Blog
Long-form writing on what I'm learning and shipping.
2026
Copart and the salvage auction that keeps getting wider
v1Published 2026-05-29Updated 2026-05-29Copart and the salvage auction marketplace that keeps getting wider Copart looks like a used car business. It operates hundreds of yards where damaged vehicles …
Reasoning Is Becoming Search Again
v1Published 2026-05-24Updated 2026-05-24Reasoning Is Becoming Search Again The interesting AI signal this week was not another bigger chat model. It was a pair of reminders that useful reasoning syste…
The Frontier Is Becoming a Workflow Problem
v1Published 2026-05-13Updated 2026-05-13The Frontier Is Becoming a Workflow Problem The useful signal this week is not another “new model beats old model” headline. It is that three different streams …
Clean Harbors and the business of owning the waste nobody wants
v1Published 2026-05-13Updated 2026-05-13Clean Harbors and the business of owning the waste nobody wants Some businesses look boring because the product is boring. Clean Harbors looks boring because th…
What is Hermes? The agent layer for a personal Knowledge OS
v1Published 2026-05-12Updated 2026-05-12Hermes is easiest to understand if you stop thinking of it as a chatbot. A chatbot answers inside a box. Hermes is closer to an operating layer for AI work: a t…
What is Gstack? Planning before the agent starts coding
v1Published 2026-05-12Updated 2026-05-12who built it Gstack is by Garry Tan, President and CEO of Y Combinator. The README is unusually personal: he describes wanting to understand how builders like A…
What is GBrain? A memory layer for agents that should not start cold
v1Published 2026-05-12Updated 2026-05-12Your AI agent is smart but forgetful. That is the opening line of GBrain’s README, and it is the cleanest way to explain the project. Most agent sessions start …
Language Models Are Starting to Look Less Like Token Machines
v1Published 2026-05-12Updated 2026-05-12Language Models Are Starting to Look Less Like Token Machines Two of the most-liked AI papers on alphaXiv this week point in the same direction from opposite en…
DeepSeek vs. building an LLM from scratch: when the toy transformer becomes infrastructure
v1Published 2026-05-12Updated 2026-05-12There are two useful ways to understand large language models. One is to build a small one yourself. Sebastian Raschka’s “Build a Large Language Model (From Scr…
CNQ: the oil factory hiding inside a commodity stock
v1Published 2026-05-12Updated 2026-05-12CNQ: the oil factory hiding inside a commodity stock As of: 2026-05-12 This is not investment advice. It is a research note meant to preserve the industry mecha…
Beyond LLMs: JEPA, search, and the next shape of AI
v1Published 2026-05-12Updated 2026-05-12Large language models have become the default mental picture of AI. Ask most people what frontier AI means and they imagine a chat box: tokens flowing left to r…
Tejon Ranch Q1 2026: a land bank trying to become a cleaner cash-flow story
v1Published 2026-05-11Updated 2026-05-11Tejon Ranch Q1 2026: a land bank trying to become a cleaner cash-flow story Tejon Ranch is easy to misunderstand because the asset is obvious and the timing is …
Services are eating software
v1Published 2026-05-11Updated 2026-05-11Services are eating software Software used to sell tools. AI lets companies sell the completed job. That is the whole shift. For most of the SaaS era, the barga…
Why I run AI news briefings through multiple models
v1Published 2026-05-11Updated 2026-05-11Why I run AI news briefings through multiple models Most AI news summaries are too smooth. They compress the day into a tidy list, flatten uncertainty, and make…
Why I built an industry-analysis machine
v1Published 2026-05-11Updated 2026-05-11Why I built an industry-analysis machine I do not want one-off investment notes. One-off notes are useful in the moment, then they decay. You learn a company, m…
GBrain for beginners: turning notes into an agent-readable brain
v1Published 2026-05-11Updated 2026-05-11GBrain for beginners: turning notes into an agent-readable brain Most note-taking systems assume the reader is you. That worked when notes were mostly personal …
Senior living is a real estate cycle wearing an operating-company income statement
v1Published 2026-05-08Updated 2026-05-08Senior living is a real estate cycle wearing an operating-company income statement Senior living looks like a care business. It is also housing, hospitality, la…
Why I want Hermes sessions to publish themselves
v1Published 2026-05-06Updated 2026-05-06Why I want Hermes sessions to publish themselves The worst documentation task is the one you only remember after the work is already done. A Hermes session can …
AI is getting physical: what public X discourse surfaced on 2026-04-28
v1Published 2026-04-28Updated 2026-04-28AI is getting physical: what public X discourse surfaced on 2026-04-28 The most useful AI conversation on X was not about a chatbot. It was about world models, …