2026

• Communication Between Runners and Agents

  Oct 2

• Static and Dynamic Reconciliation of an Agent Team from a Description

  Sep 27

• Agent Description

  Sep 27

• Agent Sessions

  Sep 27

• Boss session pattern

  Sep 27

• Social Networks as a Tool for Honing Special Knowledge

  Sep 26

• A Text Format for Drawing UX Components

  Sep 26

• Plain Text Health

  Sep 26

• Research as Finding a Path Through a Graph

  Sep 26

• Separating Core Work from Technical Work

  Sep 25

• Overengineering: The Silent Killer

  Sep 25

• Independent research instead of a single llmwiki

  Sep 25

• Jevax: Jev-driven EMACS

  Sep 24

• What if every password were correct?

  Sep 23

• Claude Opus 5.5 and GPT 6 Sol

  Sep 23

• Testing Sending an Email to the Blog

  Sep 23

• Test Post from Email Publisher

  Sep 23

• Public Researcher

  Sep 19

• coldteadotai/pr-lens

  Sep 16

• fx.sh

  Sep 15

• Social media strategy

  Sep 14

• Reflection.dev and Zeno

  Sep 14

• Test post

  Sep 13

• Trying Zulip instead of Discourse

  Sep 10

• Zeno, an always-on orchestrator for ephemeral agents

  Sep 9

• Long-lived agents vs ephemeral agents

  Sep 9

• An interest profile as a product

  Sep 9

• Ephemeral agents

  Sep 6

• An auto-researcher built on my blog

  Sep 6

• Architecture review

  Sep 5

• The pain of auto-mode agents in OMP

  Sep 5

• SecretSpec

  Sep 4

• UTM

  Sep 4

• Installing NixOS on a UTM VM on macOS

  Sep 4

• Hyprland

  Sep 4

• 16:9 screenshots

  Sep 4

• Omarchy

  Sep 4

• Numtide/llm-agents

  Sep 3

• How and why to keep agent logs

  Sep 3

• vLLM

  Sep 3

• Lightpanda

  Sep 2

• BB app

  Sep 2

• Bitnami/sealed-secrets

  Sep 2

• A devbox on a remote server

  Sep 2

• Onboarding agents

  Sep 1

• Castle architecture

  Sep 1

• Pijul

  Sep 1

• Rethinking the vision for Reflection Castle

  Sep 1

• Discourse as a platform for an AI-native company

  Aug 31

• Lisp as an interface

  Aug 31

• Reconciling documentation

  Aug 31

• Mesh LLM

  Aug 26

• Tauri 2.0

  Aug 26

• Snapcompact: SoTA compaction, instant, local, free. Pick 3

  Aug 26

• Microsandbox, a computer for anything, anywhere

  Aug 26

• Cloudflare Mesh

  Aug 26

• An org-mode based harness

  Aug 23

• Shottr, a screenshot app for pixel professionals

  Aug 22

• Nostr - Notes and Other Stuff Transmitted by Relays

  Aug 21

• Buzz: a hive mind communication platform

  Aug 21

• Omp -- a coding agent with the IDE wired in

  Aug 21

• OpenSSF Best Practices Badge Program

  Aug 21

• Crossplane Is the Cloud-Native Framework for Platform Engineering

  Aug 21

• The Open Source LLM Router for AI Agents | Manifest

  Aug 21

• A Programming Paradigm for Spatiotemporal Composability

  Aug 20

• Self-description and a permission sandbox for agents in LISP

  Aug 19

• AI doesn't work in a team, and how to fix it

  Aug 19

• A personal stream, with voice as the main interface

  Aug 19

• An experience lake, and keeping personal data safe

  Aug 19

• Collect data about yourself seamlessly, don't lean on discipline

  Aug 18

• Kanban is the interface between a human and an agent, not between agents

  Aug 18

• The shared context problem on teams that use AI

  Aug 17

• Infrastructure as a formal system

  Aug 17

• Auto-publishing to my Telegram channel

  Aug 16

• A personal AI assistant as a butler

  Aug 16

• A context/reaction tree as a ZK proof of expertise

  Aug 16

• A kanban board for my AI agents

  Aug 16

• Why DevOps matters

  Aug 16

• Context as a separate stream of events

  Aug 16

• A universal voice assistant

  Aug 16

• The company as a system

  Aug 15

• Setting up an open-source reading stack

  Aug 15

• Infinite context

  Aug 15

• What if I used my own social media to learn?

  Aug 15

• Looking at markets from first principles

  Aug 15

• A stream of reactions as a user profile

  Aug 14

• A personal public EDU stream

  Aug 14

• Training a personal LLM on what you actually do

  Aug 13

• Is there any demand for owning your own data?

  Aug 13

• PostHog: monitoring through the user's eyes

  Aug 12

• Tags on my blog

  Aug 11

• A speech recognition tool

  Aug 11

• Speech to text

  Aug 11

• Blogging in the age of AI

  Aug 11

• A blog for author-based writing

  Aug 11

• Lazy MCP: Dispatch and Discovery

  May 25

• AI Won't Teach You — Unless You Make It

  Jan 30

• Short Sessions Beat Long Ones in Claude Code

  Jan 30

• The Real Metric for Autonomous AI: Quality Without Hand-Holding

  Jan 29

• Small Steps Beat Big Leaps: Why Familiar Interfaces Win

  Jan 29

• Clawdbot, Capsules, and Self-Evolving Agents

  Jan 28

• Botlord: The Human as Corporation

  Jan 27

• Product Over Technology

  Jan 27

• Capsules: Isolated Environments for AI Agents

  Jan 22

• From Solo Sessions to Agent Orchestras

  Jan 21

• Robots Should Do Everything

  Jan 20

• How Many Tokens Did You Spend on This Task?

  Jan 20

• No Way Back: Why Constraints Drive Better Workflows

  Jan 15

• My Digital Twin Starts with Claude Code

  Jan 15

• Unrealized Risks Are Still Real Costs

  Jan 15

• Everything Is Text

  Jan 14

• Done Today Beats Perfect Never

  Jan 14

• Better Written Than Lost

  Jan 14

• Command: Controlling the World Through Text

  Jan 10