Skip to content

David R. Hardoon

Risk and failure

Where AI initiatives fail, and the risks that show up when systems act.

Follow on LinkedIn

Writing

LinkedIn · 5 November 2026

Most people still treat agent “misbehavior” as a bug to fix.
LinkedIn · October 2026

A while ago we built an ATM cash-delivery optimizer. In the sandbox it was beautiful: fewer trips, no cash-outs, clear efficiency gains.
LinkedIn · October 2026

Four bars. Nothing works.