
I Audited a Week of 'AI Productivity' — What Actually Saved Time
A rigorous time-and-error log of AI in real knowledge work — separating genuine wins from rework and the hidden review tax.

A rigorous time-and-error log of AI in real knowledge work — separating genuine wins from rework and the hidden review tax.

When a RAG pipeline gives bad answers, the LLM usually isn't the problem. Retrieval is. Here's where chunking, ranking, and recall actually break.

Stuffing a million tokens into a prompt degrades reasoning more than it helps. The real skill is curating what the model actually attends to.

A practical decision framework for choosing retrieval, full-context, or hybrid approaches based on data volatility, cost, and accuracy.

Email threads encode decisions, commitments, and relationships better than any wiki — making your inbox the highest-signal grounding source for an AI assistant.

Pure vector search over personal files stumbles on recency, permissions, and exact terms. Hybrid retrieval plus metadata is the fix.

Frontier models aren't always the answer. For inbox and search work, routing to small fine-tuned models is quietly becoming the default architecture.

Memory isn't one feature. A practical breakdown of episodic, semantic, and working memory for AI agents — and how to wire them into real workflows.

A diagnostic framework for the quiet retrieval failures that degrade RAG quality — from chunking strategy to embedding mismatch.

Retrieval failures aren't one bug — they're three. A diagnostic framework for isolating chunking, embedding, and reranking problems instead of guessing.

Bigger context windows don't guarantee better recall. Here's where models actually lose information — and how to structure prompts so they don't.

Why keyword search fails in modern workspaces, and how semantic search, metadata, and permission-aware retrieval combine to make file search actually usable.
Get started
Claim your address before someone else does — free to start, with an AI-native inbox built in.