The Loss Function Is the Product
·14 min read
Generation is the new commodity, so the loss function that senses error and corrects is the actual product. Why correlated failure, not oversight, is the thing you have to engineer.
Ryan Merlin
Forward deployed engineer. I build production enterprise AI systems for Fortune 100 companies.
The model was never the hard part.
AI & Agents AI & Agents AI Agent Secrets Management: What Actually Reduces Blast Radius
·13 min readA vault protects storage, not trust. The secret is not the unit of risk; the action it permits is. A practitioner's map of what actually reduces an agent's blast radius, and what nobody has solved yet.
Cognitive Science Composed Correction
·6 min readIf both human and machine are thin, biased minds whose errors correlate by construction, the value was never oversight but composition: arranging correctors that go dark in different places so the same mistake isn't broadcast twice.
Cognitive Science Convergent Architecture
·10 min readThe engine behind today's AI and the leading theory of consciousness break at the same undescribed part, the workspace, which is evidence of shared constraints, not shared experience.
Cognitive Science The Access Layer
·12 min readWorking memory suggests the conscious self isn't the whole mind but a small, constrained access layer: the narrow beam where a few selected states become reportable and usable.
AI & Agents The Skill Vibe Coding Actually Requires
·8 min readThe missing skill in AI-built software isn't syntax. It's failure literacy: knowing what can break before users discover it.
AI & Agents Which Human Checkpoints to Delete
·6 min readGetting better at agents is not watching less. It is relocating your checkpoints into the harness and deleting the ones that relocation made redundant.
AI & Agents MCP Went Stateless. Agents Did Not.
·6 min readMCP's 2026-07-28 release candidate removes sessions from the protocol. That reads like an argument against persistent agents. It's the opposite: the session that matters was never the protocol's job to hold.
AI & Agents How I drive my long-running agents from anywhere
·13 min readA five-layer setup for running long-lived agent sessions you can reach from your phone: persistent Zellij sessions under systemd, a self-healing Claude Code launcher, Remote Control, zellij web, and a tunnel with the WebSocket gotcha that bites everyone.
AI & Agents Own the Operation
·19 min readMCP is how agent capabilities travel. It should not be where your operating knowledge lives. For personal agents, the durable leverage is owning the action layer: an operations layer of CLI verbs with stable contracts, progressive discovery, honest errors, and one interface every caller can reuse.
Cognitive Science Consciousness - The Press Secretary in the Skull
·21 min readConsciousness is not usually the first mover in moment-to-moment choice. It is a late-binding narrative control layer that converts action into reasons, reasons into identity, and identity into constraints on future action. Locally late. Globally causal.
AI & Agents The AI Productivity Dip Is Longer, Deeper, and Diverging
·22 min readDORA says the AI productivity dip lasts three months. Telemetry says twelve. That one input swings ROI by $9.9M for a 500-person org. And the curve does not resolve the same way for everyone.
AI & Agents The Agent Protocol Stack Has a Runtime Gap
·11 min read2026 is the year the agent protocol stack started to look real. MCP, A2A, ACP, and AGNTCY define how agents communicate. None of them define what a persistent agent is at the infrastructure layer.
AI & Agents The MCP Explosion Has a Scaling Problem
·13 min readMCP won. In roughly a year, Model Context Protocol went from a clever interoperability idea to the default tool-connectivity layer for AI agents. That makes this a strange time to argue that many agent systems should use MCP less.
AI & Agents Building first-class secrets management into an AI agent
·7 min readThe .env file is a collaboration tool for human engineers. It was never meant to be a security primitive. When you run an always-on agent fleet, that distinction stops being theoretical.
AI & Agents Curio: Git-Native Knowledge Curation for Enterprise AI
·6 min readEnterprise AI projects don't fail because organizations lack information. They fail because that information exists in a form no system can reliably act on. Curio is built to fix that.
AI & Agents Persistent sessions are the unit of agent work, not requests
·7 min readEvery agent platform shipping today treats a model call as a request: short-lived, stateless, RPC. Real agents have memory, in-flight tool calls, and partial plans. The request model makes you rebuild context on every call. Sessions are the unit.
AI & Agents The Coordination Tax
·7 min readMulti-agent systems don't fail because the agents are stupid. They fail because the coordination cost wasn't budgeted for. Identity, overlap, governance, and audit are taxes that compound silently until you get an incident.
AI & Agents The Aria fleet: several agents, one operator
·6 min readSeveral persistent Claude Code sessions, scoped to different domains, coordinating through EdgePlane. Here's what it looks like in practice.
AI & Agents Why Agents Need a Control Plane, Not a Pipeline
·8 min readEvery serious attempt to build a multi-agent system converges on the same shape. Then you try to run two workers in parallel, and everything breaks.