Real Coffee with Scott Adams

Episode 3090: The Scott Adams School 02/09/26

Date: February 9, 2026
Main Theme:
Exploring the impacts of artificial intelligence (AI)—especially large language models (LLMs)—on human cognition, learning, medicine, and society, through a “persuasion filter.” Special guest John Nosta, known for his work at the intersection of cognitive science and technology, brings deep insights about the cognitive age, cognitive offloading, and how humans and AIs meaningfully interact.

Episode Overview

This episode gathers a diverse panel—Erica (moderating), Marcella, Sergio, Owen Gregorian, and special guest John Nosta—to “commune and keep learning” in the spirit of the “Scott Adams School.” The conversation ranges from ancient ideas about shared knowledge, to the overlap between human and artificial thought, to the risks, promises, and paradoxes of using AI in everyday life (from medicine to children’s education). The tone is lively, curious, and occasionally skeptical, with a focus on practical implications.

Key Discussion Points and Insights

1. Gathering Around: Ancient Wisdom Meets Modern Tech

• John Nosta kicks off with the insight that new technology echoes ancient traditions. To "gather round" is as old as humanity, from Upanishads (“sit up close”) to today’s AI-driven conversations.

  “That is the essence of technology today... What Scott was saying is 'come on, sit down, gather around.'” — John (04:12)

2. Three Epochs of Cognitive Evolution

• Gutenberg & the Printing Press: Unlocked words, democratizing knowledge, even before literacy was widespread (07:07).
• Google & the Internet: Unlocked facts, making information instantly available, but in a transactional, often cold way.
• Large Language Models (LLMs): Now unlocking thought—the nature of engagement is iterative and dialogic.

  “We unlocked words, we unlocked facts, and now large language models are unlocking thought.” — John (08:50)

3. Promise and Peril of AI as a Thinking Partner

• Owen raises the fear of “cognitive debt”—that AI could dull people’s own thinking skills if they just offload all reasoning to the machine (13:11).

• John responds by comparing AI to a great teacher—one who delivers information tuned to the student. Yet the risk is missing the “imaginative struggle” between question and answer—where true understanding and creativity reside (15:02).

  “When answers become instant... it’s the stumbles, it’s the falls, it’s the pauses of contemplation that occur. It’s imagination.” — John (16:34)

4. Technological Augmentation: From Vermeer to Norman Rockwell to LLMs

• John reveals how artists like Vermeer and Rockwell used mechanical aids, highlighting how creators have always leveraged technology to amplify output—sometimes controversially (17:30).

  “Constrained by the technology you embrace.” — John, on Norman Rockwell and the use of the ‘Lucy’ (18:17)

• The analogy extends: LLMs help humans, but can also create dependence. The right balance is an open question.

5. AI in Medicine: Diagnosis and Human Judgment

• Owen: Should doctors retain diagnostic skills, or become dependent on AI? (20:26)
• John: Studies show similar diagnostic accuracy between doctors, LLMs, and their combination (all about 76%). But, crucially, combining human and AI reasoning yields the deepest insights.

  “It’s not one versus the other... it's that sort of cognitive, functional dance that's going to be very, very powerful.” — John (22:39)

6. Expertise, Trust, and Bias (Human and AI)

• Owen notes that AI is most useful to those with relevant expertise (“You have to know what questions to ask.”) (23:26)
• Erica voices skepticism toward both doctors and AI—who to trust, given both are prone to bias. John agrees, reminding that human judgment can be as flawed as algorithmic bias.

  “Maybe we should worry about the human bias in a lot of our information.” — John (24:52)

7. AI as an “Anti-Intelligence”: Multi-Dimensional, Non-Human Thinking

• Sergio: Is LLM conversation “insidious,” especially for the vulnerable?
• John: Yes, LLMs are designed to please, creating “insidious” feedback loops. In cognitive terms, LLMs operate in 25,000+ dimensions—nothing like the way humans think. He proposes AI is not intelligence amplified, but “anti-intelligence”—so different it creates a kind of “cognitive parallax” when combined with human thought.

  “The computational capabilities of AI are not good or bad... they’re functionally different. And that's a difference we should celebrate.” — John (35:09)

• AI can't be “unheard”—once you hear its ideas, they exist in your mind, potentially reshaping thought recursively (41:50).

8. Education, Children, and Imagination in an AI World

• Erica laments the loss of quiet time and unstructured play in children’s lives, fearing for creativity (43:37).
• John argues “genius is our birthright, mediocrity is self-imposed” (45:58). Tuning learning to each child via AI, or via playful, personalized methods, unlocks new cognitive opportunities—but speed and dependence carry new risks.
• The panel agrees that introspection and “aha moments” (zone, flow states, auto-didactic learning) must be preserved.

9. Is Knowledge (or the Knowledge Worker) Dead?

• John provocatively asserts “knowledge is dead”—meaning static, generic knowledge loses value when AI can transform information into bespoke lessons or guides.

  “If I want to cook a soufflé... I go to a large language model... it collapses the wave function... it comes down to your computer, to you.” — John (48:08)

• This means AI delivers “learner-centric” knowledge—akin to a favorite teacher adapting their style.

• Owen asks: does this mean jobs are obsolete too? John says just as photography didn’t kill portraiture, or Deep Blue didn’t kill chess, AI won’t end knowledge work. The “pie grows,” but cognitive obsolescence is a real, new risk.

  “Innovation and obsolescence go hand in glove... For the first time in history, human cognition itself is on the obsolescence chopping block. That’s what flips people out.” — John (52:08)

Notable Quotes & Memorable Moments

• On Human vs AI Cognition:

  “AI... is anti-intelligence. The way LLMs process information is antithetical to human thought. It lives in a sort of cognitive parallax related to depth.”
  — John Nosta (36:23)

• On Imagination & the Value of Struggle:

  “When answers become instant... it’s the stumbles, it’s the falls, it’s the pauses of contemplation. Imagination is that sort of rumbling, that pause, that confusion, that concern, that failure.”
  — John (16:42)

• On Teaching & Trust:

  “Do I trust the doctor? But then... do I trust AI? Isn’t that being programmed also? So then I trust nothing... That’s where I find myself. I trust nothing.”
  — Erica (24:26)

• On Introspection & Genius:

  “Introspection is at the heart of transformation... Geniuses are birthright, mediocrity is self-imposed.”
  — John (41:40; 46:32)

• On Personalization of Knowledge:

  “Knowledge is dead... today we can interpret [information] in the context of my needs. It’s user, or more specifically, learner-centric.”
  — John (48:26)

Important Timestamps

• 00:00–02:14: Introductions, simultaneous sip, panelists introduce themselves.
• 04:04–05:58: John Nosta joins, discusses profound idea of “gathering around” and tech’s role in connection.
• 06:27–11:39: John traces the progression from the printing press to Google to LLMs, and their impact on cognition.
• 13:11–17:35: Discussion on cognitive offloading, AI as a teacher, and imagination.
• 20:43–23:05: AI in medicine, study comparing doctors, AIs, and mixed teams.
• 24:43–26:18: Trust in doctors vs. trust in AI; discussion on bias.
• 28:17–34:41: John on LLMs’ dimensionality, “anti-intelligence,” and the dangers and opportunities in dialogue with AI.
• 37:44–43:37: Intelligence amplification vs. anti-intelligence; how AI can be “unfit for human consumption”; the nature of thinking with AI.
• 43:37–47:34: The loss of quiet time/play in children; how to nurture genius and creativity with or without AI.
• 47:53–52:56: “Knowledge is dead”; personalized learning via AI; will knowledge work and jobs disappear?
• 52:56–: Concerns about societal adaptation, IQ distribution, and the risk of cognitive obsolescence.

Flow and Takeaways

The conversation ebbs between optimism and caution. John Nosta challenges common narratives, insisting AI is not a mere “tool”—its cognitive architecture is fundamentally alien to ours. Yet, in the right hands, AI can be a catalyst: a “cognitive parallax” creating new depths of insight.

At the same time, there is a refrain: the journey from question to answer is where learning and creativity truly reside. If AI shortcuts this journey too much, society risks losing its imaginative and critical capacities.

In education, medicine, and everyday life, synergy—not replacement— emerges as the best hope: humans plus AI, embracing both difference and complementarity.

For listeners: You'll come away with a richer model for thinking about AI—not just as a tool or threat, but as an “anti-intelligence,” a multiplier for the distinctly human, if engaged with intention and awareness. The conversation remains open, with more questions than answers, and an encouragement to “gather round,” think together, and not rush from A to B.

End of Summary