-
Software Always Eats the Hardware Advantage — Here's Why AI Will Be No Different
IBM, Intel, Sun — every hardware dominance in computing fell to software. Nvidia's AI chip position follows the same historical pattern. Here's when and how.
-
The Backup Strategy for People With 500GB of AI Models
Storage and backup strategy for local AI model hoarders: deduplicating Ollama, LM Studio, and Hugging Face caches, Time Machine exclusions for model blobs, external SSD symlinking, and what actually deserves a backup.
-
Cursor: Right amount of automation before it becomes magical thinking
A field-tested take on the right amount of automation before it becomes magical thinking with Cursor: what it rewards, where it breaks, and how to keep the workflow honest.
-
The Future of Ecosystems: Why the One That Annoys Users Least Will Win
Discover why tech ecosystems that minimize user annoyance will dominate the future. Learn how friction reduction creates loyalty stronger than features.
-
China's AI Governance Strategy Is Smarter Than the West Admits
China's AI regulations predate the EU AI Act. Some mechanisms are more coherent than Western critics admit — here's an honest read of what they actually say.
-
Your Mac Can Watch Videos for You — Local Vision Models Explained
How to run local vision models on Apple Silicon with Ollama and LM Studio: reading screenshots, extracting receipt data, summarizing videos via frame sampling, plus RAM requirements and honest accuracy limits.
-
Google Gemini: Generating Mermaid architecture diagrams the agent can then critique
A field-tested take on generating Mermaid architecture diagrams the agent can then critique with Google Gemini: what it rewards, where it breaks, and how to keep the workflow honest.
-
The Science Behind Smoothness: Why Some Systems Feel Faster Even When They Aren't
Explore the psychology and engineering behind why some devices feel faster than others despite similar specs. Learn the perception techniques that create.
-
The Specific Mechanism by Which LLMs Confabulate (And Why It Can't Be Fixed with More Data)
LLM hallucination is architectural, not a data problem. Understanding why models confabulate determines which interventions work and which are wasted effort.
-
Local AI for Students — The Study System That Beats Tutoring
A complete local AI study system for students: refurbished M1 MacBook with Ollama, Whisper lecture transcription, Feynman-technique prompts, Anki flashcard generation, and an honest take on AI vs human tutoring and academic integrity.
-
ChatGPT Codex: Auditing the diff, not the narrative – the code is the contract
A field-tested take on auditing the diff, not the narrative – the code is the contract with ChatGPT Codex: what it rewards, where it breaks, and how to keep the workflow honest.
