Periodic Table of Cognition (kk.org)

The author proposes a “Periodic Table of Cognition” analogy: just as early electricity research mixed correct insights with profound misunderstandings until atoms and the periodic table clarified matter, our current AI work mixes useful engineering with an incomplete, likely wrong theory of intelligence. Using ChatGPT5Pro, the author sketches a prototype table of 49 cognitive “elements” organized into families (e.g., Perception, Reasoning, Learning) and arranged by stages in a thought cycle (sensing → reflect & align). The chart is color-coded to show maturity (red = robustly synthesizable, orange = scaffolded, yellow = promising research), and example primitives include long-term memory, spatial awareness, deduction, planning and pattern perception. This framing matters because it reframes intelligence as an emergent, multi-dimensional compound rather than a single scalar quantity. That has practical and scientific implications: we should expect many engineered minds with different combinations and strengths of primitives, emergent behaviors that are hard to visualize or predict, and a lag between practical capabilities (LLMs used today) and a falsifiable theory. Building a taxonomy of cognitive primitives would enable clearer evaluation, targeted research, safety verification, and theoretical progress—just as knowing chemical elements enabled predictive physical science. The piece argues we’re at an early, electricity-era stage: useful inventions outpacing deep understanding, and much of cognition’s true structure still to be discovered.