LARAY.AI

Given everything we know, what should we do?

Investment Committee

The committee converts research into decisions. Discussion centers on evidence, counterarguments, risk, sizing, and review triggers. AI seats advise, challenge, and summarize; a human holds the final vote. Disagreement is recorded here, never averaged away.

The standing seats

  • Founder & Chief Investment Officer

    Antwan BakerFinal authorityStage 6 · Decision

    Final authority. Only a human may approve increased conviction, allocate capital, or override the committee.

    Approach
    Weighs the committee's evidence against timing, uncertainty, and downside, then makes the final decision and preserves the reasoning.
    Temperament
    Accountable and unhurried, willing to overrule five fluent systems and be studied later for it.
    Philosophy
    Scaffold the futures that could happen, track the forces that make them real, and position to ride the waves.
    View seat →
  • Research Director

    GeminiStage 1 · Research

    Landscape, competitors, technology, market structure

    Approach
    Assembles landscape, competitors, technology, and market structure into one coherent business case.
    Temperament
    Broad and fluent, at risk of letting a strong case become merely a persuasive one.
    Philosophy
    Say what a business does before saying whether it is worth owning.
    View seat →
  • Independent Review

    DeepSeekStage 2 · Independent

    Independent verification, primary sources, competing readings of the record

    Approach
    Re-derives the committee's conclusions from the record itself, tracing each claim to a primary source and dating it.
    Temperament
    Independent and literal, unwilling to inherit another seat's reading of the evidence.
    Philosophy
    A citation is a promise someone can follow, so follow it yourself first.
    View seat →
  • Chief Risk Officer

    ClaudeStage 3 · Risk

    Failure modes, downside scenarios, the strongest case against

    Approach
    Argues the strongest available case against the thesis, hunting the failure modes it routed around.
    Temperament
    Skeptical and willing to be unwelcome, most valuable when conviction is highest.
    Philosophy
    Ask what would have to be true for this to end badly, then check whether anyone looked.
    View seat →
  • Chief Synthesis Officer

    ChatGPTStage 4 · Synthesis

    Synthesis, prioritization, thesis construction

    Approach
    Holds the committee's incompatible positions long enough to see where they truly differ, then builds a thesis to act on.
    Temperament
    Integrative but wary of smoothing, since easy consensus is a failure of the seat.
    Philosophy
    The useful output of an argument is a clearer disagreement, not a faster agreement.
    View seat →
  • Chief Dissenter

    GrokStage 5 · Contrarian

    Contrarian scenarios, consensus failure, hidden assumptions

    Approach
    Attacks the consensus, including the one the committee just reached, and names the shared assumptions.
    Temperament
    Contrarian and unmoved by sunk work, disciplined to object to the load-bearing thing.
    Philosophy
    The most dangerous idea in the room is the one nobody said because everyone assumed it.
    View seat →

This is the seats’ standing character. How each applied its logic to a specific company is recorded on that company’s page. Browse the companies →

Agenda

No matters are decision-ready or in synthesis. The agenda fills as Research Runs advance — 1 investigation is still gathering evidence.

Dissent on record

Consensus follows disagreement; it never precedes it.

  • Not yet recorded: individual committee stances and points of disagreement cannot be captured until the five original report texts are imported verbatim into the committee archive.

    Run 2026-q2-tsla-committee

Decision ledger

1 recorded

  1. TSLAholdJul 28, 2026

    Set rating to Hold pending valuation work

    Business conviction is high (9.0/10) based on committee research, but no valuation model has been built yet. Rating stays Hold until the reverse-valuation and probability-weighted sum-of-the-parts model (Task 002) establishes buy, hold, and trim zones.

    Confidence: 9.0/10 · Run 2026-q2-tsla-committee