Prompt: "reason step by step"
text-to-image
Turning research into products people actually use.
Picked things I'm genuinely into.
Thoughts on AI, leadership, and where this is all going.
The tools I reach for every day.
Why LLMs sometimes print tokens like [control_28] mid-conversation, why it's a serving stack problem rather than a model problem, and why the model can't recover from it on its own.
An interactive essay on tabular and relational foundation models, why token streams lose table structure, and how TabFM and KumoRFM read it natively.
Enterprise AI's hidden cost is the human labour of verifying and correcting AI output. The fix is a semantic and governance layer that gives models real context.
