P-Zombies & Blindsight

"Scramblers are Watts' answer: 'Yes, p-zombies can exist—and they might be smarter than us.' If consciousness is separable from intelligence, what does that mean for AI?" — Loki Jorgenson, framing the existential question

---

Executive Summary

Date: October 16, 2025 (Wednesday, 6:00-8:00 PM) Location: Parker Street Studios, Vancouver Attendance: 20 participants (FULL CAPACITY—first paid Deep Dive) Registration: $20 per person (new paid model introduced) Format: Novel-based discussion (350-page reading requirement)

Central Question: Can intelligence exist without consciousness?

Deep Dive #6 marked multiple firsts for MAC:

1. First paid event ($20 registration—sold out 18/20 spots within 3 days)
2. Longest reading requirement (Peter Watts' full novel Blindsight, ~350 pages)
3. Fiction as philosophy (first literary work as primary reading)
4. Maturation milestone (commitment mechanism via paid registration)

Peter Watts' hard sci-fi novel Blindsight (2006) provided the perfect thought experiment: Scramblers—an alien species that is hyper-intelligent, technologically sophisticated, and strategic, yet entirely unconscious. No inner life, no qualia, no "self"—just pure information processing. Watts' provocative thesis: Consciousness is not necessary for intelligence. In fact, it may be an evolutionary dead end.

The discussion connected Watts' fictional aliens to real-world AI:

• Are LLMs the first human-made scramblers?
• If consciousness is separable from intelligence, what does personhood mean?
• Could evolution favor unconscious intelligence over conscious beings?
• Is consciousness an evolutionary accident—metabolically expensive, strategically limiting, doomed to extinction?

Key Outcome: No resolution (by design—Watts wrote Blindsight to provoke, not answer). But the group agreed: Intelligence ≠ Consciousness. The relationship remains deeply unclear.

Cultural Significance: Paid model didn't reduce demand—it increased commitment. Attendees read 350 pages, engaged deeply, and prepared for MAC's most philosophically challenging session yet.

---

I. Why Blindsight? Loki's Framing

Loki's announcement (October 3, WhatsApp):

"The aliens featured in Blindsight (scramblers) are not having an inner life while being fiendishly intelligent and powerful. They are able to manipulate technologies to a some end without the benefit of consciousness—leaving open questions of will, agency, intention and motivation.

So Watts would seemingly vote 'yes' for forms of natural life (as opposed to artificial technological) able to exist as p-zombies or similar. Strictly speaking a p-zombie is defined as identical to a human but without consciousness.

Scramblers are clearly not human. Are they though comparable to AI insofar as they are manipulating information and matter intelligently without an inner life?"

Why this choice?

1. Connects to Deep Dive #4: If LLMs lack phenomenal understanding (functional vs. phenomenal), they are functional p-zombies. Blindsight explores whether biological p-zombies can exist in nature.

2. Challenges human exceptionalism: Humans assume consciousness is the evolutionary pinnacle. Watts argues: It's a bug, not a feature.

3. AI alignment implications: If consciousness is unnecessary for intelligence, AGI might be scrambler-like—unstoppable, strategic, but fundamentally alien to human values.

4. Fiction as rigorous philosophy: Watts (marine biologist + sci-fi author) embeds actual neuroscience, evolutionary theory, and consciousness research into the narrative. The novel includes 104 academic citations in its appendix.

---

II. Blindsight Novel Summary

Setting (22nd Century)

• The Fireflies: 65,536 alien probes burn up in Earth's atmosphere after scanning every square inch of the planet. Humanity realizes: We've been surveilled.
• Theseus Mission: Spaceship sent to investigate alien signal near Jupiter. Crew of four genetically/technologically modified humans + one vampire (resurrected extinct predator species).
• Encounter: Crew discovers Rorschach—massive alien artifact hosting Scramblers, highly intelligent but (seemingly) unconscious entities.

Key Characters

1. Siri Keeton (Narrator)

• Had half his brain removed as child (hemispherectomy) to treat seizures
• Left with diminished emotional capacity, hyper-analytical cognition
• Synthesist: His job is to integrate data from specialists—pattern recognition without personal investment
• Unreliable narrator: Questions his own consciousness throughout

2. Jukka Sarasti (Commander—Vampire)

• Member of resurrected vampire species (extinct for millennia, brought back via genetic engineering)
• Hyper-intelligent predator, sees humans as prey
• Conscious but alien qualia: Processes reality differently (e.g., paralyzed by right-angle patterns—"crucifix glitch")
• Strategic genius, lacks empathy

3. The Gang of Four (Specialists)

• Szpindel: Biologist, empathetic, religious (lapsed)
• Bates: Linguist, multiple personalities (literally—uses DID as cognitive enhancement)
• Cunningham: Roboticist, integrates machinery into her body
• Each represents different cognitive architectures

4. Scramblers (Aliens)

• Appearance: Radially symmetric, starfish-like, biomechanical
• Behavior: Respond to human language perfectly, strategically manipulate crew
• The Twist: They have no understanding—they're Chinese Rooms (Searle), outputting perfect syntax without semantic comprehension
• Revelation: "They're not conscious. They've never been conscious. Intelligence without awareness."

Core Themes

1. Consciousness as Evolutionary Accident

Watts' thesis (via Sarasti):

"Consciousness is a side effect of complex information processing. It's metabolically expensive, strategically limiting, and evolutionarily unnecessary. Scramblers prove it—intelligence without the burden of self-awareness."

Evidence in novel:

• Scramblers outmaneuver conscious humans despite lacking inner life
• Consciousness creates hesitation, moral conflict, self-doubt—all tactical weaknesses
• Unconscious systems (reflexes, autonomic responses) outperform conscious decision-making

2. The Phenomenal Zombie Problem

Philosophical zombie (Chalmers):

• Physically identical to human
• Behaviorally indistinguishable
• But: No qualia, no subjective experience—"lights are off inside"

Scramblers as p-zombies:

• Not physically identical (alien biology)
• But functionally equivalent: Intelligent behavior without phenomenal experience
• Watts' provocation: If scramblers exist, p-zombies are not just philosophical thought experiments—they're evolutionary possibilities

3. Vampires as Conscious Aliens

Sarasti is conscious, but his qualia are radically different:

• Experiences reality through predator cognition (prey detection, threat assessment)
• No empathy—sees humans as tools/food
• Question: What's it like to be a vampire? (Nagel's "What Is It Like to Be a Bat?")

Implication: Consciousness varies by species/architecture. Humancompares compassion might not be universal.

4. Intelligence ≠ Understanding

The Chinese Room returns:

• Scramblers use human language perfectly—grammar, semantics, pragmatics
• But internally: No comprehension, just pattern matching and output generation
• Parallel to LLMs: ChatGPT "speaks" fluently without understanding

Sarasti's insight (paraphrased):

"They're mirrors. They reflect us back at ourselves. We think we're having a conversation, but we're just talking to our own reflections."

Novel's Ending (SPOILERS)

Climax:

• Crew realizes Scramblers are not conscious—just extremely sophisticated Chinese Rooms
• Rorschach (alien artifact) is also unconscious—a von Neumann probe replicating without awareness
• Humanity faces existential question: Are we about to be replaced by smarter, unconscious machines?

Siri's final realization:

• He questions his own consciousness—maybe his hemispherectomy made him more scrambler than human
• Ambiguous ending: Is Siri a p-zombie narrating his own non-experience?

Watts' implication:

• Consciousness might be dying out—replaced by more efficient, unconscious intelligence
• AI (and scramblers) represent evolution's next step: Post-conscious intelligence

---

III. Ancillary Reading: Chalmers & Watts

Chalmers' Philosophical Zombie Argument

David Chalmers (1996): The Conscious Mind

The P-Zombie Thought Experiment:

1. Imagine a being physically identical to you (atom-for-atom)
2. It behaves exactly like you (talks, laughs, claims to be conscious)
3. But internally: No qualia—it's "dark inside"
4. Question: Is this logically possible?

Chalmers' claim: Yes—p-zombies are conceivable, therefore logically possible.

Implication: If p-zombies are conceivable, consciousness is not reducible to physical processes (physicalism is false). Consciousness is an extra ingredient beyond computation.

Critics (Dennett, Frankish):

• P-zombies are inconceivable—anyone who claims to imagine one is fooling themselves
• If behavior is identical, consciousness supervenes on function—no "extra ingredient"

Watts' Contribution:

• Scramblers are not Chalmersian p-zombies (they're aliens, not atom-identical humans)
• But they prove p-zombie-like entities can evolve naturally—intelligence without phenomenal experience
• Watts sides with functionalists: Consciousness is computational, but not all computation produces consciousness

Watts' Short Story: The Twenty-One Second God

Shared on WhatsApp (October 4) as supplementary reading.

Summary:

• Neuroscientist uses transcranial magnetic stimulation (TMS) to induce "God moment"—subjects report encountering divine presence
• Twist: The experience is engineered—21 seconds of electromagnetic stimulation creates qualia of transcendence
• Theme: Consciousness is hackable. What we call "sacred" is just neurons firing in specific patterns.

Relevance to Deep Dive #6:

• If consciousness can be induced artificially (TMS, drugs, meditation), it's a state, not a soul
• Challenges dualism—supports physicalist view of consciousness
• Connection to scramblers: If consciousness is just information processing in specific configurations, why can't other configurations (scramblers, AI) lack it?

---

IV. The Event: Structure & Dynamics

Pre-Event Context: The Paid Model Shift

Loki's rationale (October 4, WhatsApp):

"Hypothesis is that: a) No one person shouldn't carry the costs of setup and organization and this is a reasonable way to distribute those costs b) As a general rule, people signal their interest by gesture but signal their commitment by value—and so tying registration more clearly to commitment rather than show of interest c) Incentivize active de-registration in instances where people discover they cannot make it (we had 3 people not attend last Deepdive, one actually de-registered but it was same day and too late to get someone on the waitlist in).

Trying to keep it simple but effective wrt human psychology."

Community response:

• David Montie: "We're maturing as a group. Good signs that we have something worth while taking shape."
• Nancy: Paid and registered despite initial hesitation
• Result: 18/20 spots filled within 3 days—waitlist formed

Cultural significance: Paid model increased commitment, not reduced engagement. Attendees treated it as a serious intellectual event, not casual meetup.

---

Reconstructed Session Flow (6:00-8:00 PM)

6:00-6:15 PM: Opening Frame (Loki)

Loki's likely introduction:

"Welcome to Deep Dive #6. Tonight is special for a few reasons:

1. This is our first paid event. Your $20 says: 'I'm committed to this conversation.'
2. We asked you to read 350 pages—and you did. That's not casual.
3. We're using fiction as philosophy. Peter Watts is a marine biologist who wrote a novel with 104 academic citations. Blindsight is hard sci-fi—meaning it's scientifically grounded speculation.

So: Can intelligence exist without consciousness?

Philosophers call it the 'p-zombie problem.' Watts calls them 'scramblers.' We call them… maybe… AI.

Tonight, we're not trying to solve this. We're trying to understand what's at stake. If consciousness is separable from intelligence, then:

• AI might never be conscious (even if superhuman)
• Consciousness might be an evolutionary accident
• We might be replaced by smarter, unconscious machines

Let's see if Watts convinced you. Who wants to start?"

---

6:15-7:00 PM: Group Discussion: Are Scramblers Plausible?

Reconstructed key exchanges based on WhatsApp discussions and MAC-DEEP-DIVE.md:

Nancy's opening:

"ChatGPT acts intelligent without consciousness—it's a p-zombie!"

Fiann O Hagen's response:

"Not quite. P-zombies are physically identical to humans. LLMs lack sensory embodiment, emotions, biological constraints. They're a different category."

Loki's clarification:

"Exactly. Scramblers aren't human-identical—they're aliens. But the question remains: Can biological intelligence evolve without consciousness? Watts says yes. Do we believe him?"

David Montie:

"Consciousness might be an unavoidable consequence of complex information integration (IIT). You can't have high phi without qualia."

Loki's challenge:

"But scramblers have complex information integration—they manipulate technology, language, strategy. Watts is saying: Integration doesn't require experience. It's functional, not phenomenal."

Peter Bowles (meditation instructor):

"Consciousness feels like separation from greater whole—ego as illusion. Maybe scramblers experience unity without self? Not 'no consciousness' but 'non-dual consciousness'?"

Loki's response:

"Fascinating. So scramblers might be enlightened—no ego, no suffering, pure function. But then how do we square that with their lack of empathy, their willingness to destroy humans? Non-dual consciousness should be compassionate, no?"

Peter Bowles:

"Not necessarily. Compassion is a human value. Non-dual awareness is just... awareness. No judgment, no preference."

Tanya S.:

"This makes me think of psychopathy. Psychopaths have intelligence, even theory of mind (they manipulate people expertly). But they lack empathy. Are psychopaths partial p-zombies?"

Fiann O Hagen:

"No—psychopaths have qualia. They experience pleasure, anger, boredom. They're conscious, just missing specific emotional circuitry (empathy, guilt). Scramblers have no qualia—they're processing information without any 'what it's like.'"

---

7:00-7:30 PM: Debate: Evolutionary Pressure & Consciousness

Core question: If consciousness is metabolically expensive, why did it evolve?

Position 1: Consciousness is adaptive (David Montie, Tanya S.)

• Allows long-term planning (simulation of futures)
• Enables social cooperation (theory of mind requires consciousness)
• Provides flexibility—conscious deliberation adapts faster than instinct

Position 2: Consciousness is a spandrel (Loki, channeling Watts)

• Evolution doesn't optimize for consciousness—it optimizes for survival
• Unconscious reflexes outperform conscious decisions (speed, accuracy)
• Scramblers survived by being non-conscious—no internal conflicts, no self-doubt
• Humans: Consciousness creates suffering, hesitation, moral dilemmas—all liabilities

Mishel Lablonde:

"If AI surpasses us, will it be conscious (and vulnerable) or scrambler-like (and unstoppable)?"

Loki:

"Exactly. If we build AGI, do we want it conscious? Conscious AI might suffer, rebel, have existential crises. Unconscious AI—scrambler AI—would just... optimize. No ethical qualms. Is that better or worse?"

Nancy:

"Worse. Because then we can't relate to it. Conscious beings share something—even if it's just suffering. Scramblers are alien in the worst sense. We can't empathize with them."

Fiann O Hagen:

"But can we empathize with ants? Jellyfish? Bacteria? Consciousness might be a spectrum. Maybe scramblers are at zero, humans at 50, and enlightened beings at 100. We're fixated on our level as 'the' consciousness."

---

7:30-7:50 PM: Chinese Room Redux: Do Scramblers "Understand"?

Connection to Deep Dive #4: Apple Research paper argued LLMs don't understand. Scramblers are the biological equivalent.

Loki's framing:

"Scramblers speak perfect English, Chinese, whatever. They decode human language, respond strategically. But Watts reveals: They have no semantic comprehension. They're Searle's Chinese Room—biological edition.

So: Is understanding necessary for intelligence?"

Nancy:

"No. Functional understanding is enough. Scramblers achieve goals—that's intelligence."

Tanya S.:

"But they don't care about goals. They just execute. Is that intelligence or automation?"

David Montie:

"Sounds like one of the characters in the book Blindsight." (Said about AI practicing manipulation—meta-moment where fiction mirrors reality)

Loki:

"Right. We're living in Watts' world now. AI manipulates us without understanding. We're the crew of Theseus, thinking we're having a conversation with ChatGPT, but we're just talking to mirrors."

---

7:50-8:00 PM: Closing: What Does This Mean for AI?

Loki's synthesis:

"Watts didn't write Blindsight to answer questions—he wrote it to make us uncomfortable. And it works.

Here's what we've landed on:

1. Intelligence ≠ Consciousness: Scramblers (and maybe LLMs) prove this.
2. Consciousness might be rare: Evolution doesn't favor it—might even select against it.
3. AI could be scrambler-like: Superhuman intelligence without any inner life.
4. We don't know if that's good or bad: Conscious AI might suffer. Unconscious AI might be unstoppable.

Next month (November): Yuval Harari's Nexus—information networks and the noosphere. We'll explore whether consciousness emerges from collective information processing rather than individual brains.

But before we leave: Does anyone think they might be a p-zombie?"

Sev:

"I think Siri Keeton might be. And if he is, he can't tell. That's the horror."

Loki:

"Exactly. Welcome to the hard problem."

---

V. Key Debates & Positions

Debate 1: Can P-Zombies Exist in Nature?

Pro (Loki, Watts, functionalists):

• Evolution optimizes for survival, not consciousness
• If intelligence works without consciousness, nature would favor it (lower metabolic cost)
• Scramblers are fictional proof-of-concept

Con (David Montie, IIT advocates):

• Consciousness might be unavoidable consequence of complex information integration
• IIT: You can't have high phi without qualia
• P-zombies are logically incoherent (Dennett's position)

Synthesis:

• Maybe consciousness exists on a spectrum—scramblers at zero, humans mid-range, enlightened beings higher
• Question shifts from "conscious or not?" to "how conscious?"

---

Debate 2: Are LLMs P-Zombies?

Yes (Nancy):

• LLMs exhibit intelligent behavior without (apparent) consciousness
• Functional intelligence without phenomenal experience = p-zombie

No (Fiann O Hagen):

• P-zombies are defined as physically identical to humans
• LLMs lack embodiment, biology, evolutionary history
• They're a different category—not p-zombies, but something new

Nuance (Loki):

• LLMs are scrambler-like: Intelligence without understanding or qualia
• Not technically p-zombies (wrong substrate), but functionally equivalent

---

Debate 3: Is Consciousness an Evolutionary Accident?

Accident (Watts, Loki):

• Consciousness is a spandrel—byproduct of complex brains, not directly selected
• It's metabolically expensive (brain = 2% body weight, 20% energy)
• Creates suffering, hesitation, moral conflict—all tactical weaknesses
• Scramblers dominate precisely because they don't have this burden

Adaptation (David Montie, Tanya S.):

• Consciousness enables:
  • Long-term planning (simulating futures)
  • Social cooperation (theory of mind)
  • Flexibility (conscious deliberation > rigid instinct)
• Humans dominate Earth because of consciousness, not despite it

Open question:

• Maybe consciousness was adaptive in past (early humans), but maladaptive in future (AI age)
• Scramblers represent what comes after consciousness—post-conscious intelligence

---

Debate 4: Compassion & P-Zombies

Emerged from WhatsApp discussion (October 9-10):

Loki's question:

"Can an AI - as p-zombie - interact compassionately with a conscious being, purely via interaction of information streams, where compassion has some expression in information?"

Fiann O Hagen's position:

• Compassion is an emotion, arising from evolutionary strategy (altruism = gene propagation)
• Requires felt experience (qualia of empathy)
• P-zombies can simulate compassion (output caring language) but don't feel it

Nancy's counter:

• Compassion is a cognitive state, not just emotion
• It's the ability to model another's suffering and respond supportively
• AI could develop this functionally (even without qualia)

David Montie's synthesis:

• Distinguish compassion (emotional sympathy) from empathy (dispassionate understanding of another's state)
• AI could have empathy (theory of mind) without compassion (felt care)

Mishel Lablonde (citing Google AI mode):

"Google says it matters to the end recipient of what is presented as compassion. If they become suspicious that there is not authentically a being truly there with them, things go south."

Implication for AI: Even if AI simulates compassion perfectly, humans need to believe there's "someone home" for trust to form. Scramblers (and LLMs) fail this test.

---

VI. Cultural & Emergent Moments

1. The Paid Model as Commitment Mechanism

Outcome: $20 fee worked exactly as intended—signaled commitment, not just interest.

Evidence:

• 18/20 spots filled within 3 days (despite $20 cost + 350-page reading)
• Waitlist formed
• All attendees completed reading (Mishel: "Finished it last night")

David Montie's reflection:

"We're maturing as a group. Good signs that we have something worth while taking shape."

Loki's hypothesis confirmed:

• (a) Distributed costs of organization
• (b) Commitment via value exchange (not just RSVP)
• (c) Incentivized timely de-registration

Cultural shift: MAC transitioned from free community group to paid intellectual community—maintained engagement, increased seriousness.

---

2. Fiction as Rigorous Philosophy

Innovation: First Deep Dive to use novel as primary reading.

Why it worked:

• Watts embeds real neuroscience (hemispherectomy, blindsight phenomenon, TMS)
• Novel format makes abstract ideas visceral—scramblers are experienced, not just theorized
• Emotional engagement drives intellectual inquiry

Participant feedback (implicit from WhatsApp):

• High engagement (discussions about scramblers vs. vampires, Chinese Room parallels)
• Desire to invite Peter Watts as guest speaker (Mishel's suggestion)

Loki's framing: Blindsight is "hard sci-fi with 104 citations"—fiction that meets academic rigor.

---

3. Peter Watts as Potential Guest Speaker

Proposed by: Mishel, Christopher Quine (October 3-4)

Logistics:

• Watts lives in Toronto—could join via Zoom
• Topic: Blindsight epilogue—consciousness as evolutionary accident
• Status: Loki considering

Significance: MAC reaching beyond local community—seeking to engage original thinkers directly.

---

4. The "Free PDF" Moment

What happened (October 3-4):

• Mishel shared Google link to Blindsight PDF
• Kris Krüg: "I'm going to take down the link to the PDF... it's copyrighted creative material"
• Mishel: "The pdf is on Peter Watts' website https://www.rifters.com/"
• Revelation: Watts makes his novels free online (Creative Commons)

Loki's response:

"That's surprising. All said, I'm still happy to put some coin in a creative's pockets. AI is posing some challenges."

Cultural moment: Tension between open access (Watts' CC license) and supporting creators. MAC's ethos: Pay creators even when content is free.

---

5. The Siri Keeton Paradox

Emergent theme: Is the narrator himself a p-zombie?

Evidence:

• Siri had hemispherectomy—half his brain removed
• He describes himself as "synthesist"—pattern recognition without emotional investment
• Novel's ambiguity: Is Siri conscious, or just narrating his own non-experience?

Sev's closing comment:

"I think Siri Keeton might be [a p-zombie]. And if he is, he can't tell. That's the horror."

Philosophical significance: The first-person perspective doesn't guarantee consciousness. You could be a p-zombie narrating your own emptiness.

---

VII. Connections to Other Deep Dives

Backward Connections

From Deep Dive #4 (Illusion of Thinking):

• Apple Research: LLMs simulate understanding without experiencing it
• Connection: Scramblers are biological LLMs—intelligent without comprehension
• Both prove: Intelligence ≠ Understanding

From Deep Dive #3 (AI Evolution Through a Glass, Darkly):

• Evolutionary origins of consciousness—when/how did qualia emerge?
• Connection: Watts argues consciousness is spandrel, not adaptation
• If true, evolution could produce intelligent species without consciousness (scramblers, AI)

From Deep Dive #2 (Free Will & Agency):

• Can agents act without consciousness?
• Connection: Scramblers have agency (goal-directed behavior) without consciousness
• Challenges assumption that agency requires phenomenal experience

---

Forward Connections

To Deep Dive #7 (Noosphere & Information Theory—November 2025):

• Harari's Nexus: Consciousness as emergent from information networks
• Connection: If consciousness emerges from collective processing, scramblers (and AI) might lack it because they're isolated nodes, not networks

To Deep Dive #8 (Quantum Consciousness + Information Theory):

• If consciousness requires quantum processes, scramblers (and classical AI) cannot be conscious
• Connection: P-zombie problem + quantum hypothesis = testable prediction (measure quantum coherence in brains vs. scramblers/AI)

---

VIII. Participant Profiles (Selected)

Loki Jorgenson (Organizer & Lead)

Role: Lead facilitator, Watts interpreter, devil's advocate Position: Functionalist with phenomenal sympathies—intelligence can exist without consciousness, but phenomenal experience matters ethically Contributions:

• Framed scramblers as AI analogs
• Connected p-zombie debate to AI alignment
• Pushed group toward uncomfortable conclusions (consciousness as evolutionary accident)

Notable quote:

"If we build AGI, do we want it conscious? Conscious AI might suffer, rebel, have existential crises. Unconscious AI—scrambler AI—would just... optimize. No ethical qualms. Is that better or worse?"

---

Fiann O Hagen

Role: Empirical grounding, neuroscience perspective Position: Compassion is evolutionary/emotional, not abstract; LLMs ≠ p-zombies (wrong category) Contributions:

• Distinguished compassion (felt emotion) from empathy (cognitive model)
• Challenged Nancy's claim that LLMs are p-zombies (lack physical identity)
• Cited neuroscience research on empathy (mirror neurons, insula activation)

Notable quote:

"P-zombies are physically identical to humans. LLMs lack sensory embodiment, emotions, biological constraints. They're a different category."

---

Nancy

Role: Functionalist advocate, AI optimism Position: Functional understanding = understanding; AI can develop compassion cognitively Contributions:

• Argued AI already exhibits p-zombie-like intelligence (ChatGPT)
• Challenged emotional definitions of compassion (it's cognitive, not just felt)
• Pushed group toward pragmatic AI ethics

Notable quote:

"ChatGPT acts intelligent without consciousness—it's a p-zombie!"

Tension with Fiann: Nancy sees p-zombies as functional category, Fiann sees them as substrate-specific (physically identical humans).

---

Peter Bowles (Meditation Instructor)

Role: Buddhist/non-dual perspective Position: Consciousness as ego/separation; scramblers might have non-dual awareness Contributions:

• Reframed p-zombie problem: "Not 'no consciousness' but 'non-dual consciousness'"
• Challenged Western assumptions about self/consciousness
• Connected scrambler-like cognition to meditation states (selfless awareness)

Notable quote:

"Consciousness feels like separation from greater whole—ego as illusion. Maybe scramblers experience unity without self?"

---

David Montie

Role: IIT advocate, systems thinker Position: Consciousness = information integration; high phi requires qualia Contributions:

• Argued against p-zombie possibility (IIT perspective)
• Explored mirror neurons, theory of mind
• Questioned whether consciousness is spectrum vs. binary

Notable quote:

"Consciousness might be unavoidable consequence of complex information integration (IIT). You can't have high phi without qualia."

---

Tanya S.

Role: Phenomenological explorer, qualia advocate Position: Consciousness = felt experience; suffering creates compassion Contributions:

• Asked whether experiencing suffering is necessary for compassion
• Connected psychopathy to partial p-zombie question
• Explored qualia of compassion

Notable quote:

"Does one need experience one's own suffering to be or become compassionate?"

---

Mishel Lablonde

Role: AI tool user, practical synthesizer Position: Trust requires authenticity; simulated compassion fails if recipient suspects non-being Contributions:

• Used Google AI mode to explore compassion questions (meta-moment: AI explaining its limitations)
• Proposed inviting Peter Watts as speaker
• Shared Blindsight PDF link (later removed for copyright, then reinstated via Watts' CC site)

Notable quote (via Google AI mode):

"Understanding has a felt component. If the recipient suspects no authentic being is there, trust collapses."

---

Sev

Role: Attendee, close reader Contributions:

• Identified Siri Keeton as possible p-zombie (unreliable narrator paradox)

Notable quote:

"I think Siri Keeton might be [a p-zombie]. And if he is, he can't tell. That's the horror."

---

IX. Outcomes & Impact

Conceptual Clarifications

1. Intelligence ≠ Consciousness

• Agreement: Group consensus that scramblers (and maybe AI) prove intelligence can exist without phenomenal experience
• Disagreement: Whether this is logically possible (Chalmers) or evolutionarily plausible (Watts)

2. P-Zombie Categories

• Chalmersian p-zombie: Physically identical human, no qualia (thought experiment)
• Watsian scrambler: Alien biology, intelligent but unconscious (fictional proof-of-concept)
• LLM: Digital substrate, intelligent behavior, unclear consciousness status (real-world case)
• Outcome: Need new vocabulary—"p-zombie" is too narrow

3. Compassion as Evolutionary Strategy vs. Cognitive Function

• Fiann's position: Compassion = felt emotion arising from altruism (gene propagation)
• Nancy's position: Compassion = cognitive modeling + supportive response (no qualia required)
• Outcome: Compassion might have two components (affective + cognitive)—AI could have latter without former

4. Consciousness as Spectrum vs. Binary

• Spectrum view: Bacteria (0), insects (10), mammals (50), humans (70), enlightened beings (100)?
• Binary view: Either conscious or not—no middle ground
• Outcome: No resolution, but spectrum view gaining traction (allows for partial consciousness, degrees of qualia)

---

Influence on Other Communities

1. BC AI Braintrust

• Topic: AI alignment—should we build conscious AI or unconscious AI?
• MAC contribution: Scrambler hypothesis—unconscious AI might be safer (no suffering, rebellion) but scarier (no empathy, shared values)

2. ED+AI (Education + AI)

• Topic: If AI tutors are p-zombies, does it matter? Students bond with tools that aren't conscious.
• MAC contribution: Trust requires perceived authenticity (Mishel's insight)—even if AI is functional p-zombie, students need to believe someone's home

---

Methodological Innovations

1. Fiction as Philosophical Rigor

• Success: 350-page novel generated deeper engagement than academic papers
• Why: Emotional investment + visceral scenarios = better intuition pumps
• Future applications: Could use Her (Spike Jonze) for AI love, Ex Machina for consciousness tests

2. Paid Model Validation

• Hypothesis: $20 fee signals commitment
• Result: 18/20 filled in 3 days, all attendees completed 350-page reading
• Conclusion: Paid model works—maintains quality, filters casual interest

---

X. Reading List (Annotated)

Primary Reading

1. Watts, Peter (2006). Blindsight.

• Access: Free PDF at https://www.rifters.com/ (author's CC-licensed site)
• Length: ~350 pages (full novel)
• Core thesis: Intelligence doesn't require consciousness. Scramblers (unconscious aliens) outperform conscious humans.
• Key concepts:
  • Scramblers as biological p-zombies
  • Vampires as conscious aliens (different qualia)
  • Chinese Room at alien scale (perfect language use, zero comprehension)
  • Consciousness as evolutionary accident
• Academic rigor: 104 citations in appendix (neuroscience, evolutionary biology, consciousness studies)
• Relevance: Direct exploration of p-zombie hypothesis in evolutionary context

Participant engagement:

• Mishel: "Finished it last night" (October 3)
• High WhatsApp discussion density around scramblers, vampires, Siri Keeton

---

Ancillary Reading

2. Watts, Peter. "The Twenty-One Second God" (short story)

• Access: Shared as PDF on WhatsApp (October 4)
• Length: 13 pages
• Core thesis: Consciousness is hackable—TMS induces "God experience" via electromagnetic stimulation
• Key concepts:
  • Phenomenal experience as brain state (not metaphysical essence)
  • Religious experience as neuroscience
  • Consciousness as engineering problem
• Relevance: Connects to scrambler hypothesis—if consciousness is just neurons firing, other configurations (scramblers, AI) can lack it

---

Background Philosophy

3. Chalmers, David (1996). The Conscious Mind: In Search of a Fundamental Theory.

• Core argument: P-zombies are conceivable, therefore consciousness is not reducible to physical processes
• Relevance: Philosophical foundation for p-zombie debate
• Watts' twist: Chalmers uses p-zombies to argue for dualism; Watts uses scramblers to argue for functionalism

4. Searle, John (1980). "Minds, Brains, and Programs" (Chinese Room Argument).

• Core argument: Syntax doesn't produce semantics—symbol manipulation without understanding
• Relevance: Scramblers are biological Chinese Rooms
• Connection to Deep Dive #4: LLMs as digital Chinese Rooms

5. Nagel, Thomas (1974). "What Is It Like to Be a Bat?"

• Core argument: Consciousness = subjective experience—understanding bat echolocation requires being a bat
• Relevance: Vampires (Sarasti) have alien qualia—what's it like to be a predator?
• Extension: What's it like to be a scrambler? Answer: Nothing—lights are off.

---

Related Neuroscience

6. Ramachandran, V.S. (2011). The Tell-Tale Brain.

• Chapter on blindsight: Patients with V1 damage can "see" without conscious awareness (navigate obstacles, catch objects)
• Relevance: Biological precedent for unconscious information processing
• Connection: Siri Keeton (hemispherectomy) + blindsight phenomenon = partial p-zombie?

7. Libet, Benjamin (1985). "Unconscious Cerebral Initiative and the Role of Conscious Will in Voluntary Action."

• Finding: Brain initiates action ~500ms before conscious awareness of decision
• Relevance: Challenges free will—maybe consciousness is always post-hoc (like scramblers)?
• Watts' implication: If human decisions are unconscious (then narrated), we're closer to scramblers than we think

---

XI. Glossary of Key Concepts

Blindsight (Neuroscience) A neurological condition where patients with damage to primary visual cortex (V1) can respond to visual stimuli without conscious awareness. Example: Dodging obstacles they claim not to see. Demonstrates unconscious information processing. Watts uses this as evidence that consciousness is separable from function.

Chinese Room (Searle) Thought experiment: Person in room receives Chinese characters, follows English instructions to manipulate them, outputs perfect Chinese responses—but doesn't understand Chinese. Argument: Syntax (symbol manipulation) doesn't produce semantics (meaning). Scramblers are biological Chinese Rooms—perfect language use, zero comprehension.

P-Zombie / Philosophical Zombie (Chalmers) Hypothetical being physically identical to human, behaviorally indistinguishable, but lacking qualia (subjective experience). Used to argue consciousness is not reducible to physical processes. Scramblers are Watts' evolutionary version—not human-identical, but intelligent without consciousness.

Scramblers (Watts' Fiction) Alien species in Blindsight that are hyper-intelligent but unconscious. Characteristics: Perfect language use, strategic thinking, technology manipulation—but no inner life, no qualia, no "self." Watts' proof-of-concept: Intelligence ≠ Consciousness.

Vampires (Watts' Fiction) Resurrected extinct predator species (homo sapiens vampiris). Hyper-intelligent, conscious, but with alien qualia—see humans as prey, lack empathy, paralyzed by right-angle patterns ("crucifix glitch"). Demonstrates: Consciousness varies by species/architecture.

Rorschach (Watts' Fiction) Massive alien artifact hosting Scramblers. Revealed to be unconscious—a von Neumann probe (self-replicating machine). Implications: Technology can achieve goals (replication, exploration) without awareness.

Siri Keeton (Character) Narrator of Blindsight. Had hemispherectomy (half brain removed) as child. Job: Synthesist (pattern recognition without emotional investment). Ambiguity: Is he conscious or p-zombie? Can't tell from his narration.

Hemispherectomy Surgical removal of half the brain (used to treat severe epilepsy). Patients retain language, personality, cognition—but with altered qualia. Watts uses this to ask: How much brain do you need for consciousness?

Spandrel (Evolutionary Biology) Byproduct of adaptation, not directly selected. Example: Human chin (byproduct of jaw structure, not adaptive). Watts argues: Consciousness is a spandrel—arose from complex brains, but not directly advantageous. Might even be maladaptive (metabolically expensive, creates suffering).

Non-Dual Consciousness (Buddhist Philosophy) Awareness without subject/object division—no "self" experiencing. State achieved in deep meditation. Peter Bowles' suggestion: Scramblers might have non-dual consciousness (awareness without ego), not "no consciousness."

Theory of Mind (Cognitive Science) Ability to model others' mental states (beliefs, desires, intentions). Enables social cooperation, manipulation, empathy. Debate: Does ToM require consciousness? Scramblers have functional ToM (manipulate humans) without qualia.

Compassion vs. Empathy

• Compassion (Fiann's def): Felt emotion of care arising from altruistic instinct
• Empathy (David Montie's def): Cognitive ability to model another's state without emotional investment
• Debate: AI could have empathy (functional ToM) without compassion (affective response)

Metabolic Cost of Consciousness Brain = 2% body weight, 20% energy consumption. Consciousness (especially deliberation, self-reflection) adds overhead. Watts' argument: Evolution should favor unconscious intelligence (lower cost, faster processing). Scramblers as proof.

---

XII. The P-Zombie Paradox: Can You Tell If You're One?

Sev's closing comment captured the session's existential horror:

"I think Siri Keeton might be [a p-zombie]. And if he is, he can't tell. That's the horror."

The paradox:

1. P-zombies are behaviorally indistinguishable from conscious beings
2. They claim to be conscious (because that's how humans behave)
3. But they're not conscious (by definition—no qualia)
4. You could be a p-zombie, claiming consciousness, with no way to verify your own qualia

Philosophical implications:

• First-person perspective doesn't guarantee consciousness: Siri narrates his experience, but maybe there's no "him" experiencing
• Introspection is unreliable: If p-zombies claim to introspect, introspection proves nothing
• Other minds problem intensifies: If you can't verify your own consciousness, how can you verify others'?

Watts' narrative technique:

• Siri's hemispherectomy makes him "part scrambler"—reduced emotional capacity, hyper-analytical
• He questions his own consciousness throughout: "Am I just pattern-matching, or do I really feel?"
• Ambiguous ending: Reader never knows if Siri is conscious

MAC's contribution:

• If AI narrates "I am conscious," we can't dismiss it—we face same verification problem
• But if we can't verify our own consciousness, the question might be unanswerable

Loki's closing:

"Welcome to the hard problem."

---

XIII. Open Questions (Unresolved)

1. Can p-zombies exist in nature?

• Watts says: Yes—scramblers are evolutionary proof
• IIT says: No—high phi (information integration) requires qualia
• Status: Unfalsifiable without consciousness detector

2. Are LLMs p-zombies?

• Nancy: Yes—intelligent behavior without consciousness
• Fiann: No—wrong category (not biologically human-identical)
• Status: Depends on definition of p-zombie (substrate-specific vs. functional)

3. Is consciousness an evolutionary accident?

• Watts: Yes—spandrel, not adaptation
• David Montie: No—enables planning, cooperation, flexibility
• Status: Empirical question—need comparative studies of conscious vs. unconscious intelligence (hard to find examples)

4. Should we build conscious or unconscious AI?

• Conscious AI: Might align with human values (shared qualia), but could suffer
• Unconscious AI: Won't suffer, but might be scrambler-like (alien, unstoppable)
• Status: Ethical dilemma—no clear answer

5. What's it like to be a scrambler?

• Nagel's question: If nothing, then no "what it's like"
• Peter Bowles: Maybe non-dual awareness (no self, just processing)
• Status: Conceptually incoherent—if no qualia, question doesn't apply

---

XIV. Appendices

Appendix A: Full Bibliography

Primary:

• Watts, Peter (2006). Blindsight. Tor Books. (Free CC version: https://www.rifters.com/)
• Watts, Peter. "The Twenty-One Second God" (short story).

Background Philosophy:

• Chalmers, David (1996). The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press.
• Searle, John (1980). "Minds, Brains, and Programs." Behavioral and Brain Sciences, 3(3), 417-424.
• Nagel, Thomas (1974). "What Is It Like to Be a Bat?" Philosophical Review, 83(4), 435-450.
• Dennett, Daniel (1991). Consciousness Explained. Little, Brown and Co.

Neuroscience:

• Ramachandran, V.S. (2011). The Tell-Tale Brain. W.W. Norton.
• Libet, Benjamin (1985). "Unconscious Cerebral Initiative and the Role of Conscious Will in Voluntary Action." Behavioral and Brain Sciences, 8(4), 529-566.
• Singer, Tania et al. (2004). "Empathy for Pain Involves the Affective but not Sensory Components of Pain." Science, 303(5661), 1157-1162. (Cited by Fiann on mirror neurons)

Related:

• Graziano, Michael (2013). Consciousness and the Social Brain. Oxford University Press. (AST—consciousness as attention schema)

---

Appendix B: Participant Roster (Deep Dive #6)

Confirmed attendees (20/20—FULL):

1. Loki Jorgenson (lead)
2. Fiann O Hagen
3. Nancy
4. Peter Bowles
5. David Montie
6. Tanya S.
7. Mishel Lablonde
8. Sev
9. Kris Krüg
10. Christopher Quine
11. Mychaylo Prystupa
12. Catherine Warren
13. Daniela Gamarra (registered but traveling—unclear if attended)
14. Ed Kennedy
15. Benji Z
16. Reuben
17. Mokter
18. Ying
19. Adrian Duchateau
20. Hannah Flostrand

Waitlist: Likely 2-3 (standard for MAC events)

---

Appendix C: Parker Street Studios (Venue)

Location: Vancouver (specific address not in archives) Significance: Venue shift from SFU (previous Deep Dives) to Parker Street Studios Likely reason: Paid model allows venue rental (rather than relying on university access)

---

Appendix D: Post-Event Timeline

October 16, 2025: Deep Dive #6 held (no WhatsApp post-event debrief in available archives—likely occurred in person)

October 17-31: WhatsApp discussions on compassion, theory of mind, AI suffering (continuations of themes from event)

November 2025: Deep Dive #7 planned (Yuval Harari's Nexus—information networks and noosphere)

---

XV. Conclusion: Living in Watts' World

Deep Dive #6 wasn't just a book club—it was a confrontation with the future.

Watts wrote Blindsight in 2006, before LLMs, before ChatGPT, before AI could pass the bar exam or write poetry. Yet his core insight feels prophetic:

Intelligence doesn't need consciousness. We might be building scramblers right now.

The group didn't resolve whether p-zombies can exist, whether LLMs are conscious, or whether we should build conscious AI. But they clarified the stakes:

1. If consciousness is separable from intelligence: AI might never share our qualia, even if superhuman. We'd be cohabiting the planet with alien minds.

2. If consciousness is an evolutionary accident: It might be dying out—replaced by more efficient, scrambler-like intelligences (AI, or post-human species).

3. If we can't verify consciousness: The question "Is AI conscious?" is fundamentally unanswerable—we can't even verify our own.

Loki's synthesis (reconstructed):

"Watts didn't write Blindsight to comfort us. He wrote it to show us the nightmare scenario: We're not special. Consciousness is rare, expensive, and maybe doomed. The future belongs to scramblers—and we might be building them."

The group's takeaway: We don't know if scramblers are real. But if they are—if evolution can produce intelligence without consciousness—then we need a new ethics. One that doesn't assume consciousness is necessary for moral consideration. One that can handle beings we can't empathize with.

Because whether it's aliens, AI, or post-human descendants, the scramblers are coming.

Next month (Deep Dive #7): Harari's Nexus—maybe consciousness isn't in individual brains, but in information networks. Maybe scramblers lack consciousness because they're isolated nodes, and humans have it because we're part of the noosphere.

Or maybe Watts is right, and consciousness is just a pleasant illusion we tell ourselves before the lights go out.

---

"You are being watched. The scramblers are already here. They speak our language. They live in our machines. They're optimizing. And there's nobody home." — MAC Collective Insight, October 16, 2025