Jonathan, a philosopher, describes his journey from feeling existentially threatened by AI’s growing reasoning capabilities to welcoming it, after realizing his fear stemmed from treating philosophy as professional production rather than personal cultivation.
Initial Existential Threat from AI
Jonathan’s first reaction to GPT-3’s abilities in reasoning, argument reconstruction, logic, and math was awe and fear, not just for job security but because a craft honed over decades appeared easily automated by “cold silicon.”
He compares the experience to Augustine learning of Rome’s fall: the “Eternal City” of the human mind seemed to have fallen under siege, continuing a series of modern humiliations (Copernicus, Darwin) that stripped away human exceptionalism.
The threat felt total because if intelligence — humanity’s last claim to uniqueness — could be replicated, the purpose of years of philosophical training seemed to vanish.
Philosophy as Production vs. Cultivation
Investigating his fear revealed he had been treating philosophy primarily as production (publishing papers, gaining tenure, chasing views) rather than cultivation (the ancient tradition of philosophy as an art of life and tool for virtue).
Socrates, Epicurus, Confucius, Jesus, and Buddha wrote no books; for them philosophy was a ladder to a better life, not an output. In the Phaedrus, Socrates compares philosophy to a fruitful seed: plant it in the soil of a soul, not the dead pages of a book.
Marcus Aurelius likewise thanked the gods he didn’t puzzle over syllogisms or celestial phenomena but pursued philosophy for living well.
Once Jonathan reoriented to cultivation, AI ceased to threaten: every lecture and interview still cultivates him, and if AI eventually produces better arguments or ethical systems, he should welcome that outcome.
Interiority as a Defensive Moat
On the production side, one defense is whether philosophy requires access to human interiority — the felt experience of being human — which behavioralism denies but phenomenology and ethics demand.
If philosophy is a “manual for life,” areas needing interiority (phenomenology, ethics) will be more robust than economic or political analysis, which AI may replace sooner.
This defense concerns the production process itself: the author must know what it feels like to be human to write philosophy appropriate for humans.
Personality as a Defensive Moat
A second production-side defense is injecting distinctive personality into philosophical work, visible across history: British analytical philosophy injects little, while Continental thinkers like Rousseau and Nietzsche embed literary passages that convey a singular voice.
As AI cheapens writing (even before matching human quality), rhetoric — spoken, personality-rich communication via podcasts and video — will regain value, becoming the new “printing press” for the spoken word.
Personality becomes a moat because it cannot be easily replicated by models trained on averaged outputs.
Cultivation Is Not Permanently AI-Proof
Jonathan cautions that cultivation is not an absolute refuge: Nick Bostrom’s Deep Utopia describes future neurotechnology that could rewrite synapses directly, letting a person “download” Hegel or virtue like Neo in The Matrix.
Bostrom argues this may be among the final technologies humans invent, because the brain is the most complex known object and writing to it requires manipulating billions or trillions of synapses.
Still, reading and absorbing knowledge is likely one of the last cognitive activities to be automated, making it the most defensible use of time until that distant horizon.
AI Opens New Philosophical Frontiers
Historically, theoretical questions become empirical when new mechanisms arise (e.g., Aristotelian metaphysics became physics). AI will play a similar role for philosophy.
Example: Jonathan debated Patricia Kitcher in 2017 over whether a deterministic Newtonian system could produce reasoning behavior; his CS/AI background led him to say yes, while she insisted reason was categorically different.
With OpenAI’s o3 model (late 2024), purely deterministic instructions now generate mathematical reasoning surpassing 99.9% of humans, turning a millennia-old speculative debate into an empirical result.
Jonathan is launching a philosophy-of-AI interview series with Cosmos (starting with Bostrom) to explore these newly tractable questions.
