Ideas Worth Exploring

Current PKI infrastructure in defence and logistics supply chains assumes RSA/ECC remains computationally hard. What does a practical migration path look like when the adversary has a near-term quantum capability? The problem isn't the algorithm — it's the operational transition.

Automation platforms that run tasks and leave recoverable state create persistent attack surface. Can we design execution environments that process sensitive workflows without retaining any recoverable artefacts — and what does that cost operationally?

When an automated system makes a consequential decision, who is accountable? In high-consequence environments — procurement, logistics, access control — every automated action needs a tamper-evident record. How do you build that without the audit system itself becoming the attack surface?

Biological systems repair under load without full shutdown. Industrial and defence logistics chains typically don't — failure cascades. Are there structural patterns from systems biology that translate into practical process design for high-tempo operational environments?

Cloud AI is unusable in classified or air-gapped environments. But modern local models are increasingly capable. What does a practical, security-audited, local-first AI stack look like for operational use — and where are the genuine capability gaps today?

Defence procurement is notoriously complex, slow, and expensive. Most improvement attempts add more process. What if the approach was purely subtractive — identifying the minimum viable constraints and stripping everything else? Has it ever been done? Does it work?