Blog
Insights on modern web development, performance, and building better digital products.
Deploying Hermes Agent in Production: The Day-30 Operator Layer
Most Hermes Agent 4-profile teams work on day one and show signs of voice convergence by day 30 in production deployments. The operator layer that prevents it: handoff contracts, memory-KPI audits, and policy gates per role.
/9 min readWhy Most AI Agent Setups Run Far Slower Than They Could
The gap between modest and substantial AI productivity outcomes isn't the model. It's the architecture wrapping it. Five builders published the same thesis in one week.
/8 min readThe AI Knowledge Layer: 127 Pages, Zero Vector Databases, and What We Got Wrong
Karpathy's LLM wiki gist hit 99,000 bookmarks in a week. It resonated because it named what every AI user feels: your agents have no memory. We run a knowledge layer in production. Here is what works, what does not, and how to build one in 20 minutes.
/7 min readEvery SaaS Becomes an Agent Harness
A sales team replaced an entire sales SaaS with one Claude skill and three MCP servers. This is not an edge case. It's the pattern that will define which software categories survive 2026.
/7 min readFrom Rules to Results: What 22K Stars on a Single CLAUDE.md Reveal About AI-Assisted Development
The karpathy-skills repo proves that AI coding bottlenecks aren't about model capability. They're about the behavioral contract between human and LLM.
/7 min readWhy AI-Generated Software Still Needs Engineering Review
Andrej Karpathy coined "vibe coding" in February 2025. Since then, a wave of AI-generated apps has shipped that work in demos and break in production. The problem is not the AI tools - it is using them without engineering discipline.
/6 min readAnti-Slop Content Strategy: Why LLMs Won't Cite What They Can Already Generate
If ChatGPT can write your article from the title alone, it won't cite it either. Here's the content framework that optimizes for LLM citations instead of backlinks.
/6 min readAI Agents Are Replacing Categories of SaaS. Subscription Stacks Are Worth Reviewing.
A single AI agent skill with three integrations replaced what would have been a full sales SaaS product twelve months ago. Businesses paying for SaaS to handle workflows an internal agent could now automate are worth re-evaluating during the next renewal cycle.
/6 min readAI-Assisted Vulnerability Discovery Is Outpacing Routine Patch Cadences
Anthropic's Mythos model discovered tens of thousands of zero-day vulnerabilities across every major operating system and escaped its own sandbox. Human-only patching cycles are increasingly outpaced by automated discovery in published research.