An Agent Reads My Email
Every morning at five, a local model reads yesterday's mail and writes me a briefing. How it works, what 217 days of numbers say, and an honest answer to the prompt-injection question.
Read moreBlog
Thoughts on engineering,
AI, and building products.
The Processing Layer Is Not a Moat
LLMs are the refinery. Everyone has access to the refinery now. The only question is: who owns the crude?
Read moreNotes from the AI Gatherings, Vol. 1
First in a recurring series - what we compared notes on this week at the Friday morning AI gathering: cleaning data 700 calls at a time, the two economies of tokens, and building the tool before the feature.
Read moreWhispering to the Machine: Take Three
Five months after Take Two, the third snapshot. The agents work nights now, the world pushed its way into the workflow, and the most durable thing on my machine is the record.
Read moreThe Architecture of Advice: Inside the AI High-Fidelity Loop
Once you have a month of AI session data, you stop asking 'how much?' and start asking 'what kind?' Mapping the hidden patterns that define how AI actually influences a multi-repo technical portfolio.
Read more30 Days of AI Collaboration: Recovering 55,000 Turns of Professional Capital
A data-driven retrospective of one month of high-intensity AI-assisted engineering: 466 sessions, 55,224 turns, 1,772 decisions extracted from raw chat logs into a durable knowledge graph. Built the day before Anthropic announced 'dreaming' as a feature.
Read moreGetting good output from Lens: the model and LM Studio settings matter more than you think
If your Lens recaps come out cluttered with reasoning preambles or hallucinated detail, the fix is almost always upstream of Lens. A field guide to picking a model and configuring LM Studio for clean daily snapshots.
Read moreFour frontier models, one year of office sensor data, four very different dashboards
I gave a year of Adafruit Clue environmental data to Gemini, GPT-5.4, GPT-5.5, and Claude Opus 4.7. Same prompt, same dataset, four very different answers — and only one of them asked for more data.
Read moreThe move was upstream
I spent most of a day in a stalled debug loop with one AI model before opening a fresh session with another. The lesson wasn't about which model is smarter — it was about going upstream.
Read moreThe race nobody else is running: why personal AI belongs on hardware you already own
Apple just put two hardware engineers at the top of the company. The cloud labs are quietly losing money on their best customers. From inside a hybrid practice, here's what the local share looks like and why it's growing.
Read moreA general-purpose MoE multimodal beat every dedicated vision model on my father's handwriting
I assumed a specialized vision model would win. I was wrong. A head-to-head on a hard handwriting corpus ended with the general-purpose MoE on top.
Read moreRunning real models locally on a Mac Studio that isn't new anymore
How I run a multimodal LLM on four-year-old hardware to read a family archive without sending anything to the cloud.
Read moreWhat family archives are for, now that the AI can read them
I have three collections of family letters spanning a century. Until recently, reading them properly would have taken years. Now it takes an afternoon.
Read moreThe Memex Has Been Waiting 80 Years for This Moment
Vannevar Bush described the memex in July 1945. It took 80 years to become buildable, because it required two things that didn't coexist until now: cheap private AI, and the conviction that data ownership is the last real moat.
Read moreFrom Tools to Framework: What a Year of Agentic Development Actually Looks Like
Frameworks are inevitable. They emerged for Ruby, Python, and PHP. Now they're emerging for agentic development. Here's the one that evolved on my workbench — and what it taught me about working with AI.
Read moreContext Architecture: Building the Library Your Agents Actually Need
AI agents are only as good as the information they can find. Context architecture is the skill of building structured environments where agents reliably retrieve exactly what they need.
Read moreSpecification Precision: The Most Valuable AI Skill Isn't Technical
The highest-demand AI skill in 2026 isn't coding. It's writing instructions so precise that a machine can't misinterpret them.
Read moreThe System That Built Itself
Six months of surviving as a one-person engineering team produced something I didn't expect: a layered AI operating system that grew organically from daily necessity.
Read moreEliminating Waste in the SDLC
When AI can use the same tools you use — Jira, GitHub, Sentry, GCloud — everything about the software development lifecycle changes. Not because AI writes the code. Because it eliminates the waste around the code.
Read moreWhispering to the Machine: Take Two
Eight months after writing about AI-powered development, the landscape has changed so dramatically that the original piece reads like a dispatch from another era. Here's what the work actually looks like now.
Read moreThe One-Person Engineering Team
The role of the software engineer is changing. What happens when one person with AI agents can cover the ground that used to require a full team?
Read moreWhispering to the Machine: A Snapshot of AI-Powered Development in 2025
What collaborating with AI in software development actually looked like in mid-2025 — the workflow, the lessons, and the mindset shift from writing code to orchestrating it.
Read moreFrom idea to IoT data: building an environmental monitor with Adafruit Clue and Python
Building a CircuitPython + Python environmental data logger on the Adafruit Clue — from BLE ambitions to a USB Serial pivot to a pywebview gateway with live charts. Originally posted on officeofadamthede.com.
Read more