AI Adoption is a Trust Problem
AI Adoption is a Trust Problem
Thought Eddies
An experimental digital garden
Latest posts
Claude Code and Agent Conventions
Claude Code and Agent Conventions
Chatbots are not input validation
Chatbots are not input validation
Direnv env var additions, overrides, and removals
Direnv env var additions, overrides, and removals
Computer Use
Computer Use
Stateful Agent Collaboration
Stateful Agent Collaboration
Externalize Your Thought Process
Externalize Your Thought Process
Personal Software
Personal Software
Stateful Agents and Basic Memory
Stateful Agents and Basic Memory
Director, Not Performer
Director, Not Performer
I've Stopped Learning Programming Languages
I've Stopped Learning Programming Languages
Reviving a Raspberry Pi with Claude Code
Reviving a Raspberry Pi with Claude Code
Why Write
Why Write
Year in Review: 2025
Year in Review: 2025
Images as Context
Images as Context
On Writing, Speaking and Thinking
On Writing, Speaking and Thinking
Embeddings with sqlite-vector
Embeddings with sqlite-vector
Debugging DSPy token usage and prompts
Debugging DSPy token usage and prompts
Intro to DSPy
Intro to DSPy
Running a Mile with LLMs
Running a Mile with LLMs
Claude Code Custom Slash Commands Hierarchy
Claude Code Custom Slash Commands Hierarchy
Running `uv` with Environment Variables
Running `uv` with Environment Variables
Datadog and the OpenAI API Spec
When Datadog's magic breaks your API calls
Migrating To Astro
A short, fast rollercoaster
Find the Groove
Using Checkpoints with Coding Agents
State Of Code Agents - July 2025
Agents can make progress independently, but they also make messes
Gemini Hidden Reasoning
The performance costs of thinking and provider defaults
With MCP, Dates are a Footgun
With MCP, Dates are a Footgun
Gemini 2.5 Uses Thinking By Default
Gemini 2.5 Uses Thinking By Default
Amazon Bedrock On Demand Throughput Error
Amazon Bedrock On Demand Throughput Error
A few strategies to make LLMs useful for you
A few strategies to make LLMs useful for you
Agent Coding Strategies
Agent Coding Strategies
Who Finds LLMs Useful and Why
Who Finds LLMs Useful and Why
OpenAI Compatible API Implementation Variance
OpenAI Compatible API Implementation Variance
Note Taking
Note Taking
Using `llm` to handle large input context
Using `llm` to handle large input context
Git Worktree
Git Worktree
Create a Link Blog
Create a Link Blog
Alfred as a Text Editor
Alfred as a Text Editor
LLM Chat UI Concepts
A few concepts for LLM chat UIs
RSS Club
RSS Club
Models as Force Multipliers and their Limits in Software Development
You need to use models to build software to really understand their limits
Vibe Coding in early 2025
Vibe Coding in early 2025
Goose as a Task Runner
Goose as a Task Runner
Adding an llms.txt file to Hugo
Adding an llms.txt file to Hugo
Claude Code
Claude Code
Cursor Triple Backticks Stop Sequence
Cursor Triple Backticks Stop Sequence
Developing a Mental Model for using Models
Developing a Mental Model for using Models
Astro Code Toggle Component
Astro Code Toggle Component
Document Citations
Using Claude Citations to annotate the sources for document Q&A