Supersoftware: software that is itself symbolic AI

Today’s software paradigm emerged during the pre-networked era, to say nothing of the LLM age. ‘The network is the computer’ remains more industry slogan than reality. While the current paradigm has been inadequate in its own right for years, its inability to meet the demands of LLMs let alone help realise their full value makes the problems all the more palpable. And costly.

We introduce a new paradigm — software that is itself symbolic AI, poised to support intelligence between not merely within. Reactive and reflective symbolic systems are capable of inspecting and modifying their own form, and so help govern their own behavior. The simplified, behaviour-free model system brings significant advantages in terms of speed, resilience, adaptability, and security.

We call it supersoftware.

Supersoftware provides the missing half of the AI equation. The LLM provides intuition, intent, and context, and supersoftware provides a world model, grounding, validation, security, and architectural guarantees. LLMs generate accurate, explainable, and policy-compliant responses when guided and constrained by the supersoftware's verifiable logic. Supersoftware supports the requisite complexity with unparalleled simplicity — a true symbiosis.