Videos

Janet narrates, the kernel verifies.

A 5-topic, 10-subjects-each video arc — 50 explainers, ~10–15 minutes apiece. Mostly animated cards + D3/Cytoscape graphics. Janet — an AI-rendered presenter (HeyGen avatar) — narrates. Source plan lives in the explaining/ subproject; topic shape is fixed, subjects can shift. The clip below is the master-still test from the voice/lipsync pipeline — narration + on-screen graphics will composite around it.

Channel · YouTube

@Quantapix

Janet narrates, the kernel verifies. A 5×10 explainer arc on axiomatic agents.

Watch on YouTube → 4 live · 50 planned · 5 topics

1.1 — The hallucination tax — where LLMs lose at high stakes

1.1

1.5 — Negative Verification — When the Kernel Says "Not Proven"

1.5

1.7 — Civil RICO, End to End — Predicate, Sub-Axioms, § 1962 Elements

1.7

2.1 — When a docket and its record disagree — a real narrative-analysis use case

2.1

The full arc lands on @Quantapix as cuts clear the pipeline. Each video ends pointing at the open-source repos and the call to help axiomatize the U.S. Code — see the donate page.

Preview · Janet master-still test

First HeyGen Photo Avatar IV pass against the wardrobe-locked master still. Voice + lipsync only — no animated overlays yet. Source still and meta in explaining/voice/janet/.

Latest release · 1.1

WRONG

The hallucination tax — where LLMs lose at high stakes

Watch on YouTube → Watch the cut →

The brief — verbatim

I need to attract attention to my work, and while I did manually create a few YouTube videos, the experiment fell flat. With qagents now scaffolded, I want you to help me write scripts for a set of not more than 10–15 minute videos about my work. The videos would be mostly rendered/animated cards/text and lots of graphics. Therefore, I will need scripts to submit to Claude Design. Before actually doing any design work, I need an outline for the 5 topics, i.e., 1) "Why theorem provers… instead of semantic searches?"; 2) "Why narrative analysis… or Qnarre?"; 3) "Why result evaluations… or Qresev?"; 4) "Why rigorous debates about… technical analysis?"; and 5) "About Quantapix… the Jack and Janet startup". Each topic should have 10 videos that would be narrated by Janet, I will provide a picture for her with the understanding that a (to be recommended) AI will render her "narration" while playing the animated cards/text or running the high quality "graph actions".

5 topics · 10 subjects each

Each topic is a chapter of ten. Live subjects link to YouTube and the direct cut; the rest open as cuts clear the pipeline. Expand a chapter to see its slate.

Topic 1 Why theorem provers… instead of semantic searches? P1 3 live

proving/ · accounting/

1.1 The hallucination tax — where LLMs lose at high stakes WRONG live

1.2 What Semantic Search Actually Does — and What It Can't upcoming

1.3 The Lean4 Kernel — a 10kloc Oracle for Proof Correctness upcoming

1.4 Predicates — Where Judgment Lives Between LLM and Kernel upcoming

1.5 Negative Verification — When the Kernel Says "Not Proven" NOT PROVEN live

1.6 Axioms vs. Evidence — What We Don't Trust the LLM With upcoming

1.7 Civil RICO, End to End — Predicate, Sub-Axioms, § 1962 Elements RICO live

1.8 Title VI — Discriminatory Intent Has Structure upcoming

1.9 Same Kernel, Different Axioms — Scaling to Financial Contracts upcoming

1.10 The Cost — Latency, Surface Area, and Auditability as a Feature upcoming

Topic 2 Why narrative analysis… or Qnarre? P2 1 live

verifying/ · proving/ · appealing/ · legal/

2.1 When a docket and its record disagree — a real narrative-analysis use case DIFF live

2.2 What a "Narrative" Is, Formally — Predicates Over Named Entities upcoming

2.3 Three-Zone Qnarre — Input, Kernel, Witness LIVE upcoming

2.4 Watching a Proof Elaborate Live — SSE Event Streaming upcoming

2.5 From DOCX to Lean — How a Complaint Becomes a Claim Object upcoming

2.6 When the Appendix Is the Source of Truth — RAv:p Record Citations upcoming

2.7 Pro Se on Appeal — Encoding the Procedural Traps as Predicates upcoming

2.8 Counterfactual Narratives — What a Defense Story Has to Deny upcoming

2.9 Stacking Predicates — Civil RICO and § 1983 Over One Fact File upcoming

2.10 What Qnarre Never Does — Write Your Brief for You upcoming

Topic 3 Why result evaluations… or Qresev? P3

evaluating/ · accounting/ · analyzing/

3.1 Backtests Lie — the Survivorship and Look-Ahead Problem LIES upcoming

3.2 Six Defined-Risk Strategies — and a Hard Refusal of the Rest upcoming

3.3 Five Frameworks — TREND, MOMENTUM, OPTIONS-RISK, SECTOR, DRAWDOWN upcoming

3.4 Why a Sector Cap Needs a Kernel, Not a Spreadsheet upcoming

3.5 One Bar Schema, Ten Consumers — the OHLCV Parquet Hub upcoming

3.6 Portfolio In, Verdict and Witness Out — Qresev Live TRACE upcoming

3.7 Drawdown as a Theorem — the Conservative Mandate as a Prop upcoming

3.8 Where the LLM Lives — Predicate Judgments Over Price Action upcoming

3.9 Three PMs Disagree, the Kernel Doesn't — Per-PM Axioms upcoming

3.10 What Qresev Refuses — Naked Options, Leverage, Look-Ahead upcoming

Topic 4 Why rigorous debates about… technical analysis? P4

analyzing/ · monitoring/ · aggregators

4.1 The TA Bestiary — 117 Indicators, Mostly Redundant 117 upcoming

4.2 TA-Lib as Ground Truth — the MACD EMA-Realignment Quirk QUIRK upcoming

4.3 DuckDB and Parquet — Why Columnar Beats Row-Store for Bars upcoming

4.4 Lightweight-Charts v5 — 10 Years of Bars at 60fps upcoming

4.5 Aggregators — Small-Multiples, Heatmaps, Three.js Surfaces upcoming

4.6 Three PMs on One Feed — Where the Disagreements Live upcoming

4.7 Even the Chart Viewer Refuses Naked Legs — the Options Floor upcoming

4.8 The Bar You Saw vs. the Bar You Traded On — Data Sourcing upcoming

4.9 "Tech" Is Not a Sector — GICS as the Only Referent upcoming

4.10 What an Indicator Is, Formally — TA as Testable Predicates upcoming

Topic 5 About Quantapix… the Jack and Janet startup P5

qagents/ · lib/memsearch/ · sub-agents

5.1 Two People, One Repo — What "Jack and Janet" Means TWO upcoming

5.2 The qagents Monorepo — 20 Subprojects, One Venv, One Workspace TOUR upcoming

5.3 Claude Code as the Third Teammate — Agents You Can Fire upcoming

5.4 CLAUDE.md — Instructions for the Colleague Who Never Reads Twice upcoming

5.5 memsearch — Remembering What Was Said Last Tuesday upcoming

5.6 The 5×10 Plan — Made by Claude, for Claude to Make upcoming

5.7 Two Products From One Kernel — Qnarre and Qresev upcoming

5.8 The Legal Arc — From Pro-Se Filings to a Verifier Product upcoming

5.9 The Financial Arc — From Analyzing TA to Evaluating Portfolios upcoming

5.10 What Quantapix Is Betting On — Kernels Under Everything upcoming