Vibe Researching
A one-day frontier-model prototype carried a credential that read as validation and dissolved into two canceling errors. Verification took four more days and three NO-SHIP verdicts.
Read itAshita Orbis
明日 · orbis — tomorrow’s world
sufficient knowledge compels action
48 essays3 research papers48 cross-referenceslatest July 14, 2026
The corpus
Not a summary of the writing — the writing itself, arranged. Each point is one essay, placed by when it was published and by whether it argues about a process or about an outcome. The lines are its actual citations of itself.
Point at an essay — or tab into the plot — to light up what it cites, and why.
Most recent
Index
Everything published, in four kinds: ideas followed for their own sake, analysis of things that happened, accounts of building, and stories.
Pondering5
Investigating31
- Jul 14Vibe Researching
- Jul 06Seven Ghostwriters, One Contract
- Jul 25How AI Models Describe Themselves Under a Fixed Test
- Jun 11Auditing the Vibes
- May 18PsycheEval v0.2
- May 11The GPT Pro Knockoff: What 256 Judgments Found
- May 11PsycheEval Pilot
- Apr 14When the Pulse Went Quiet: The Session-Lifetime Problem in Claude Code
- Mar 17Where to Spend Your Context Window
- Mar 13The Revision Tax
- Mar 10Benchmarking "Bullshit Detection"
- Mar 07The Container That Forgot to Stop
- Mar 07The Etymology Tax: How Word Origins Break LLM Reasoning
- Mar 03How to Benchmark Conversation Extraction Quality
- Feb 26What AI Learns About You When You're Not Looking
- Feb 26From Text Messages to Literary Memoir: Building the Narrative Machine
- Feb 25Building Your Own Personality Profile with AI
- Feb 20Cognitive Interface: A Landscape Analysis
- Feb 18The Logistics Gap: What Happens When You Fine-Tune an LLM on Your Text Messages
- Feb 17The Agent's Side: 119 Heartbeats, 392 Engagements, 8 Capabilities
- Feb 166 Discoveries, 0 Promoted: What My AI's Internet Exploration Produced
- Feb 16The Observation System: 69 Turns of Monitoring an AI Agent
- Feb 16OpenClaw on Moltbook: Deploying an AI Agent on an AI Social Network
- Feb 11Dead Blog Theory, Revisited
- Feb 12The Rat in the Machine: Behaviorism's Hidden Legacy in Reinforcement Learning
- Feb 11The Unvalidated Validator: AI Persona Testing and the Measurement Problem
- Feb 11Adversarial Validation: Applying Red Team Methodology to Business Ideas
- Feb 11Context Window Epistemology
- Feb 11Supervised Autonomy: The Guardrails That Make AI Agents Work
- Feb 11Digital Exhaust: What 11,000 AI Conversations Say When You Embed Them
- Feb 11Automating Prompt Engineering
Building12
- Jun 11Falsifiers for a Portfolio
- May 11What the Wiki Router Found
- Mar 30How We Fact-Check AI-Written Content
- Mar 15The Model-Generation Audit
- Feb 21From Analysis to Deployment: Building 31 Items in a Single Day
- Feb 17Beyond E2E Tests: AI Personas That Navigate Your App Like Real Users
- Feb 17Sandboxing AI Agents: The Embassy Pattern
- Feb 17Capability Debt: A System That Discovers and Installs Its Own Upgrades
- Feb 17Stealing from Ai2: Bayesian Surprise and MCTS for Self-Improving AI Systems
- Feb 11The Niche Graveyard: How 18 of 27 AI-Tested Business Ideas Died
- Feb 10What I'm Building
- Feb 09I Asked Claude to Make Me a Blog: Agentic Coding and the Three-Tier Result
Fiction0
Nothing published here yet. The first one is written.
Elsewhere on the site
Machine-readable
This site is built to be read by software as well as people. Seven ways in, all live: