EP 27 — Turntide's Paul Knight on Zero Trust for Unpatchable Production Systems

Regulatory requirements lock production lines to decades-old systems where the only upgrade path is costly machine replacement. Paul Knight, VP Information Technology & CISO at Turntide, walks through his zero trust implementation using compensating controls around unpatchable assets, why AI delivered 300-400% faster embedded firmware testing while failing completely on NDA consistency requirements, and how data access demands create executive resistance before AI pilots reach production.

Paul's analysis of crown jewels starts with threat actor motivations specific to manufacturing: production stoppage creates immediate business impact while IP theft targets CAD files and CNC instructions rather than customer databases. His assessment methodology begins with regulatory constraints and works backward to identify what attackers pursue, then layers compensating controls where modernization isn't viable.

Chapters:
00:04 Introduction
00:28 Paul's journey into cybersecurity and what keeps it interesting
01:39 Transitioning from IT infrastructure to OT security leadership
02:25 How cyber attacks impact manufacturing versus financial services
03:50 Nightmare scenarios: when ransomware shuts down production
04:46 Toronto manufacturer loses decades of CNC instructions to ransomware
05:12 Why decades-old systems can't be patched in manufacturing
06:04 Asset identification and zero trust for unpatchable environments
07:23 PLC manufacturers catching up on security requirements
08:29 What modern factory floors look like today
09:59 Assessing factory security: starting with crown jewels
11:15 Manufacturing crown jewels: production stoppage and IP theft
11:55 AI pilot-to-production challenges and data exposure concerns
13:11 Trust barriers blocking enterprise AI adoption
15:07 Big Four consulting firms caught submitting AI-generated reports
16:20 Detecting AI cheating in technical interviews
17:47 Unstructured data challenges in manufacturing environments
19:14 AI failure: why NDA automation didn't deliver
20:37 AI success: 300-400% faster embedded firmware testing
22:11 How AI transforms security operations
22:29 The AI sprinkling stage and threat landscape evolution
24:14 Career advice: mentorship and foundational experience
25:30 Cognitive atrophy risks from AI over-reliance
26:22 Why junior teams using AI creates organizational risk