Posts

List-based and process-based approaches to AI risk classification both fail in predictable ways. The failure mode depends on which you chose.

Every AI governance framework demands human-in-the-loop. Nobody does the maths on what that means at enterprise scale.

Most AI maker-checker implementations capture the correction but not the reason. That's a feedback loop with no signal.

The most useful reframe I've found for AI governance in financial services

The difference between protecting a commitment and hoarding optionality

When your model's labels come from human decisions rather than reality, you're not measuring what you think you're measuring.

Governance isn't about catching every failure — it's about proving your process was reasonable when one happens. The real skill is knowing what to deliberately not monitor.

It's not enough to say humans are in the loop. You need to show the loop is in the system.

Mathematical impossibility results are the best meeting-room weapons I know.

24MB download, <100ms per batch, nothing leaves the machine. For banks with air-gapped environments, this changes the conversation.

Your AI budget question isn't which model — it's which phase of the workflow needs depth, and which just needs speed.

What does the Frontier Alliance actually mean for advice quality?

Search returns summaries. Get returns detail. The model decides what to expand. 75% context savings.

Open-source agent adoption can outpace enterprise security controls by weeks. Governance teams need a policy before the agents arrive uninvited.

The most dangerous AI failures are the ones that look fine on the surface.

Every AI fairness debate is secretly a values debate disguised as a technical question.

Interaction, inference, orchestration, tooling. The boundaries between them must be enforcement points, not design principles.

MCP decouples the tool from the model. Once that happens, the durable asset isn't the model — it's which systems you've exposed.

The consulting question isn't how to build AI — it's how to get it past the 62% graveyard.

Adding detail to a deliverable doesn't fix credibility — it creates new interrogation targets.

A strand of multi-agent research — latent-space inter-agent communication — is thriving in Chinese labs and almost invisible in the West.

A 3-year AI roadmap designed around today's model capabilities may be solving last year's problem by year 2.

Monitoring systems need consumers before they need features

Most AI writing tools want to chat with you. The better model is annotation — AI that reads what you wrote and leaves margin notes.

The real risk of building a personal AI operating system isn't that a better tool appears — it's that your system's complexity becomes the thing you maintain instead of the thing that maintains you.

When AI agents can grep, read, and reason iteratively, most RAG infrastructure becomes unnecessary middleware.

When you have both a credential and real work in the same domain, route effort through the real work.

The best use of AI in investing isn't picking stocks — it's building the pipeline that teaches you why you can't.

Most people use AI transactionally. The real unlock is conversational — thinking with the model, not through it.

In a world where AI has perfect recall, the skill that matters is rebuilding frameworks from first principles — not memorising them.