H
Howardismvol. 03 · quiet corner of the web
Howardism · Vol. 03Masthead · No. 00

A wiki, set in plates.

CuratorHoward TaiIssueVol. 03 · no. 99CadenceBatched, not scheduledBasedTaiwan, 23°NContacthoward@howardism.dev

Working notes on interaction-era AI, kept publicly. Each plate is a topic; each topic carries its brightest notes and the source I'm currently digesting. 99 notes across 25 sources, last filed 23 May 2026. Updated whenever a batch finishes — not on a schedule.

● CurrentsWhat's moved recently.
  1. 23 May 2026
    • lintFull audit (post-worktree-merge)
    • compileAndrej Karpathy — From Vibe Coding to Agentic Engineering
    • compileFiona Fung — Running an AI-Native Engineering Org
    • compileJohn Glasgow — Campfire AI-Native ERP
    • compileDeepMind — AI-Driven Formal Proof Search (new domain)
  2. 22 May 2026
    • lintFull audit
  3. 21 May 2026
    • ingestHow I AI: Thariq Shihipar on Replacing Markdown with HTML (ChatPRD / Claire Vo, May 2026)
    • compileCross-link sweep
    • lintFull audit
    • queryDoes the human-facing harness (HTML artifacts) hit its own bloat ceiling?
    • maintenanceVerify raw-source `published` dates against live articles
  4. 18 May 2026
    • ingestThe Founder's Playbook: Building an AI-Native Startup
Plate · 01
12
filed under interaction
Activity · 8 wks

Interaction, in order.

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

  1. 01Orchestration vs Employee Framing: Reconciling the Founder's Playbook with HBR's Accountability Evidence23 May 2026 · 14′
  2. 02Compute Allocator21 May 2026 · 6′
  3. 03Disposable Micro-Apps21 May 2026 · 5′
  4. 04Living Design System21 May 2026 · 4′
  5. All 12 notes in interaction →
Plate · 02
21
filed under architecture
Activity · 8 wks

Architecture, in order.

Model internals: encoder-free fusion, training, scaling, evals.

  1. 01The AI-Native Safe-Choice Inversion23 May 2026 · 4′
  2. 02Building Is Cheap, Arguing Is Expensive23 May 2026 · 4′
  3. 03Code as Source of Truth23 May 2026 · 4′
  4. 04Founder-Led Sales Discipline23 May 2026 · 5′
  5. All 21 notes in architecture →
Plate · 03
25
filed under harness
Activity · 8 wks

Harness, in order.

Agent loops, tools, orchestration, and Claude Code patterns.

  1. 01Agent-Native Infrastructure23 May 2026 · 4′
  2. 02Agentic Loops Overtake Bespoke Systems23 May 2026 · 5′
  3. 03AI-Driven Formal Proof Search23 May 2026 · 6′
  4. 04Evolutionary Proof Search23 May 2026 · 5′
  5. All 25 notes in harness →
Plate · 04
6
filed under alignment
Activity · 8 wks

Alignment, in order.

RLHF, character, interpretability, and safety case studies.

  1. 01Agentic Misalignment (AM)8 May 2026 · 4′
  2. 02Alignment Fine-Tuning (AFT)8 May 2026 · 3′
  3. 03Chain-of-Thought Monitorability8 May 2026 · 3′
  4. 04Deliberative Alignment8 May 2026 · 4′
  5. All 6 notes in alignment →
Plate · 05
35
filed under orgs
Activity · 8 wks

Orgs, in order.

Labs, people & products: Anthropic, OpenAI, TML, Claude Code.

  1. 01AlphaProof Nexus23 May 2026 · 5′
  2. 02Andrej Karpathy23 May 2026 · 4′
  3. 03Campfire23 May 2026 · 5′
  4. 04Dogfooding as Product Discipline23 May 2026 · 5′
  5. All 35 notes in orgs →
Plate · ∞

The desk.

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

Talk
6
Article
14
Paper
4
Repo
1