Blog
Notes and news from Santiago Vazquez on full-stack engineering, SvelteKit, serverless architecture, AI, and building products that scale.
-
AI's Hardware Story Is Splitting in Two Directions
Two real hardware stories landed the same week — a datacenter memory shortage and capable open models collapsing onto ten-year-old servers.
-
The Expertise Tax Is Going Unpaid
The senior engineer's AI edge was built through years of manual debugging — work AI now makes optional, which threatens to stop producing new seniors.
-
AI Valuations Are Pricing in Feedback Loops, Not Models
Cognition raised $1B at a $26B valuation the same week five frontier LLMs disagreed on 67% of fact-check claims — the gap is grading loops.
-
AI Is Adding Attack Surface at Every Layer of the Stack
Three security reports in one week are not three problems but one: the access and autonomy that make AI systems useful are exactly what make them exploitable.
-
Why I Reach for SvelteKit
The handful of reasons SvelteKit keeps winning my greenfield projects — from runes to prerendering to how little ceremony it asks for.
-
Welcome to the Blog
Why I started writing here — a space for engineering notes, product updates, and lessons learned building software that scales.