Hi, I'm Sean Moran.

What attention maps, residual streams, and hidden states can tell you about a transformer's internal computation, and the limits of current mechanistic-interpretability techniques.

Practical lessons from time-series forecasting: split design, leakage, calibrated baselines, drift detection, and the recurring failure modes of naive models.

A working notebook on Claude Code in 2026 — hooks, skills, sub-agents, plan mode, and the workflow patterns that have proven useful in practice.

A reading of the released OpenAI emails as a record of how AGI strategy looked in the year before the transformer architecture reshaped the field.

The 1952 statistical model behind chess ratings, A/B test ranking, and the reward model in RLHF — derived from first principles with a working implementation.

Two years of Y Combinator batches read as a signal about the AI market: which categories are growing, which are commoditising, and where founders are positioned against consensus.

An argument that current LLM-agent frameworks are at an early, exploratory stage analogous to phone phreaking — and a sketch of the more durable engineering patterns underneath them.

A prototype shipped end-to-end by an agent while I was at the coffee shop — what the IDE becomes when the inner loop runs without you, and what the developer's job becomes instead.

A walkthrough of the computational "circuits" that emerge inside transformer attention layers, with a hand-traced example of how a model performs multi-step reasoning step by step.

A retrieval evaluation metric that compares more favourably than accuracy or recall for ranking RAG and agent systems, with implications for how to choose between embedders.

A common failure mode in production RAG systems: queries whose answer cannot be reached by dense similarity alone, and the hybrid retrieval patterns that address it.

Self-evaluation, uncertainty calibration, and abstention as first-class design choices in multi-agent systems, and how each one contributes to operational reliability.

An analysis of the 17× error-compounding effect in "bag-of-agents" architectures and the structural choices that limit it.

A close reading of the NeurIPS 2025 Best Paper on attention gating: what the ablations show, the behaviour of attention sinks, and the implications for inference cost.

An introduction to the Distance-Dependent Chinese Restaurant Process for nonparametric clustering of deep embeddings, for settings where the number of clusters is not known in advance.

How to keep your engineering judgment intact when an AI assistant is doing more of the writing — a personal "internal KL divergence" check for when you're drifting from your own thinking.

A historical thread from Victor Lavrenko's relevance models through cross-encoder reranking to modern attention, drawing out continuities between classical IR and the transformer architecture.

Kahneman's "Noise" applied to AI-assisted decision making — where calibrated models reduce variance, where they amplify it, and how to tell which one is happening to you.

Pairing an LLM with a symbolic proof assistant to do mathematics that neither side could do alone — what the architecture looks like and where it still fails in surprising ways.

The JPMorgan engineering blog on Senatus — an internal code-to-code recommendation engine built on De-Skew LSH that scales sub-linearly across very large codebases.