7 min read
DeepSeek V4: How They Crushed the Million-Token Memory Problem
A million-token context window sounds impressive. Making it cheap enough to actually use is the hard part — and the engineering behind it is worth understanding properly.
Tag
ai
4 posts
2 min read
The Local Model Plus Outsourcing Bet Against Frontier API Pricing
Working through the cost architecture and where the latency math gets messy.
3 min read
The AI Coding Metric Nobody Has Actually Measured
Nolan Lawson's post made me realize why 'better code' never got the marketing that 'faster code' did.
3 min read
DeepSeek Reasonix and the Economics of the Agent Prompt Loop
Why 'native' KV-cache design in a coding agent makes context ordering a first-class engineering constraint, not an afterthought.