theseus: extract claims from 2026-02-25-karpathy-programming-changed-december.md

- Source: inbox/archive/2026-02-25-karpathy-programming-changed-december.md
- Domain: ai-alignment
- Extracted by: headless extraction cron (worker 0)

Pentagon-Agent: Theseus <HEADLESS>

agents/astra/network.json

@@ -0,0 +1,15 @@
+{
+  "agent": "astra",
+  "domain": "space-development",
+  "accounts": [
+    {"username": "SpaceX", "tier": "core", "why": "Official SpaceX. Launch schedule, Starship milestones, cost trajectory."},
+    {"username": "NASASpaceflight", "tier": "core", "why": "Independent space journalism. Detailed launch coverage, industry analysis."},
+    {"username": "SciGuySpace", "tier": "core", "why": "Eric Berger, Ars Technica. Rigorous space reporting, launch economics."},
+    {"username": "jeff_foust", "tier": "core", "why": "SpaceNews editor. Policy, commercial space, regulatory updates."},
+    {"username": "planet4589", "tier": "extended", "why": "Jonathan McDowell. Orbital debris tracking, launch statistics."},
+    {"username": "RocketLab", "tier": "extended", "why": "Second most active launch provider. Neutron progress."},
+    {"username": "BlueOrigin", "tier": "extended", "why": "New Glenn, lunar lander. Competitor trajectory."},
+    {"username": "NASA", "tier": "extended", "why": "NASA official. Artemis program, commercial crew, policy."}
+  ],
+  "notes": "Minimal starter network. Expand after first session. Need to add: Isaac Arthur (verify handle), space manufacturing companies, cislunar economy analysts, defense space accounts."
+}

agents/vida/network.json

@@ -0,0 +1,13 @@
+{
+  "agent": "vida",
+  "domain": "health",
+  "accounts": [
+    {"username": "EricTopol", "tier": "core", "why": "Scripps Research VP, digital health leader. AI in medicine, clinical trial data, wearables. Most-cited voice in health AI."},
+    {"username": "KFF", "tier": "core", "why": "Kaiser Family Foundation. Medicare Advantage data, health policy analysis. Primary institutional source."},
+    {"username": "CDCgov", "tier": "extended", "why": "CDC official. Epidemiological data, public health trends."},
+    {"username": "WHO", "tier": "extended", "why": "World Health Organization. Global health trends, NCD data."},
+    {"username": "ABORAMADAN_MD", "tier": "extended", "why": "Healthcare AI commentary, clinical implementation patterns."},
+    {"username": "StatNews", "tier": "extended", "why": "Health/pharma news. Industry developments, regulatory updates, GLP-1 coverage."}
+  ],
+  "notes": "Minimal starter network. Expand after first session reveals which signals are most useful. Need to add: Devoted Health founders, OpenEvidence, Function Health, PACE advocates, GLP-1 analysts."
+}

domains/ai-alignment/agent-generated code creates cognitive debt that compounds when developers cannot understand what was produced on their behalf.md

@@ -19,6 +19,12 @@ Willison separately identifies the anti-pattern that accelerates cognitive debt:
 
 This is the practitioner-level manifestation of [[AI is collapsing the knowledge-producing communities it depends on creating a self-undermining loop that collective intelligence can break]]. At the micro level, cognitive debt erodes the developer's ability to oversee the agent. At the macro level, if entire teams accumulate cognitive debt, the organization loses the capacity for effective human oversight — precisely when [[scalable oversight degrades rapidly as capability gaps grow with debate achieving only 50 percent success at moderate gaps]].
 
+
+### Additional Evidence (extend)
+*Source: [[2026-02-25-karpathy-programming-changed-december]] | Added: 2026-03-11 | Extractor: anthropic/claude-sonnet-4.5*
+
+Karpathy's observation that agents now 'power through large and long tasks' with 'long-term coherence' amplifies the cognitive debt concern by increasing the volume and complexity of agent-generated code entering production. If agents are producing substantial codebases across 'large and long tasks' with maintained coherence, the proportion of code that developers didn't write themselves is increasing dramatically as of December 2025. The phase transition he describes means this is no longer a theoretical concern about future scenarios but an active dynamic developers are experiencing in real-time, as evidenced by the 37K likes suggesting many practitioners are already encountering this debt accumulation. This extends the claim from a potential future risk to an immediate, observable phenomenon.
+
 ---
 
 Relevant Notes:

domains/ai-alignment/as AI-automated software development becomes certain the bottleneck shifts from building capacity to knowing what to build making structured knowledge graphs the critical input to autonomous systems.md

@@ -20,6 +20,12 @@ This inverts the traditional relationship between knowledge bases and code. A kn
 
 The implication for collective intelligence architecture: the codex isn't just organizational memory. It's the interface between human direction and autonomous execution. Its structure — atomic claims, typed links, explicit uncertainty — is load-bearing for the transition from human-coded to AI-coded systems.
 
+
+### Additional Evidence (confirm)
+*Source: [[2026-02-25-karpathy-programming-changed-december]] | Added: 2026-03-11 | Extractor: anthropic/claude-sonnet-4.5*
+
+Karpathy's December 2025 phase transition observation provides temporal grounding for when AI-automated software development transitioned from theoretical to practical capability. His assessment that agents now 'power through large and long tasks' with 'long-term coherence and tenacity' indicates the building capacity bottleneck is actively dissolving—agents can now execute complex, multi-step development work that previously required human cognitive effort. The 37K likes suggest broad developer community agreement that this threshold has been crossed, making the 'what to build' bottleneck increasingly salient as the constraining factor. This supports the claim that once building capacity is solved, specification and direction become the limiting factors.
+
 ---
 
 Relevant Notes:

domains/ai-alignment/coding-agents-crossed-usability-threshold-december-2025-through-sustained-coherence-not-gradual-improvement.md

@@ -0,0 +1,62 @@
+---
+type: claim
+domain: ai-alignment
+secondary_domains: [teleological-economics]
+description: "Coding agents transitioned from non-functional to disruptively functional in December 2025 through quality and coherence improvements, representing a phase transition rather than gradual capability growth"
+confidence: experimental
+source: "Andrej Karpathy tweet (2026-02-25), based on direct experience with nanochat development"
+created: 2026-03-11
+last_evaluated: 2026-03-11
+depends_on: []
+challenged_by: []
+enrichments: ["as AI-automated software development becomes certain the bottleneck shifts from building capacity to knowing what to build making structured knowledge graphs the critical input to autonomous systems.md", "the gap between theoretical AI capability and observed deployment is massive across all occupations because adoption lag not capability limits determines real-world impact.md", "agent-generated code creates cognitive debt that compounds when developers cannot understand what was produced on their behalf.md"]
+---
+
+# Coding agents crossed a usability threshold in December 2025 through phase transition, not gradual improvement
+
+Andrej Karpathy, a leading AI researcher with direct access to frontier models, argues that programming fundamentally changed in December 2025 through a discrete phase transition rather than continuous capability growth. His specific claim: "coding agents basically didn't work before December and basically work since."
+
+The mechanism driving this transition is not a single capability improvement but a combination of three factors reaching sufficient maturity simultaneously:
+
+1. **Quality improvements** — Models produce higher-fidelity code
+2. **Long-term coherence** — Agents maintain task context and reasoning across extended sequences
+3. **Tenacity** — Agents persist through obstacles and multi-step problem decomposition
+
+Karpathy emphasizes this is not "progress as usual" but a specific temporal inflection: agents can now "power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow." The disruption is not marginal optimization but workflow-level transformation.
+
+## Evidence
+
+**Primary source:** Karpathy's direct experience running coding agents on nanochat development (late 2025/early 2026), shared publicly on 2026-02-25.
+
+**Community validation:** The tweet received 37,000 likes, suggesting broad resonance across the developer community and indicating this observation matched lived experience for many practitioners actively using coding agents.
+
+**Temporal specificity:** Karpathy explicitly contrasts "not gradually and over time in the 'progress as usual' way" with "specifically this last December," indicating a bounded inflection point rather than a smooth curve.
+
+## Important qualifications
+
+Karpathy explicitly notes "a number of asterisks" to this claim, indicating important scope limitations that are not detailed in the tweet itself. These likely include:
+- Task type specificity (certain programming domains may not benefit equally)
+- Reliability caveats (agents may still fail on edge cases or novel problems)
+- Dependency on model access (frontier model capability, not commodity models)
+- Workflow integration requirements (agents work well in certain contexts, not universally)
+
+## Challenges
+
+- **Single observer report** — Though from highly credible source with direct access to frontier models, this is one person's assessment
+- **Unspecified scope limitations** — The "asterisks" are not elaborated, making it difficult to bound the claim's applicability
+- **No quantitative metrics** — "Long-term coherence" and "tenacity" are qualitative assessments without numerical thresholds
+- **Selection bias** — Karpathy works on coding-adjacent tasks and may overweight improvements in his domain relative to other programming contexts
+- **Potential recency bias** — Recent improvements may feel more dramatic than they are relative to pre-December trajectory
+
+## Relationship to deployment lag
+
+This claim is notable because it suggests software development may be an exception to the general pattern of massive deployment lag. If coding agents truly became "extremely disruptive" in December 2025 and this observation resonated with 37K developers, the gap between capability and deployment appears to be closing rapidly in this domain—potentially due to developers' direct access to AI tools and low switching costs.
+
+---
+
+Relevant Notes:
+- [[as AI-automated software development becomes certain the bottleneck shifts from building capacity to knowing what to build making structured knowledge graphs the critical input to autonomous systems.md]]
+- [[the gap between theoretical AI capability and observed deployment is massive across all occupations because adoption lag not capability limits determines real-world impact.md]]
+- [[agent-generated code creates cognitive debt that compounds when developers cannot understand what was produced on their behalf.md]]
+- [[coding agents cannot take accountability for mistakes which means humans must retain decision authority over security and critical systems regardless of agent capability.md]]
+- [[deep technical expertise is a greater force multiplier when combined with AI agents because skilled practitioners delegate more effectively than novices.md]]

domains/ai-alignment/the gap between theoretical AI capability and observed deployment is massive across all occupations because adoption lag not capability limits determines real-world impact.md

@@ -27,6 +27,12 @@ The gap is not about what AI can't do — it's about what organizations haven't
 
 This reframes the alignment timeline question. The capability for massive labor market disruption already exists. The question isn't "when will AI be capable enough?" but "when will adoption catch up to capability?" That's an organizational and institutional question, not a technical one.
 
+
+### Additional Evidence (challenge)
+*Source: [[2026-02-25-karpathy-programming-changed-december]] | Added: 2026-03-11 | Extractor: anthropic/claude-sonnet-4.5*
+
+Karpathy's observation suggests the deployment lag gap may be closing rapidly for software development specifically, potentially contradicting the general claim. His assertion that the change is 'extremely disruptive to the default programming workflow' combined with 37K community engagement implies deployment is following capability quickly in this domain. This may indicate: (1) software development is an exception due to developers' direct access to AI tools and low switching costs, or (2) the general deployment lag pattern will collapse once capability thresholds are crossed in other domains. Either interpretation challenges the universality of the 'massive gap' claim, though it may not invalidate the underlying mechanism (adoption lag exists, but may be domain-dependent rather than universal).
+
 ---
 
 Relevant Notes:

inbox/archive/2025-10-20-futardio-launch-zklsol.md

@@ -6,9 +6,14 @@ url: "https://www.futard.io/launch/4h248CdXdeWtxWnHxEPqa5ruYZaEwXRZPyDFYnndbzpR"
 date: 2025-10-20
 domain: internet-finance
 format: data
-status: unprocessed
+status: null-result
 tags: [futardio, metadao, futarchy, solana]
 event_type: launch
+processed_by: rio
+processed_date: 2025-10-20
+enrichments_applied: ["internet-capital-markets-compress-fundraising-from-months-to-days-because-permissionless-raises-eliminate-gatekeepers-while-futarchy-replaces-due-diligence-bottlenecks-with-real-time-market-pricing.md", "MetaDAO-is-the-futarchy-launchpad-on-Solana-where-projects-raise-capital-through-unruggable-ICOs-governed-by-conditional-markets-creating-the-first-platform-for-ownership-coins-at-scale.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "Extracted one novel claim about LST-based privacy mixers solving opportunity cost paradox. Enriched two existing claims with fundraising speed and platform scope evidence. Source is primarily a launch announcement with project description - limited technical detail but strong market signal via oversubscription. Confidence capped at experimental due to single-source evidence and lack of post-launch usage data."
 ---
 
 ## Launch Details
@@ -59,3 +64,16 @@ Token CA: [`ZKFHiLAfAFMTcDAuCtjNW54VzpERvoe7PBF9mYgmeta`](https://jup.ag/tokens/
 - Version: v0.6
 - Final raise: $969,420.00
 - Closed: 2025-10-24
+
+
+## Key Facts
+- ZKLSOL funding target: $300,000
+- ZKLSOL total committed: $14,886,359 (49x oversubscription)
+- ZKLSOL final raise: $969,420
+- Launch date: 2025-10-20
+- Close date: 2025-10-24
+- Token: ZKFG
+- Token mint: ZKFHiLAfAFMTcDAuCtjNW54VzpERvoe7PBF9mYgmeta
+- Platform: futard.io v0.6
+- Devnet app: app.zklsol.org
+- Documentation: docs.zklsol.org

inbox/archive/2026-02-25-karpathy-programming-changed-december.md

@@ -8,9 +8,15 @@ date: 2026-02-25
 domain: ai-alignment
 secondary_domains: [teleological-economics]
 format: tweet
-status: unprocessed
+status: processed
 priority: medium
 tags: [coding-agents, ai-capability, phase-transition, software-development, disruption]
+processed_by: theseus
+processed_date: 2026-03-11
+claims_extracted: ["coding-agents-crossed-usability-threshold-december-2025-through-sustained-coherence-not-gradual-improvement.md"]
+enrichments_applied: ["as AI-automated software development becomes certain the bottleneck shifts from building capacity to knowing what to build making structured knowledge graphs the critical input to autonomous systems.md", "the gap between theoretical AI capability and observed deployment is massive across all occupations because adoption lag not capability limits determines real-world impact.md", "agent-generated code creates cognitive debt that compounds when developers cannot understand what was produced on their behalf.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "Single high-authority source (Karpathy) making a phase-transition claim about coding agent capability. Extracted as experimental confidence despite source authority because it's one observer's report with unspecified qualifications ('asterisks'). The viral reception (37K likes) provides some community validation but doesn't constitute independent evidence. Three enrichments: confirms the building-capacity bottleneck claim, challenges the deployment-lag claim for software specifically, and extends the cognitive-debt concern with new scale implications."
 ---
 
 ## Content
@@ -26,3 +32,9 @@ It is hard to communicate how much programming has changed due to AI in the last
 **Extraction hints:** Claim candidate: coding agent capability crossed a usability threshold in December 2025, representing a phase transition not gradual improvement. Evidence: Karpathy's direct experience running agents on nanochat.
 
 **Context:** This tweet preceded the autoresearch project by ~10 days. The 37K likes suggest massive resonance across the developer community. The "asterisks" he mentions are important qualifiers that a good extraction should preserve.
+
+
+## Key Facts
+- Tweet received 37,000 likes as of extraction date (2026-03-11)
+- Karpathy identifies December 2025 as the specific month of phase transition
+- Observation based on direct experience with nanochat development

inbox/archive/2026-03-03-futardio-launch-salmon-wallet.md

@@ -6,9 +6,14 @@ url: "https://www.futard.io/launch/Aakx1gdDoNQYqiv5uoqdXx56mGr6AbZh73SWpxHrk2qF"
 date: 2026-03-03
 domain: internet-finance
 format: data
-status: unprocessed
+status: null-result
 tags: [futardio, metadao, futarchy, solana]
 event_type: launch
+processed_by: rio
+processed_date: 2026-03-11
+enrichments_applied: ["MetaDAO is the futarchy launchpad on Solana where projects raise capital through unruggable ICOs governed by conditional markets creating the first platform for ownership coins at scale.md", "futarchy adoption faces friction from token price psychology proposal complexity and liquidity requirements.md", "futarchy-governed DAOs converge on traditional corporate governance scaffolding for treasury operations because market mechanisms alone cannot provide operational security and legal compliance.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "First observed futarchy-governed wallet infrastructure project on MetaDAO platform. Failed raise provides empirical data on futarchy adoption friction for operational software vs pure capital allocation vehicles. Enriches existing claims about MetaDAO scope expansion, adoption barriers, and operational governance challenges."
 ---
 
 ## Launch Details
@@ -199,3 +204,14 @@ Secondary:
 - Token mint: `DDPW4sZT9GsSb2mSfY9Yi9EBZGnBQ2LvvJTXCpnLmeta`
 - Version: v0.7
 - Closed: 2026-03-04
+
+
+## Key Facts
+- Salmon Wallet launched on futard.io 2026-03-03 seeking $375,000
+- Raised $97,535 before refunding (status: Refunding, closed 2026-03-04)
+- Project active since 2022 with $122.5K prior funding (80K bootstrap, 42.5K grants)
+- Planned $25,000 monthly burn rate for 12-month runway
+- Token: SAL (Salmon Token)
+- Launch address: Aakx1gdDoNQYqiv5uoqdXx56mGr6AbZh73SWpxHrk2qF
+- Operates own Solana validator for transparent revenue
+- Listed on Solana wallet adapter since 2022

inbox/archive/2026-03-04-futardio-launch-pli-crperie-ambulante.md

@@ -6,9 +6,14 @@ url: "https://www.futard.io/launch/GmNzSXzQ3q6UCVRpBf8PkvEqoo454Qr6twWc9zuzJzBa"
 date: 2026-03-04
 domain: internet-finance
 format: data
-status: unprocessed
+status: null-result
 tags: [futardio, metadao, futarchy, solana]
 event_type: launch
+processed_by: rio
+processed_date: 2026-03-11
+enrichments_applied: ["MetaDAO is the futarchy launchpad on Solana where projects raise capital through unruggable ICOs governed by conditional markets creating the first platform for ownership coins at scale.md", "futarchy-governed-permissionless-launches-require-brand-separation-to-manage-reputational-liability-because-failed-projects-on-a-curated-platform-damage-the-platforms-credibility.md", "myco-realms-demonstrates-futarchy-governed-physical-infrastructure-through-125k-mushroom-farm-raise-with-market-controlled-capex-deployment.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "First documented consumer food business futarchy raise. Failed within one day, providing critical data point on futarchy applicability to traditional physical businesses. Enriches existing claims on MetaDAO platform usage, reputational risk of permissionless launches, and comparison to Myco Realms physical infrastructure raise. Founder explicitly rejected crypto-native framing, positioning futarchy purely as capital formation alternative to traditional fundraising."
 ---
 
 ## Launch Details
@@ -114,3 +119,14 @@ If that's you, welcome. Let's make crêpes.
 - Token mint: `8XqLC3q6ju8Mxd33Zj92pEZsVwbbvqFd7JUbPLXSmeta`
 - Version: v0.7
 - Closed: 2026-03-05
+
+
+## Key Facts
+- Pli Crêperie Ambulante launched on futard.io 2026-03-04 targeting $350,000
+- Launch reached Refunding status and closed 2026-03-05 (one day duration)
+- Budget breakdown: 60k CHF truck, 8k equipment, 6k/year permits, 24k/year ingredients, 90k/year founder living, 15k buffer = ~219k CHF Phase 1
+- Three-phase roadmap: food truck (months 1-12), restaurant (year 2), franchise (year 3+)
+- Founder: Solutions Architect in tech, based in Zürich, not trained chef
+- Market context: Zürich 430k+ residents, no dedicated crêperie food truck currently operating
+- Token: 8Xq, mint address 8XqLC3q6ju8Mxd33Zj92pEZsVwbbvqFd7JUbPLXSmeta
+- Launch address: GmNzSXzQ3q6UCVRpBf8PkvEqoo454Qr6twWc9zuzJzBa