Synthetic Philosophy and Deductive Engineering
This document is a sketch of a development strategy for SPaDE, which is oriented towards enabling AI to contribute effectively at all levels of the project, from the highest levels of philosophy and architecture down to the foundations, code, testing and application. Speed is important, so an important aim of the strategy is to identify those aspects of the continuing development which will have the greatest impact on accelerating progress.
An important element of the strategy is to identify and prioritise the development of key singularities in the focal tower, which will enable rapid advancement through focal methods targeted at these singularities. Some further articulation of the focal stack is therefore desirable and will be linked to from this document when available.
There is a bootstrap problem, the problem of getting this project off the ground. I’m going to take that as the problem of getting to the point at which the deductive intelligence component is in a position to work continuously (24/7) on the first singular focus. This does of course depend on the other components advancing sufficiently to provide an appropriate context and support for deductive intelligence to get on with the job. This includes advancing the collaboration with not focal AGI which will contribute to the work both by work on the SPaDE repo and by interaction with the SPaDE system through the MCP server interface.
SPaDE is like no other project I have ever undertaken, and therefore the methods at every level are incompletely thought out and poorly documented.
A crucial and dominant aspect of the project is that it is intended to be developed with the help of artificial intelligence, and therefore the methods will be evolving as I learn how to use AI effectively and as the AI itself evolves. It is relatively easy to get agentic AI to code and test from detailed design documents, but it is essential to this project that as soon as possible AI is effectively contributing at all levels of the project, from the highest levels of philosophy and architecture down to the foundations, code, testing and application. Whether or not that is achievable with the AI currently available to me is moot, but making the most of what we now have is part of what it takes to be ready for the future when more powerful AI is available.
There is a very great distance between the high level philosophical vision and the code which will implement the various subsystems of SPaDE. Any amount of time could be spent on articulating and elaborating the philosophical aspects of the project, which determine but do not materially progress the aims of the project.
In order to enable a balance to be struck in progressing the different levels, it will be useful for me say a little about my thinking in putting them together into the SPaDE project.
I am a retired software engineer, with a lifetime interest in philosophy, logic and the foundations of mathematics, and in artificial intelligence. I have long wished to progress foundational approaches to artificial intelligence.
Having retired and wishing to spend my time productively, I have proceeded in the following way. I have sought to understand the directions that the future may take so that I can play into the more desirable trajectories, on the off chance of making a difference.
Life on earth has taken over 3 Billion years to reach human level intelligence. It is not unreasonable to consider how intelligent systems will evolve over the next billion years. On these timescales evolution is the driving force, even though the characteristics we associate with the evolution of intelligence in earth are now being disrupted by the intelligence which has evolved.
SPaDE is intended as a key resource underpinning the proliferation of intelligence across the universe. Its advancement through focal methods targeted at key singularities in the focal stack is essential to success in this aim.
Rapid advancement depends on effective exploitation of the capabilities of agentic AI, in facilitating the continuous application of focal AI through continuous deductive exploration of the relevant formal theories, 23/7/365, primarily by the the deductive intelligence component of SPaDE. Before this can begin, we need a functional knowledge representation subsystem, key elements (yet to be identified) of the deductive kernel, key foundational reflexive theories and meta-theories, and a first-cut, layered alpha-zero approach to deductive reasoning.