Hatim El JazouliCatalog

Hatim El Jazouli

RFormal methods,PPoetry,FForecasting,NNotes,QSide quests

Entries
34
Tracings
20
Crossing class
10
Last accession
2026-10-14

Casablanca · Email · Substack · GitHub · LinkedIn · Metaculus · X

Tracing trafficcounts off the diagonal cross a class boundaryRFORMAL METHODSPPOETRYFFORECASTINGNNOTESQSIDE QUESTS
From class, to classRFORMAL METHODSPPOETRYFFORECASTINGNNOTESQSIDE QUESTS
RFORMAL METHODS··1··
PPOETRY···3·
FFORECASTING··21·
NNOTES1·281
QSIDE QUESTS···1·
  1. F104Viva connects to N106The Noise is the System

    Viva is built on the essay's point that you only learn where you stand from contact with reality. Here that means running the code instead of asking a model.

  2. N108Steeped in Paradoxes connects to F101Ensemble AI forecasting bot

    The VAR argument is the one I care about in forecasting too: random error averages out, systematic bias does not.

  3. N107Calibrated Exposure connects to Q101Hull Tactical market prediction

    The essay says I finally started building market models. This is one of them.

  4. Q102Protein function prediction for CAFA 6 connects to N107Calibrated Exposure

    The essay explains why I left bioinformatics. CAFA 6 is where I came back to it with different tools.

  5. N105Why I’m interested in AI safety connects to R101Context Tree Switching implementation

    AIXI is a mathematical model of an ideal agent. Implementing one of its building blocks was my way into the theory behind the systems this essay worries about.

  6. N105Why I’m interested in AI safety connects to F103Intelligence Workbench

    The essay ends with wanting to help people understand and forecast AI behavior. The workbench gathers the evidence and leaves the judgment to the forecaster.

  7. R101Context Tree Switching implementation connects to F101Ensemble AI forecasting bot

    Both combine many predictors instead of trusting one. CTS mixes context models as data arrives; the bot combines forecasts from different personas.

  8. P118Clean as Vile connects to N107Calibrated Exposure

    The poem sits in the murk waiting for an epiphany. The essay ends by refusing to wait: act before understanding is complete.

  9. P117Nods in Hays connects to N106The Noise is the System

    The poem asks the noisy world for a tranquil mind. Three years later the essay stops asking: the noise is the sound of the system working.

  10. P116Nuit de suie connects to N102B

    The poem writes through the ennui that follows a black night. Two years later B takes the other exit: sleep as escape.

  1. Viva

    Forecasting. — 2026-09-26.

    An oral-style exam generated from a student's own JavaScript: predict what small mutations of the code return, say how sure you are, and be graded by running the code.

    TracingsI.N106Viva is built on the essay's point that you only learn where you stand from contact with reality. Here that means running the code instead of asking a model.

  2. Intelligence Workbench

    Forecasting. — 2026-09-23.

    A Rust and Python service that turns a research question into a sourced, as-of-queryable evidence graph and an analytical briefing that never forecasts.

    Tracingsnone recorded

  3. Betomcat

    Forecasting. — 2026-09-21.

    A Metaculus forecasting bot, forked from the Metaculus bot template, that researches each question through the Intelligence Workbench and forecasts it with two models drawn from a pool.

    TracingsI.F103The workbench is built to feed this bot: it gathers and grades the evidence, the bot makes the forecast.II.F101

  4. Exert, the Pursuit

    Notes. — 2026-06-29.

    On agency.

    TracingsI.N106The same argument in two registers: competence comes from contact, one step or one closed loop at a time.

  5. Steeped in Paradoxes

    Notes. — 2026-03-31.

    On football

    TracingsI.F101The VAR argument is the one I care about in forecasting too: random error averages out, systematic bias does not.

  6. Calibrated Exposure

    Notes. — 2026-03-24.

    On excavation

    TracingsI.N108This essay ends by choosing football analytics. A week later, this is what that interest sounds like on the page.II.Q101The essay says I finally started building market models. This is one of them.

  7. The Noise is the System

    Notes. — 2026-02-10.

    On exposure.

    TracingsI.N107The next essay takes the same word, exposure, and turns it on the past instead of on work.