On May 31st, I presented a talk on "Evaluating and reasoning with and about GPT", and Arun Majumdar presented a demo that shows how the technology developed by Permion.ai LLC supports those methods.. These methods also show that ideas developed by Charles Sanders Peirce, especially in the last decade of his life, are still at the forefront of some of the latest developments in AI. The Permion technology uses tensor calculus for relating the Large Language Models (LLMs) of GPT to more traditional computational methods of AI (logic, conceptual graphs, computational linguistics, neural networks, and statistics) . With these methods, the Permion software can use GPT for its ability to translate languages (natural and artificial) and to retrieve large volumes of useful data from the WWW. By using conceptual graphs and formal deduction, Permion can detect and avoid the errors caused by the loose or nonexistent reasoning methods of GPT while taking advantage of its useful features for processing languages. For the slides by John Sowa, see EvaluatingGPT--JohnSowa_20230531.pdf (ontologforum.s3.amazonaws.com) For the Audio-Video recording of both talks, see https://ontologforum.s3.amazonaws.com/General/EvaluatingGPT--JohnSowa-ArunM… John