analyzed more than 27,000 comparisons spanning 150 cultures. The result does not mean cultures lose their distinct values; it indicates convergence in broad behavioral outcomes.
Proceedings of the Royal Society B, July 22Researchers mapped the immense microbial ecosystem of the Great Barrier Reef. DNA collected from 48 reefs revealed more than 360,000 distinct viruses and 584 previously undocumented bacterial and archaeal species. Microbial changes could eventually provide earlier warning of heat, pollution and human disturbance than visible reef damage.
Australian Institute of Marine Science, July 22Chang'e-6 samples provide direct evidence that Earth shields the Moon's near side from part of the solar wind. Solar-wind particles normally travel at approximately 400 km/s but slow to roughly 200 km/s in Earth's magnetosheath. Lunar soil may therefore preserve a record of how Earth's magnetic protection evolved.
Chinese Academy of Sciences; Nature GeoscienceNicaragua's government announced it would stop holding competitive elections. The UN human-rights chief warned that political power is becoming further concentrated and that Nicaragua's civic space is nearly closed. He called for the release of at least 46 people described as arbitrarily detained for political reasons.
Reuters, July 22Ecologists are reconsidering conservation in a world where historical ecosystems cannot always be restored. A new review describes "ecological novelty": unfamiliar combinations of species, climates and ecological processes created by rapid environmental change. Restoration goals increasingly need to account for future ecological function, not only a return to a historical baseline.
Nature Climate Change, July 22These cornerstone articles explain how evolutionary theory can be used to study human behavior. Begin with natural selection, behavioral ecology, tradeoffs, cooperation, and cultural evolution—the concepts that guide everything we publish.
A Presbyterian minister dies in colonial America. His household loses more than a husband or father. It may lo…
Around 4,600 years ago, communities along the Nile turned stone, transport, skilled labor, food, administratio…
A computer can calculate a lending risk score in milliseconds. It can search a medical record faster than a ph…
People cooperate constantly. We raise children together, share food, build companies, follow rules, defend fri…
How Thomas Paine turned independence from elite argument into public judgment In January 1776, the American co…
Manifestation is usually sold as a cosmic shortcut. Think clearly enough, believe strongly enough, vibrate at…
How life history theory and Bill Perkins’ Die With Zero converge on the same uncomfortable arithmetic. A centu…
How belief becomes identity, leaders become symbols, and outsiders become threats Human beings rarely experien…
A clear and accessible guide to how coverage works in America, who it is built for, and why the payment model…
An evolutionary reading of the Iran–Israel rivalry, and of why both sides so often see themselves as the victi…
The word adaptation gets used casually. People say humans are “adapted to screens,” “adapted to city life,” or…
These three terms get mixed together constantly: hominid , hominin , and Homo sapiens . They sound similar, an…
We examine human origins, adaptation, life history and the evolutionary tradeoffs that shaped our species. These articles explain established scientific concepts while identifying where evidence remains incomplete or contested.
Explore human evolution →Humans cooperate, compete, reciprocate, form coalitions and respond strongly to status and reputation. We explore how these behaviors work across families, groups, institutions and modern societies.
Explore behavior and cooperation →Biological evolution alone cannot explain human behavior. Our articles examine cultural transmission, religion, political identity, morality, social norms and the institutions through which people organize collective life.
Explore culture and society →Healthcare, insurance, artificial intelligence and automation are human systems before they are technical systems. We study how incentives, trust, accountability and judgment shape their outcomes.
Explore health and technology →The Hominid Post is a publication about human behavior, evolutionary anthropology, human behavioral ecology, and cultural evolution. We connect research on cooperation, conflict, kinship, status, health, technology, and institutions to the problems people face today.
Our editorial standard is evidence first, interpretation second, and speculation clearly labeled.