As impressive as Musk's solitary NeuraLink demo is, beneath the hype lies a misperception with a disturbing parallel to large language models. Both operate on statistical inference rather than scientific models of the brain, and garner attention for cherrypicked successes 1/3
