Presenter
Robert McIntyre, Nectome Inc
CEO at Nectome Inc Robert comes from a background of artificial intelligence at MIT, where he studied under Marvin Minsky and Patrick Winston. After completing an award winning master’s thesis on embodied cognition, he turned his sights to the problem of cognitive science. He realized that it would be incredibly hard to fully scan all the information in a human brain, especially the connectome, but he did realize that locking all the structures of the brain in place for long term storage might be possible, by marrying ideas from two schools of thought- some from the field of cryobiology, and others from the field of connectomics in neuroscience. After getting up to speed with the state of the art, he developed with other a novel way to preserve brains, and won a prize in 2016. Soon after, he started Nectome with Michael to further this research and help realize the potential created by a robust brain preservation technology.
Summary:
Robert McIntyre discusses the potential of uploading a human brain within five years. He emphasizes the significance of Transformer models in revolutionizing our understanding of language and suggests that they have the capability to comprehend the structural language of the brain. McIntyre proposes that scanning the nanoanatomy of the brain, including every synapse, is feasible and can be accomplished within a year at a cost of less than $100 million. He highlights the importance of conducting experiments on lower-order brains before attempting on higher-order brains and emphasizes the need for a fully differentiable pipeline to facilitate learning and refine the methodology. McIntyre also addresses the challenge of accounting for fine-grained dynamic processes and plasticity in brain processes. He suggests the development of infinite context window types of Transformers to capture these complexities. Additionally, he discusses the potential distinction between memories pre-upload and post-upload in whole brain emulation (WBE).
