GPT-5.6 Luna: the model that became my default in Codex
After nearly 10 billion tokens in Codex, why GPT-5.6 Luna became my default model: cost, performance, context, and limits.
Term
Llm
Ontologies for AI agents: the missing contract between model, memory, and tools
LLMs can converse, but they do not share a stable domain model. Ontologies, knowledge graphs, and semantic validation can give agents a common vocabulary, safer actions, and verifiable memory.
Meta Harness: the missing layer in agent engineering
The agent is no longer the whole product. The next jump is the layer above it: memory across sessions, coordination across agents, context across repos, and automatic optimization of the harness itself.
Domain-specific agents: composition against the bloated agent
The generic agent looks simple until it becomes a giant context window with too many tools, too many permissions, and too much cost. Domain-specific agents trade that accumulation for composition, clear boundaries, and cheaper execution.
Evaluating AI Agents Beyond the "Vibes Check": How to Measure What Actually Matters
Your agent nailed the demo and everyone loved it. But how do you know it actually works? If the answer is 'we tested it and it seemed fine', you are operating in vibes mode. And vibes don't scale.
Why Your AI Agent Is a Compulsive "Spender": The Truth About Token Consumption
Autonomous agents can turn a trivial task into a financial black hole because token consumption is massive, stochastic, and difficult to predict.
Why Does Your AI Fail? The Secret Is Not in the Model, But in the "Agent Harness"
You have probably experienced this: one moment, GPT or Claude solves a complex coding problem in seconds; the next, the same AI forgets basic context or invents nonexistent information.
The Illusion of Sequence: Why Hierarchy Is the Next Leap in AI Inference
There is an efficiency gulf between biological architecture and silicon. While a 12-year-old child already masters human language, models like GPT-3 require far more data.
Taking AI Engineering to the Next Level with RAG
If you are starting to venture into AI Engineering and want to go beyond the basics, you need to deeply understand what **Retrieval-Augmented Generation (RAG)** is.
Multi-Agent AI Systems: When to Use Them, How to Build Them, and Why So Many Fail
You know when you are trying to solve a complex problem with AI and it feels like a single model cannot handle the job? Multi-agent systems help, but only in the right cases.