I was fascinated by John Nosta’s post, on anti-intelligence… So, I thought about my own idea of ante-intelligence…

—

John Nosta's description of much of what we call AI as "anti-intelligence" strikes me as productively unsettling (which is the good unsettling…). Not because these systems lack capability, but because they can generate fluent, coherent outputs while bypassing the lived, effortful, consequential processes through which human understanding is formed. They can simulate knowing without memory, stakes, embodiment, or intention. I tend to defer to my instincts more than I should, but I am fascinated by the inverse.

John’s framing suggests a companion idea: ante-intelligence.

If anti-intelligence describes an inversion of human cognition, ante-intelligence is a possible naming of the human foundation that must precede and guide it.

By ante-intelligence, I mean the purposeful, context-rich, value-laden reasoning that sets the conditions for everything that follows. It is the curiosity that asks better questions before the prompt, self-scrutiny that resists easy coherence. The experiential memory and ethical judgment that determine whether a compelling output can survive contact with real patients, regulators, payers, biology, and consequence(s).

I do not see this as a challenge to John’s concept, but as its complement. Anti-intelligence helps us name the risk. Ante-intelligence helps us name the human capacity required to respond wisely.

This also connects closely to how I've been thinking about Asymmetric Learning in pharma and biotech innovation. Asymmetric Learning is not passive data absorption or faster mimicry. It is a faster, more honest, less ego-driven loop of insight generation under uncertainty. And that is precisely where ante-intelligence matters most.

Before AI runs portfolio models, virtual scenarios, or optimisation routines, ante-intelligence defines the values, constraints, and criteria for what matters: patient outcomes, scientific integrity, societal impact, and long-term trust. Before a model suggests go/no-go decisions, ante-intelligence surfaces hidden assumptions, explores counterfactuals with real curiosity, and applies normative judgment to trade-offs no model can legitimately resolve on its own. Before we accept coherent but weightless outputs, ante-intelligence insists on consequence, iteration, and feedback from reality.

In practical terms, this suggests hybrid decision systems in which human judgment does not appear only at the end as a rubber stamp, but at the beginning and throughout:

• governance structures with explicit human values and constraints

• decision protocols that use AI for breadth and speed, while reserving human judgment for trade-offs and accountability

• cultures that reward lived insight, honest questioning, and intellectual humility more than polished fluency

Seen this way, the labs hiring philosophers are not indulging a curiosity (asymmetriclearning.subs…). They may be recognizing the need for something deeper: forms of judgment, interpretation, and ethical orientation that advanced capability alone cannot provide.

Pharma and biotech, because they already operate under conditions of uncertainty, consequence, and moral complexity, are especially well positioned to lead here. The long-term winners may not be those with the most powerful models, but those with the strongest ante-intelligence: the human capacity to learn asymmetrically, decide responsibly, and align action with what truly matters.

This is not anti-AI. It is pro-intelligence. If anti-intelligence names the inversion, ante-intelligence names the human foundation that keeps us from mistaking fluent output for understanding.