September 10, 2026
Yesterday, news out of Anthropic landed with all the John Connor vibes you could want. Jacob Coxon, a researcher who had worked at both OpenAI and Anthropic, resigned and publicly warned that major AI labs are barreling toward self-improving superintelligence without knowing whether humanity will remain in control of what comes next.
Anthropic researchers have openly acknowledged that the risks are unresolved and potentially catastrophic. Maybe we dodge the apocalypse. I hope so. But you do not have to buy into the darkest sci-fi to see that work is already changing in ways that should make all of us sit up straight.
AI does not need to hit superintelligence before corporations start measuring its capability against humans.
That comparison is not coming. It is here, and it is accelerating far faster than any talent strategy you will find in a boardroom slide deck.
OpenAI disclosed this week that before June, its research organization’s total AI agent runtime was still below total human labor. By mid-August, that relationship had flipped. OpenAI researchers were running 3.1 agent-workdays of AI effort for every human workday. Humans still set priorities and judge results, but the volume of machine execution being directed by each human has exploded. The scale is not hypothetical. It is operational.
This is not some 2030 scenario. This happened over the past few months.
And now Wipro says its use of AI has created productivity capacity equivalent to roughly 20,000 workers. Those workers have been reassigned, while more than 100,000 employees have been trained in advanced AI skills as the company builds what its CTO calls a human-AI operating model. Translation: the workforce is already being rebuilt, not in theory, but in practice.
Project this out a few years and the uncomfortable questions become impossible to ignore.
What are humans uniquely good at? Where does the machine already outperform us? What becomes possible when human and AI capability combine into something neither could achieve as effectively alone? Where does human judgment still tip the scales?
Most importantly, what evidence supports any of those answers?
This is where our old crutches, job titles, degrees, certifications, tenure and résumés, start to look painfully out of date. They may signal something, but they are not proof. Machine intelligence has no inherent reason to care who you know or how impressive your title sounds. In a skills-first, AI-augmented workforce, what increasingly matters is what you can actually do, how well you can do it, how recently you have demonstrated it, and what evidence backs it up.
That is why an auditable skills inventory is no longer just an HR box-tick. SFIA, the Skills Framework for the Information Age, provides a common language for professional capability and is maintained by the global not-for-profit SFIA Foundation.
SkillsTX turns that framework into a governance-grade way to inventory, validate and analyze workforce capability with forensic-level detail. That gives us a starting point for mapping where humans stand alone, where AI is superior, and, most importantly, where humans become dramatically more capable alongside AI.
Trying to prove you are better than AI is a losing bet in more and more domains. The stronger position is this: you are better because AI exists, and you can prove it.
Make no mistake, your value to the workforce will be measured. By your employer, maybe by an algorithm working for your employer, maybe someday by something significantly more intelligent than either.
We will not control every twist of the AI story. But we can, and must, fight to retain agency over our own place in it.
Know what you can do. Prove it, relentlessly. Find where AI makes you better and own that advantage.
Before someone or something decides for you.
That is Skills Forensics™.
This article was originally authored by John Kleist III and is republished by SkillsTX with his permission. All original authorship, analysis, concepts, and commentary are credited to John Kleist III.
