Why RAG will outlive the context window debate
Longer context doesn't make retrieval irrelevant — it changes what retrieval is for.
← Portfolio
Thinking
Essays & observations.
Arguments on AI systems, engineering decisions, and the edges of what machines can do. Specific positions, not trend summaries.
The evaluation theatre problem in LLM systems
When the model can write its own tests, passing tests stops being evidence of correctness.
Agentic systems need failure modes, not just success paths
Every agent decision that can't be reviewed is a gap in your production architecture.
What hybrid retrieval actually means (and when BM25 still wins)
Dense embeddings feel like magic until you need to recall a specific invoice number.
Prompt engineering is not a job title
The skill isn't writing better prompts. It's knowing when the prompt is the wrong abstraction entirely.
Context is not memory
Treating the context window like a conversation history is the first mistake most teams make when building agents.
Structured outputs require structured thinking
JSON mode doesn't save you from an underspecified schema. It just makes the failure more consistent.
Set a latency budget before you pick a model
Most teams choose the most capable model and then discover their latency requirements. That's backwards.
Ship evals before you ship features
You can't improve what you can't measure. In LLM systems, measurement is the hard part.