Posts

All the articles I've posted.

• A confidence score is not a reason

  21 Jul, 2026·45 min read

  An extraction confidence is not an adverse-action reason. Document Intelligence returns two scores, and the gate reads the wrong one.

• The permission model was built for people holding jobs

  7 Jul, 2026·52 min read

  An HCM grants permissions to people holding jobs, and an agent is neither. Least privilege with no native expression, and the primitive that just shipped.

• Nobody owns the join between two agent platforms

  23 Jun, 2026·40 min read

  Four vendor agent platforms in one mesh, each sound alone. The seven joins that failed, starting with three catalogs where the loosest one decides.

• The LLM judge is a model too

  16 Jun, 2026·32 min read

  The judge is a model too, with the same length, position and self-preference biases. Measuring it, versioning it, and knowing when not to use one.

• Containment is measured on the calls that never needed you

  9 Jun, 2026·36 min read

  Containment counts the sessions that never needed you. A booking agent's rollback is a refund, and a surge multiple is a scalar over a rotating mixture.

• Throughput per dollar, once the traffic is bursty

  26 May, 2026·23 min read

  One roofline number sets the speculative-decoding gate, the batch-size knee and the chunked-prefill floor. Why max-num-seqs is a latency control.