Avoiding Cyber Catastrophe

The Key Issues for Cybersecurity

1. The "Original Sin" of von Neumann Hardware: Every modern computer relies on the general-purpose von Neumann architecture, created in the 1940s. This architecture relies on shared memory, blind trust, and unfair hardware privileges (such as the all-powerful "superuser" or root account). When John von Neumann designed his framework, he could ignore the scientific principles of Alonzo Church’s lambda-calculus. This shortcut created hardware privileges and vulnerabilities that led to ransomware, malware, and hacker exploits.
2. The Illusion of Software Security ("Bailing the Boat with a Sieve"): Modern cybersecurity relies on an endless cycle of patching software, adding antivirus programs, and setting up firewalls. This is like bailing a boat riddled with structural flaws: software wrappers and AI safety monitors run on top of the insecure hardware substrate they try to police. Today, software patches are increasingly complex, costly, and burdensome, yet they fail to seal the foundational hardware breach.
3. The Hardware Solution: Capability-Limited Computing & Church Machines: True security cannot be patched in software; it must be built into silicon on bare-metal hardware using  Capability-Limited/Object-Oriented/Machine Code. In the CLOOMC model, every instruction requires an unforgeable "Golden Token", the key to unlock and authorise the execution. This paradigm was proven in 1972 with the Plessey System 250 (PP250) telecommunications computer.
4. Metaphors for Perfect Computer Science: To explain how secure systems function, CLOOMC use three core metaphors:
   • Digital DNA: Software constructed with immutable "capability keys" that define an explicit, unchangeable blueprint, a living structure for what code can run and when.
   • The Babylonian Golden Rule: Tools like the Abacus where "form matches function" perfectly, eliminating extra, exploitable features ("corruptible baggage")
   • The Coiled Spring: Software actions designed as self-contained lambda-calculus function abstractions that release predictable, contained energy only when properly triggered on an approved task.
5.  The "Cybercrime Tsunami": Dire economic and societal risks exist while hardware remains unchanged. Global cybercrime costs are projected to reach $56.2 trillion by 2037 (rivalling the entire US GDP). 
6. AI Threats: Simultaneously, autonomous AI agents and state actors pose severe risks regarding bio-weapons, cyber espionage, and systemic infrastructure disruption.

---

What It Means

• For Technology & Engineering: Security must shift from an afterthought added via software to a physical invariant enforced by hardware, where unauthorised code is mathematically unnameable and unexecutable
• For AI Safety & Governance: Regulatory mandates, voluntary slowdowns, and software alignment checks are surface-level "theatre" unless AI models run on capability-constrained hardware that prevents unauthorised ambient authority.
• For Society & Democracy: Current computing paradigms push society toward centralised monopolies and an "Orwellian digital dictatorship" controlled by a few tech suppliers. Rebuilding hardware on scientific principles is presented as a necessary step to protect individual freedom and enable true digital democracy.