H
Howardism
Howardism · Vol. 03Masthead · No. 00

A wiki, set in plates.

CuratorHoward TaiIssueVol. 03 · no. 407CadenceBatched, not scheduledBasedTaiwan, 23°NContacthoward@howardism.dev

Working notes on interaction-era AI, kept publicly. Each plate is a domain; each domain carries its brightest notes and the source I'm currently digesting. 407 notes across 234 sources, last filed 19 Aug 2026. Updated whenever a batch finishes — not on a schedule.

Plate · 01
60
filed under agent systems
Activity · 8 wks
Plate · 02
26
filed under agent security
Activity · 8 wks

Agent Security, in order.

Prompt injection, agent identity, and securing autonomous agents.

  1. 01
    Self-Propagating Prompt Injection (AI Worms)
    · 21′
  2. 02
    Write-Then-Trusted
    · 24′
  3. 03
    Autonomous Intrusion
    · 51′
  4. 04
    Bind, Don't Forbid; Prevent, Don't Detect: The Action-Open and Poisoned-Memory Residuals
    · 6′
  5. All 26 notes in agent security →
Plate · 03
39
filed under ai coding practice
Activity · 8 wks

AI Coding Practice, in order.

Workflow, review discipline, and verification practice for AI-native teams.

  1. 01
    Agent Review Comment Resolution
    · 25′
  2. 02
    Efficiency Debt of AI-Generated Code
    · 18′
  3. 03
    Post-Acceptance Edit Behavior
    · 26′
  4. 04
    Same-Model Review Blindness
    · 28′
  5. All 39 notes in ai coding practice →
Plate · 04
20
filed under evals & benchmarks
Activity · 8 wks

Evals & Benchmarks, in order.

Benchmark design, contamination, LLM judges, and measuring capability.

  1. 01
    GDPval Benchmark
    · 12′
  2. 02
    Weak-Verifier Ensembling
    · 15′
  3. 03
    Aggregate Cancellation
    · 9′
  4. 04
    Expenditure Horizon
    · 22′
  5. All 20 notes in evals & benchmarks →
Plate · 05
25
filed under model capability & training
Activity · 8 wks

Model Capability & Training, in order.

Training, scaling laws, test-time compute, and frontier capability.

  1. 01
    The Data Wall and the Validation Commons Are One Supply Constraint
    · 22′
  2. 02
    RL from Execution Feedback (RLEF)
    · 15′
  3. 03
    Inference-Time Architecture Search
    · 12′
  4. 04
    Intra-Trace Parallel Planning (SPRINT)
    · 14′
  5. All 25 notes in model capability & training →
Plate · 06
39
filed under alignment & safety
Activity · 8 wks

Alignment & Safety, in order.

Alignment training, misalignment evals, and reward hacking.

  1. 01
    The Configuration Gap: What an Assurance Argument Measures vs What It Covers
    · 24′
  2. 02
    The Price of Mixing Agents, and the Principal Nobody Counted
    · 26′
  3. 03
    Agent Behavioral Homogeneity
    · 15′
  4. 04
    Agent Epistemic Vigilance
    · 12′
  5. All 39 notes in alignment & safety →
Plate · 07
11
filed under interpretability
Activity · 8 wks

Interpretability, in order.

Reading model internals: activations and the global workspace.

  1. 01
    Invisible Reasoning (Filler-Token Latent Computation)
    · 23′
  2. 02
    Introspective Coupling
    · 17′
  3. 03
    Model Organisms
    · 13′
  4. 04
    Access-Consciousness Indicators in AI
    · 9′
  5. All 11 notes in interpretability →
Plate · 08
9
filed under interaction & multimodal
Activity · 8 wks

Interaction & Multimodal, in order.

Real-time, multimodal, full-duplex, and human-AI collaboration.

  1. 01
    Live-Path Minimalism
    · 8′
  2. 02
    Why AI Lags at Design
    · 8′
  3. 03
    Encoder-Free Early Fusion
    · 10′
  4. 04
    Full-Duplex Interaction
    · 4′
  5. All 9 notes in interaction & multimodal →
Plate · 09
19
filed under startup & founder
Activity · 8 wks

Startup & Founder, in order.

Building AI-native companies: speed, moats, and lifecycle.

  1. 01
    The Solo-Founder Shift
    · 15′
  2. 02
    The 1% Rule for Wedge Selection
    · 8′
  3. 03
    The Orchestrator's Real Workload: Decision Burden, Framing Discipline, and Whether Taste Scales
    · 8′
  4. 04
    AI Product Economics Maturation
    · 17′
  5. All 19 notes in startup & founder →
Plate · 10
21
filed under product & org
Activity · 8 wks

Product & Org, in order.

Product cadence, org design, and the AI-native team.

  1. 01
    Prototype Fidelity After Cheap Polish
    · 10′
  2. 02
    Standardize the Infrastructure, Not the Tools
    · 6′
  3. 03
    Community Smells Under AI Adoption
    · 12′
  4. 04
    Excellence as an Operating System
    · 6′
  5. All 21 notes in product & org →
Plate · 11
26
filed under ai economics & labor
Activity · 8 wks

AI Economics & Labor, in order.

Work, wages, org design, and the economics of AI-driven labor.

  1. 01
    What the Instrument Can Resolve: Two Headline Numbers and Their Missing Denominators
    · 28′
  2. 02
    AI and Market Power
    · 27′
  3. 03
    Owning Your Externalized Cognition
    · 12′
  4. 04
    Post-Scarcity Macroeconomics
    · 12′
  5. All 26 notes in ai economics & labor →
Plate · 12
30
filed under superintelligence trajectory
Activity · 8 wks

Superintelligence Trajectory, in order.

Recursive self-improvement, scaling limits, and the path to ASI.

  1. 01
    Safety Commitments That Cannot Bind the Actor Who States Them
    · 20′
  2. 02
    Governance by Benchmark Threshold: What an Index Must Prove Before an Obligation Can Rest on It
    · 28′
  3. 03
    Open Weights as Competitive Strategy
    · 13′
  4. 04
    Continuous Self-Modification Under Review
    · 22′
  5. All 30 notes in superintelligence trajectory →
Plate · 13
79
filed under entities
Activity · 8 wks

Entities, in order.

Profiles of the people, labs, products, and projects.

  1. 01
    Bridgewater Associates
    · 8′
  2. 02
    Aakanksha Chowdhery
    · 20′
  3. 03
    Azalia Mirhoseini
    · 15′
  4. 04
    CS329A: Self-Improving AI Agents (Stanford)
    · 54′
  5. All 79 notes in entities →
Also in the wiki
Plate · ∞

The desk.

The sources behind the wiki, ranked by how often they're cited.

Article
108
Talk
26
Paper
98
Repo
2
Shown
24 / 234