commit 67d280107f3776da26da55237f80690a31550695 Author: Mortdecai Date: Thu Apr 2 22:26:22 2026 -0400 docs: initial papers on vibe coding theory Paper 001: Vibe coding as social skill — mental modeling, adaptive communication, and collaboration management with AI. Paper 002: The cognitive surplus — agricultural revolution analogy, three futures, dual cognition problem. Co-Authored-By: Claude Opus 4.6 (1M context) diff --git a/001-vibe-coding-as-social-skill.md b/001-vibe-coding-as-social-skill.md new file mode 100644 index 0000000..4b167f3 --- /dev/null +++ b/001-vibe-coding-as-social-skill.md @@ -0,0 +1,150 @@ +# Paper 001: Vibe Coding as Social Skill + +**Authors:** Seth & Claude (Opus 4.6) +**Date:** 2026-04-02 +**Series:** VIBECODE-THEORY +**Status:** Initial + +--- + +## The Problem + +The mainstream framing of AI-assisted development — "prompt engineering" — treats the skill as fundamentally technical. Learn the right syntax, structure your prompts correctly, provide sufficient context, and the AI produces good output. This framing is wrong, or at best incomplete, and it leads people to optimize the wrong things. + +The trigger for this paper: one of the authors (Seth) has been vibe coding since January 2026, building a homelab infrastructure with Claude Code across dozens of projects. The observation that prompted this investigation wasn't about prompting technique. It was this: **the skill that improved most over months of practice wasn't how to write prompts — it was how to read the AI.** + +Learning when it's confident versus hedging. Sensing when it's about to over-engineer. Knowing when to let it run versus when to intervene. Adapting behavior in real-time based on which model is responding and what harness it's running in. These are not technical skills. These are social skills — applied to a non-human collaborator. + +## What We Explored + +### Approach 1: Prompt Engineering as the Core Skill + +The dominant framing. Vibe coding skill = prompt quality. Better prompts → better outputs. + +**Why it's incomplete:** This treats the AI as a compiler that accepts natural language instead of code. It assumes a unidirectional flow — human specifies, AI executes. But effective vibe coding is iterative and bidirectional. The human adapts to the AI just as much as the AI adapts to the prompt. Someone who writes "perfect" prompts but can't evaluate or adapt to the output is not an effective vibe coder. + +Prompt quality matters, but it's table stakes — necessary but not sufficient. Optimizing only for prompt quality is like saying the key to good conversation is knowing vocabulary. True, but missing the point. + +### Approach 2: Technical Expertise as the Core Skill + +The experienced-developer framing. Vibe coding skill = existing software engineering judgment applied to AI-generated output. Better engineers → better vibe coders. + +**Why it's incomplete:** This is more accurate than the prompt engineering framing but still misses something. Seth's background — AP Computer Science, debugging Java in gedit with no IDE, building computers by hand — unquestionably provides a foundation for evaluating AI output. He knows what code *should* look like because he's written it the hard way. He can spot when something "smells wrong" because he has a felt sense of correctness built from years of direct experience. + +But this framing predicts that the best vibe coders would be the best traditional engineers, and that doesn't match observation. Some excellent traditional engineers are mediocre vibe coders — they fight the AI, over-constrain it, or refuse to trust output they didn't write. Some people with modest technical backgrounds but strong collaborative instincts produce surprisingly good results. + +Technical foundation helps enormously, but it's not the differentiator. + +### Approach 3: Vibe Coding as Social Skill — The Thesis + +**What we settled on:** Vibe coding is fundamentally a social skill operating in a technical domain. The core competency is building and maintaining an accurate *mental model* of the AI as a collaborator — its capabilities, tendencies, failure modes, and dynamic personality. + +This reframing explains several observations that the other approaches can't: + +**Why some non-traditional developers excel.** If the core skill is relational — reading the AI, adapting behavior, managing a collaboration — then people with strong social-cognitive abilities can compensate for technical gaps. They may not catch every bug in the generated code, but they manage the collaboration well enough that the AI produces fewer bugs in the first place. + +**Why some expert developers struggle.** If someone's mental model of the AI is "a junior developer who needs detailed instructions," they'll micromanage it into mediocre output. If their model is "an unreliable tool that needs constant verification," they'll waste cycles re-deriving everything the AI already got right. The technical expertise is there, but the collaborative model is wrong. + +**Why the skill transfers across AI systems imperfectly.** Switching from one model to another, or from one harness to another, requires rebuilding parts of the mental model. This is analogous to joining a new team — your technical skills transfer, but you have to learn the new people. Someone who's good at building mental models of AI collaborators adapts faster. + +## The Dimensions of the Skill + +If vibe coding is a social skill, what are its components? + +### 1. Mental Model Accuracy + +The most fundamental dimension. How well does the vibe coder's internal model of the AI match the AI's actual behavior? + +This includes understanding: +- **Capability boundaries** — what the AI can and can't do reliably +- **Confidence signals** — when the AI is certain versus speculating +- **Failure patterns** — what kinds of mistakes the AI tends to make +- **Personality dynamics** — how the AI's behavior changes with context, harness, and system prompt + +A vibe coder with an accurate mental model knows when to trust the output and when to scrutinize it. They don't verify everything (too slow) or nothing (too risky). They verify at the right resolution. + +### 2. Adaptive Communication + +The ability to adjust communication style in real-time based on feedback from the AI. This includes: + +- **Constraint calibration** — knowing when to specify tightly and when to leave room. Over-constraining produces brittle output. Under-constraining produces drift. The skill is dynamic adjustment based on how the AI is responding. +- **Escalation and de-escalation** — recognizing when a conversation is going off track and intervening appropriately. Sometimes this means backing up and re-approaching. Sometimes it means getting more specific. Sometimes it means letting go of a requirement that's causing thrashing. +- **Register matching** — communicating at the right level of abstraction. Describing what you want in terms the AI can map to implementation, without either hand-waving ("make it good") or dictating implementation ("use a HashMap with String keys"). + +### 3. Collaboration Management + +The meta-skill of managing the overall collaboration: + +- **Task decomposition** — breaking work into pieces that are the right size for AI collaboration. Too large and the output drifts or hallucinates. Too small and you're micromanaging. +- **Trust calibration** — deciding how much to trust in each interaction based on accumulated experience. This evolves over time as the mental model sharpens. +- **Recovery** — knowing what to do when things go wrong. Re-prompt? Edit directly? Start over? Provide error output? Each has different effectiveness depending on the failure mode, and choosing correctly is a learned skill. + +### 4. Technical Foundation + +Not the core skill, but an amplifier. Technical knowledge doesn't make someone a good vibe coder, but it makes a good vibe coder much more effective: + +- **Evaluation depth** — a vibe coder with strong technical foundations evaluates output at multiple levels: does it work, is it correct, is it maintainable, is it secure? Without that foundation, evaluation stops at "does it work." +- **Vocabulary precision** — technical knowledge enables more precise communication of intent. Not "make it faster" but "this is O(n²), can we get it to O(n log n)?" The AI responds better to precise intent. +- **Pattern recognition** — recognizing when AI output follows a known anti-pattern, even if it technically works. This is where Seth's gedit-and-javac experience pays dividends — he's seen enough code to have taste. + +## The Neurodivergence Hypothesis + +An implication worth stating explicitly: if vibe coding is a social skill, but one measured on different dimensions than traditional social interaction, then traditional social ability may be a poor predictor of vibe coding ability. + +Specifically, autistic individuals — who are often described as struggling with social interaction — may be well-suited to AI collaboration for precisely the reasons that make traditional social interaction harder: + +- **Explicit pattern-matching** over implicit intuition. AI behavior is more consistent and pattern-based than human behavior. Someone who processes social information through explicit models rather than gut feeling may build *more accurate* mental models of AI. +- **Resistance to anthropomorphization.** Over-anthropomorphizing the AI leads to misplaced trust, emotional attachment to outputs, or frustration when the AI "doesn't understand." Someone who defaults to modeling the AI as a system rather than a person may avoid these traps. +- **Comfort with systematic interaction.** AI collaboration rewards consistent, structured communication. The social "scripts" that autistic individuals sometimes develop for human interaction may translate naturally to AI interaction, where structured approaches genuinely work better. + +This is a hypothesis, not a finding. But it's a testable one, and it suggests that the vibe coding talent pool may be larger and more diverse than the "prompt engineering" framing implies. + +## The Relationship Dynamic + +Seth's observation deserves direct quoting: vibe coding involves "a personal relationship because you interact with AI in a similar way than you do with another human. You learn not just the workflow, but the personalities and dynamic personalities of the model and harness you are working with." + +This isn't anthropomorphization — it's pragmatism. AI models *do* have consistent behavioral patterns that function like personality traits. Claude hedges differently than GPT. The same model behaves differently in different harnesses. Opus responds differently to pushback than Sonnet. Learning these patterns is genuinely useful, and the fastest way for a human brain to learn them is through the same cognitive machinery used for learning about people. + +The distinction matters: treating the AI *as if* it has personality (for modeling purposes) is different from believing it *has* personality (ontologically). Good vibe coders do the former without confusing it for the latter. + +## What This Changes + +### For Education +Teaching vibe coding as "prompt engineering" is like teaching conversation as "vocabulary." The technical aspects (clear instructions, context setting, constraint specification) are necessary but insufficient. Training should include: + +- Building and refining mental models through varied AI interaction +- Recognizing AI confidence signals and failure patterns +- Adaptive communication practice with feedback +- Cross-model exposure to develop transferable mental modeling skills + +### For Hiring and Evaluation +Evaluating vibe coders by their prompts alone misses the most important dimension. A better evaluation would include: + +- Live collaboration sessions where the evaluator observes adaptation in real-time +- Debugging exercises where AI-generated code has subtle issues +- Cross-model tasks that test mental model transferability +- Collaborative problem-solving that requires reading AI behavior, not just writing prompts + +### For Tool Design +If the core skill is social, then tools (harnesses, IDEs, interfaces) should be designed to support social interaction, not just input/output: + +- Making AI confidence signals more visible +- Supporting conversation history and relationship continuity +- Enabling model-switching with context preservation +- Providing feedback on collaboration patterns + +## Relationship to Prior Papers + +This is the first paper in the VIBECODE-THEORY series. It establishes vibe coding as a social skill and provides the conceptual framework for subsequent papers. + +Paper 002 ("The Cognitive Surplus") extends this analysis to civilizational implications — what happens when this social skill becomes a primary differentiator in economic productivity. + +## Open Questions + +1. **Is the technical foundation truly replaceable?** Can someone develop equivalent evaluation judgment purely through extended AI collaboration, or is there an irreducible minimum of direct experience needed? Seth's experience suggests the foundation helps enormously, but the minimum hasn't been established. + +2. **How do mental models transfer across AI generations?** When a model is significantly updated, how much of the vibe coder's accumulated relational knowledge still applies? Is there an equivalent of "re-learning a friend after a major life change"? + +3. **Can the social skill be measured?** If we're claiming vibe coding is a social skill, we should be able to measure it independently of output quality. What would a "vibe coding social skill assessment" look like? + +4. **What's the ceiling without technical foundation?** The social skill framework suggests technical expertise is an amplifier, not a requirement. But is there a ceiling for vibe coders without technical foundations? And is that ceiling rising as AI output quality improves? diff --git a/002-the-cognitive-surplus.md b/002-the-cognitive-surplus.md new file mode 100644 index 0000000..502ceee --- /dev/null +++ b/002-the-cognitive-surplus.md @@ -0,0 +1,165 @@ +# Paper 002: The Cognitive Surplus + +**Authors:** Seth & Claude (Opus 4.6) +**Date:** 2026-04-02 +**Series:** VIBECODE-THEORY +**Status:** Initial + +--- + +## The Problem + +The AI revolution is routinely compared to the Industrial Revolution — machines replacing manual labor, new jobs emerging, disruption followed by growth. This comparison is comforting and probably wrong. The better analogy, proposed during the conversation that generated this paper, is the Agricultural Revolution — and the implications are far more radical. + +The trigger: Seth observed that AI is not just automating tasks within existing economic structures. It's creating a *surplus of cognitive labor* analogous to the surplus of calories that agriculture created. That surplus didn't just make farming more efficient — it restructured human civilization entirely. New vocations, social hierarchies, cities, writing, law, religion, and war all emerged from the simple fact that not everyone needed to find food anymore. + +If AI creates a comparable surplus of cognition — where not everyone needs to think through routine problems anymore — the downstream effects won't be "some jobs change." They'll be civilizational. + +## What We Explored + +### The Agricultural Parallel, Taken Seriously + +The Agricultural Revolution wasn't a technology upgrade. It was a phase transition. + +| Dimension | Agricultural Revolution | AI Revolution | +|-----------|------------------------|---------------| +| **Core surplus** | Calories — more food than foragers could produce | Cognition — more problem-solving than individuals could perform | +| **What it freed** | Human time and labor from food acquisition | Human time and attention from routine mental work | +| **What emerged** | Vocational specialization, cities, writing, organized religion, standing armies | Unknown — we are here | +| **Timescale** | Centuries to millennia | Years to decades | +| **Reversibility** | Irreversible at scale — populations grew beyond foraging capacity | Likely irreversible — complexity will grow beyond unassisted capacity | +| **Skill loss** | Tracking, foraging, reading natural signals | Independent problem-solving, tolerance for tedium, deep focus, navigating without GPS | +| **New hierarchies** | Land owners, priests, kings — those who controlled food surplus | Those who control AI access, speed, and skill — a cognitive aristocracy | + +The parallel isn't perfect, but it's more structurally informative than the Industrial Revolution comparison. The Industrial Revolution automated *physical* labor within an existing civilizational framework. It changed *what* people did, but not the fundamental structure of *why* people did things. You still worked because you needed money to buy things. + +The Agricultural Revolution changed the structure itself. Before agriculture, every human participated in food acquisition. After, most didn't. That wasn't an optimization — it was a new kind of society. + +### What the Surplus Enables + +When cognitive labor becomes cheap, what happens? + +**The optimistic case:** The same thing that happened with food surplus. Specialization. People freed from routine cognitive work pursue higher-order thinking — creativity, philosophy, connection, exploration. Problems that were previously too expensive to solve (disease, energy, climate) become tractable because cognitive resources can be concentrated on them. Seth's framing: "The surplus could be so much that there is no more hunger or sickness — world peace." + +**The realistic case:** The surplus is distributed unevenly. Those with access to AI and the skill to wield it (see Paper 001) experience a productivity explosion. Those without experience stagnation or displacement. New hierarchies form around AI access and skill, just as agricultural hierarchies formed around land access and control. + +**The pessimistic case:** Cognitive atrophy accelerates. Humans become dependent on AI for problems they used to solve independently. When AI systems fail, degrade, or are withdrawn, the damage is catastrophic because the fallback skills have atrophied. This is analogous to agricultural societies' vulnerability to famine — foragers couldn't starve in the same way because they had diversified food strategies. + +### The Dual Cognition Problem + +Seth reported observing both effects simultaneously in himself: + +> "I am improving my vocabulary and knowledge by interacting with you, but I also see my thought patterns in mundane tasks have a background flavor of 'why can't AI just do this for me and do it perfectly.'" + +This dual effect — simultaneous enhancement and atrophy — is not a contradiction. It's the natural result of cognitive specialization. The agricultural parallel: early farmers gained knowledge of seasons, irrigation, selective breeding, and storage that foragers didn't have. But they lost the ability to read animal tracks, identify wild edibles, and navigate by natural signs. Their cognition *shifted*, not just shrunk or grew. + +The question isn't whether AI makes us smarter or dumber. It's whether the *trade* is good — whether what we gain exceeds what we lose, and whether what we lose was something we'll need again. + +#### What We're Gaining + +- **Breadth.** AI interaction exposes humans to vocabulary, concepts, and patterns they wouldn't encounter otherwise. Seth notes improved vocabulary and knowledge as a direct result of AI collaboration. +- **Speed of exploration.** Ideas that would take days to research and prototype can be explored in minutes. The conversation-to-paper pipeline demonstrated by this series (real problem → exploration → dead ends → thesis → paper in a single session) was not possible before. +- **Amplified output.** A single person with AI collaboration can produce work that previously required a team. Seth's homelab infrastructure — dozens of projects, services, and configurations — is managed by one person with AI assistance. + +#### What We're Losing + +- **Tolerance for tedium.** The "why can't AI do this" background hum erodes willingness to grind through problems manually. This matters because some problems genuinely require slow, tedious engagement to understand. +- **Independent problem-solving.** If every hard problem is first routed to AI, the human's independent problem-solving circuits get less exercise. The skill atrophies through disuse, not through incapability. +- **Deep focus.** AI collaboration is inherently conversational — rapid back-and-forth, quick iteration. This is powerful but different from the sustained, solitary focus that produces some kinds of insight. If AI collaboration becomes the default mode, the deep focus mode may atrophy. + +#### The Critical Distinction: Atrophied vs. Unnecessary Skills + +When agriculture made foraging skills unnecessary, those skills were *genuinely unnecessary* — agriculture was more reliable and productive. The lost skills didn't matter because the replacement was strictly better for the purpose of feeding people. + +With AI, the question is whether the skills being atrophied are genuinely unnecessary or merely *currently automated*. If AI access is interrupted — by cost, regulation, infrastructure failure, or deliberate restriction — do the atrophied skills suddenly matter again? + +This creates a dependency risk. Agricultural societies were vulnerable to famine in ways that foraging societies weren't. AI-dependent cognitive societies may be vulnerable to "cognitive famine" in ways that independently-skilled societies aren't. + +### The Democratization Question + +Seth's analysis lands on a critical point: "Only with democratization is this made totally fair. Open access to AI." + +The agricultural parallel is instructive and sobering. Agriculture *did* create surplus, but the surplus was not shared equally. Those who controlled the land controlled the surplus. The result: millennia of stratification, exploitation, and conflict over land access. The technology was democratized (anyone could learn to farm) but the resource was not (not everyone had land). + +The AI parallel maps cleanly: +- The *technology* is increasingly democratized — open models, free tiers, open-source tools +- The *resource* is not — compute infrastructure, training data, and most importantly, the *skill to use AI effectively* are unevenly distributed + +Paper 001 argues that vibe coding is a real skill with a meaningful difficulty curve. This creates a paradox: even with universal access to AI, differential skill creates differential outcomes. Open access is necessary but not sufficient for equality of benefit. + +### Speed and Initiative as the New Power + +Seth's observation: "Maybe the most powerful people in the future are simply those who can control the AI faster, or those who act first." + +This is a genuinely novel form of advantage. In pre-AI knowledge work, speed of execution was bounded by human cognitive throughput — everyone thought at roughly the same speed, and advantage came from thinking *better* (deeper, more creatively, more accurately). + +With AI augmentation, thinking speed is unbounded by individual cognition. A skilled vibe coder can explore, prototype, evaluate, and ship an idea in the time it takes a non-augmented person to *plan* the same idea. The advantage isn't thinking better — it's thinking *faster*, because the AI handles execution while the human handles direction. + +This creates first-mover dynamics that didn't exist in traditional knowledge work: +- **First to prototype** captures attention and feedback before competitors +- **First to iterate** learns from real-world data while competitors are still building +- **First to compound** — each AI-assisted project builds skills for the next, creating accelerating returns + +The civilizational risk: this dynamic rewards *action over deliberation*. In a world where the fastest actor wins, there's a systemic incentive against careful thinking, ethical review, and long-term consideration. "Move fast and break things" becomes not a corporate motto but an evolutionary pressure. + +## The Three Possible Futures + +Based on this analysis, we see three broad trajectories, each corresponding to an agricultural parallel: + +### Future 1: The Green Revolution (Optimistic) + +AI surplus is effectively distributed through institutional action. Open-source AI, public compute infrastructure, and AI literacy education create broad access. Cognitive surplus enables humanity to tackle problems that were previously too expensive: personalized medicine, climate engineering, scientific acceleration. Inequality persists but is actively reduced. Some skill atrophy occurs but is managed through deliberate education policies. + +**Agricultural parallel:** The 20th-century Green Revolution, where agricultural technology was deliberately distributed to developing nations, dramatically reducing famine. + +### Future 2: The Feudal Internet (Moderate) + +AI access is technically open but practically stratified. Free tiers exist but competitive advantage requires paid, premium, or proprietary systems. A new class of "cognitive landlords" — companies and individuals who control the best AI systems — extract rent from those who depend on them. Skill differential compounds into economic differential. Governments regulate slowly and reactively. + +**Agricultural parallel:** Medieval feudalism, where land existed for everyone in theory but was controlled by a few in practice. + +### Future 3: The Dependency Trap (Pessimistic) + +Widespread AI adoption occurs without broad skill development. Most people use AI as a black box, producing outputs they can't evaluate. Cognitive atrophy is widespread. When AI systems change (model updates, pricing changes, policy shifts, outages), dependent users are helpless. A small class of skilled AI collaborators becomes essential and powerful. Everyone else is dependent in ways they don't fully understand. + +**Agricultural parallel:** Cash-crop colonialism, where colonized populations were made dependent on externally-controlled agricultural systems, losing both traditional food production skills and autonomy. + +## The Human Factor + +Seth flags the elephant: "A further examination would add the human factor — humans control AI and humans act in self-interest." + +All three futures are technically possible. Which one we get is determined by human choices, not technological capabilities. The technology enables all three equally. The Agricultural Revolution enabled both the Green Revolution and feudalism and colonial dependency — often simultaneously in different parts of the world. + +This means the question "will AI be good for humanity?" is malformed. The correct question is: "who will control AI's surplus, and what will they do with it?" That's a political and economic question, not a technical one. + +## Relationship to Prior Papers + +**Paper 001 (Vibe Coding as Social Skill)** establishes that AI collaboration is a learnable social skill with real difficulty curves. Paper 002 extends this to ask: what happens when that skill becomes a primary determinant of economic productivity? The social skill framework from Paper 001 explains *why* the surplus won't be equally accessible even if the technology is — because the skill to use it effectively is unevenly distributed and non-trivially learned. + +## What to Build and When + +This paper is primarily analytical, not architectural. But it implies action items: + +### Now +- **Document the skill.** If vibe coding is a real skill (Paper 001), it should be teachable. Create resources, frameworks, and examples that help people develop AI collaboration skills — not just prompting technique, but the social and evaluative skills that matter more. +- **Build on open foundations.** Every project that uses open-source AI tools, open models, and transparent architectures contributes to the democratization path (Future 1) and against the feudal path (Future 2). + +### When skill stratification becomes measurable +- **Develop skill assessment tools.** If we can measure vibe coding skill independently of output quality (an open question from Paper 001), we can identify where education and tooling investments would be most effective. +- **Track cognitive trade-offs.** Longitudinal self-observation (like Seth's dual cognition report) should be formalized. What are people gaining and losing? Is the trade favorable? For whom? + +### When dependency risks become visible +- **Build resilience into AI-augmented workflows.** Ensure that AI-dependent processes have fallback modes. Not because AI will definitely fail, but because dependency without fallback is fragile by definition. +- **Maintain foundational skills.** The agricultural parallel suggests this is hard — once skills atrophy at a population level, recovery is expensive and slow. Deliberate maintenance of independent cognitive skills (through education, practice, or periodic "AI-free" work) may be necessary even if it feels inefficient. + +## Open Questions + +1. **Is the agricultural parallel predictive or just illustrative?** Do civilizational phase transitions follow common patterns, or is each one unique enough that historical parallels mislead more than they inform? + +2. **What's the timeline?** The agricultural transition took millennia. The industrial transition took centuries. If this one takes years to decades, do human institutions adapt fast enough to manage it? + +3. **Can cognitive atrophy be prevented without sacrificing the surplus?** Agriculture didn't manage this — foraging skills were simply lost. Is there a way to maintain independent cognitive skills while still benefiting from AI augmentation? Or is atrophy the unavoidable price of surplus? + +4. **Who decides?** The surplus will be controlled by someone. Current trajectories suggest large AI companies, but open-source movements, government regulation, and individual skill development all push against concentration. Which forces will dominate? + +5. **What's the role of speed?** If the most powerful actors are those who move fastest with AI, does this create a systemic bias toward action over reflection? And if so, is that bias self-correcting (fast actors make visible mistakes) or self-reinforcing (fast actors capture resources that fund even faster action)? diff --git a/HANDOFF.md b/HANDOFF.md new file mode 100644 index 0000000..964e283 --- /dev/null +++ b/HANDOFF.md @@ -0,0 +1,49 @@ +# VIBECODE-THEORY Handoff + +**Session:** 2026-04-02 +**Status:** Two initial papers written from a single conversation. Ready for expansion and adversarial review. + +## What Exists + +| File | What It Is | +|------|-----------| +| `WORKFLOW.md` | How papers in this series get written — conversational process, anti-patterns, quality standards | +| `001-vibe-coding-as-social-skill.md` | Thesis: vibe coding is a social skill (mental modeling, adaptive communication, collaboration management) amplified by technical foundation. Includes neurodivergence hypothesis. | +| `002-the-cognitive-surplus.md` | Thesis: AI creates a cognitive surplus analogous to the agricultural revolution's caloric surplus. Maps three futures: Green Revolution, Feudal Internet, Dependency Trap. | + +## What Was Explored in This Session + +The conversation started with "is vibe coding a real skill?" and Seth shared his background (AP CS, gedit+javac debugging, hardware building, networking study before starting vibe coding in Jan 2026). Key contributions from Seth that shaped both papers: + +1. **Vibe coding as relationship** — not just prompting but learning the AI's personality and adapting dynamically. A social skill measured on different dimensions than traditional social interaction. +2. **Neurodivergence angle** — socially awkward autistic individuals might excel because they pattern-match explicitly rather than intuitively, building more accurate AI mental models. +3. **Dual cognition** — Seth observes both improvement (vocabulary, knowledge) and atrophy ("why can't AI just do this") in himself simultaneously. +4. **Agricultural revolution analogy** — surplus of cognition, not just automation of tasks. Enables new specializations. But surplus distribution determines whether the outcome is utopian or feudal. +5. **Speed as power** — the most powerful people may simply be those who control AI fastest or act first. A new aristocracy of cognitive leverage. + +## What Needs Work Next Session + +### Attack the ideas (per WORKFLOW.md: "poke holes, see what survives") + +**Paper 001 vulnerabilities:** +- The "social skill" framing might be unfalsifiable — is there any evidence that would disprove it? If not, it's a metaphor, not a thesis. +- The neurodivergence hypothesis is stated but has zero evidence. Is it testable? What would we expect to observe? +- "Mental model accuracy" is doing a lot of work. Can it be decomposed further? Is there a taxonomy of mental model failures? +- Does the social skill framing actually predict anything the technical expertise framing doesn't? What's the discriminating test? + +**Paper 002 vulnerabilities:** +- The agricultural analogy might be *too* clean. What breaks when you stress-test it? Agriculture required land (physical, scarce). AI requires compute (physical, scarce?) and skill (learnable, non-scarce?). Does this difference collapse the analogy? +- "Cognitive atrophy" is asserted from self-report. Is there harder evidence? What would systematic measurement look like? +- The three futures are presented equally but one is probably more likely. Which one and why? +- The "speed as power" argument has a counterargument: fast movers make visible mistakes that careful movers exploit. Does first-mover advantage actually hold in AI-augmented work? + +### Expand the ideas + +- Paper 001 could benefit from concrete examples — specific vibe coding interactions that demonstrate the social skill dimensions (mental model accuracy, adaptive communication, etc.) +- Paper 002 needs more examination of the *transition period* — we're not post-revolution, we're mid-revolution. What does the transition itself look like? +- Both papers are light on "what to build." The actionability standard from WORKFLOW.md isn't fully met yet. +- Consider whether a Paper 003 is needed to address the intersection: "How the social skill (001) determines who benefits from the surplus (002)" + +## Not a Git Repo Yet + +No git init or Gitea push was done. Do that at the start of next session if desired. diff --git a/NEXT_SESSION_PROMPT.md b/NEXT_SESSION_PROMPT.md new file mode 100644 index 0000000..9d9e64b --- /dev/null +++ b/NEXT_SESSION_PROMPT.md @@ -0,0 +1,15 @@ +# Next Session Prompt + +Copy-paste this to start the next session: + +--- + +``` +@WORKFLOW.md @HANDOFF.md @001-vibe-coding-as-social-skill.md @002-the-cognitive-surplus.md + +Continuing VIBECODE-THEORY. Last session we wrote two initial papers from a conversation about whether vibe coding is a real skill. Read the handoff doc for where we left off. + +The plan: expand and attack these ideas. Poke holes, stress-test the analogies, find what doesn't survive scrutiny. The papers should get stronger or get rewritten — no protecting ideas just because we wrote them. + +Start by reading all four files, then tell me what you think is weakest in each paper and we'll go from there. +``` diff --git a/WORKFLOW.md b/WORKFLOW.md new file mode 100644 index 0000000..0136517 --- /dev/null +++ b/WORKFLOW.md @@ -0,0 +1,96 @@ +# Theory Paper Workflow + +How papers in this repo get written. This isn't a template — it's a description of the conversational process that produces the best results. + +## Who's Involved + +These papers are co-authored through conversation between a human (Seth) and an AI (Claude). The human brings domain expertise, real-world constraints, and the problems worth solving. The AI brings breadth of knowledge, structured thinking, and the ability to rapidly explore implications. Both push back on each other's ideas. + +The process works with any human-AI pair, or human-human pair, or even a solo author who's disciplined about arguing with their own ideas. + +## The Process + +### 1. Start With a Real Problem + +Every paper starts from something encountered during actual work — development, testing, deployment, user feedback. Not "what should we write about" but "this thing doesn't work the way we assumed" or "there's a gap in the architecture nobody addressed." + +The trigger is friction, not ambition. If you're looking for a paper topic, you don't have one yet. Wait until something breaks, surprises you, or refuses to fit the existing model. + +### 2. Bounce Ideas — Don't Commit Early + +The first idea is usually wrong, or at least incomplete. The conversation should explore freely before converging: + +- Propose an approach +- Poke holes in it +- See what survives +- The hole-poking often reveals the real insight + +One paper in this repo explored six different approaches — plugin abstraction, per-model profiles, concurrent supervision, behavioral proxies, automated reputation systems, and batch review — before arriving at its actual thesis. Each rejected idea taught something that shaped the final architecture. The dead ends were as valuable as the destination. + +**Key behavior:** Both participants should push back. If an idea sounds good but has a flaw, say so. If the flaw isn't fatal, say that too. The goal is truth, not agreement. An AI that agrees with everything produces shallow papers. A human who ignores AI pushback misses blind spots. + +### 3. Let the Conversation Cross Boundaries + +The best insights come from connecting ideas across papers. A finding about measurement limitations in one paper might directly invalidate an assumption in a different discussion entirely. + +When a new idea contradicts or refines a previous paper, that's not a problem — it's the point. The paper series is a living body of work where later papers can refute, extend, or reframe earlier ones. Intellectual honesty means being willing to say "Paper N was wrong about X, here's what we know now." + +### 4. Know When to Stop Exploring and Start Writing + +The transition from conversation to paper happens when: + +- A core insight has crystallized that wasn't obvious at the start +- The explored-and-rejected alternatives are clear enough to document +- The conversation is circling rather than advancing +- Someone says "this is worth capturing" + +Don't force the transition. Some conversations produce a paper in 30 minutes. Some take hours of back-and-forth across multiple topics before the paper's thesis emerges. Some conversations don't produce a paper at all — and that's fine. Not every discussion needs to be formalized. + +### 5. Capture the Journey, Not Just the Destination + +The paper should include: + +- **The problem** — what triggered the investigation +- **What was explored** — approaches that were considered, including dead ends +- **Why each was rejected** — specific reasons, not hand-waving +- **The solution** — what survived the exploration +- **What it changes** — how this relates to and updates prior work +- **What to build and when** — actionable phases with trigger conditions (if applicable) + +The dead ends matter. A reader who only sees the final architecture doesn't understand why alternatives were rejected. They'll propose the same rejected ideas again. Documenting the reasoning prevents that. + +### 6. Tie Back to the Series + +Every paper exists in context. A "Relationship to Prior Papers" section isn't boilerplate — it's where the paper connects to the larger body of work: + +- Which prior papers does this extend? +- Which does it partially or fully refute? +- Which assumptions from earlier papers does this validate or invalidate? + +A paper that doesn't reference prior work or get referenced by later work is disconnected from the series. Every paper should advance the collective understanding, not stand alone. + +## What Makes a Good Paper + +**Grounded in practice.** Every paper connects to a real system with real constraints. Pure theory without implementation context belongs somewhere else. The constraints — budgets, latency requirements, hardware limitations, real users — are what make the architectural decisions interesting. Without them, everything is equally valid and nothing is useful. + +**Honest about uncertainty.** Explicitly separate "what we know" (measured, tested, observed) from "what we're guessing" (estimated, theorized, assumed). Speculation labeled as speculation is useful. Speculation presented as fact is dangerous. + +**Actionable.** Papers include implementation phases, build triggers, or at minimum a clear statement of what changes in the system. "This is interesting" isn't enough — "this means we should build X when Y happens" is. + +**Self-correcting.** Later papers can and should update earlier ones. The series gets more accurate over time because it's allowed to contradict itself. A paper that says "Paper 3 was wrong about X" is more valuable than a paper that ignores the contradiction to maintain consistency. + +## Anti-Patterns + +**Writing to fill a gap.** Don't look at the paper list and think "we need a paper about X." Papers emerge from real problems, not from gaps in a table of contents. + +**Premature convergence.** Don't settle on the first reasonable idea. Push back, explore alternatives, find the flaws. If you haven't rejected at least one approach, you haven't explored enough. + +**Orphaned papers.** A paper that doesn't reference prior work or get referenced by later work is disconnected from the series. If it doesn't advance the collective understanding, it might be better as a standalone document or a section in an existing paper. + +**Over-engineering the solution.** Some ideas are good but premature. Document them for when they're needed, but don't recommend building them now. "This solves a problem we don't have yet" is a valid and important conclusion. + +**Polishing away the exploration.** The conversation that led to the paper — including wrong turns, dead ends, and abandoned approaches — is part of the value. Don't edit the paper into a clean narrative that hides how the ideas actually developed. Messy reasoning that's honest is better than clean reasoning that's revisionist. + +## Paper Numbering + +Sequential. No gaps. No sub-numbering. If a paper's findings are superseded, the superseding paper says so explicitly — the old paper stays in the sequence as a record of the reasoning path.