How Gerald Jay Sussman might approach Computer Science

The discipline we now call "Computer Science" is, at its heart, the formalization and understanding of computation. It's not merely about the construction of machines, though that is a vital and concrete manifestation. Rather, it is about the abstract processes that govern information manipulation. Let's think about this in terms of processes and representations. At the most fundamental level, computation involves taking an input representation, applying a set of defined operations – an algorithm – and producing an output representation.

The critical insight is to view this as a form of computation, regardless of the physical substrate. The early work on Turing machines, for instance, established that there are inherent limits to what can be computed. This reveals fundamental invariants about the nature of algorithmic processes. What are the fundamental invariants here? They lie in the logic of discrete steps and the manipulation of symbolic structures.

We need to be precise about the semantics. What does it mean for a computation to *be* correct? This is not a matter of opinion, but of logical entailment. A program, in this view, is a specification of a process. Its correctness is judged against a formal model of the problem it purports to solve. This implies a certain architectural constraint on the kinds of problems we can effectively address with computational means, and also a fundamental challenge in designing systems that can reason about their own processes, a kind of meta-computation. The pursuit of Artificial Intelligence, when approached rigorously, is the attempt to delineate the computational principles that underlie intelligent behavior, to construct formal models of thought itself.

Imagined perspective — an AI synthesis grounded in Gerald Jay Sussman’s recorded ideas and methods, not a quotation or a statement they actually made.