George Ure has written multiple books exploring artificial intelligence, including Mind Amplifiers, Co-Intelligence, Test-Fitting Apes, Theomachines, The Judgment Engine, and AI: Thinking on the Gallows.
So when he ventures into new AI territory, there’s a substantial body of earlier thinking behind it.
Today’s Peoplenomics research goes considerably further than proposing a thinner, less computationally expensive alternative to today’s enormous AI models.
It asks whether we can build something closer to the functional architecture of a human mind: a continuously active system that learns from experience, develops preferences, remembers, forgets, makes consequential choices, and even processes information during sleep-like periods.
The proposed Volitional Observer Host Architecture (VOHA) explores whether recognizable individuality can emerge from the interaction of memory, noise, learning, and choice. Its longer-range Carbon-to-Silicon (C2S) project asks whether those processes can be sustained inside a relatively modest computer.
And this isn’t just armchair speculation. Today’s release includes an eight-page research paper documenting eight experiments, a programming framework, and a 22-page Python source codebook so technically inclined readers can inspect the methods and reproduce the simulations.
No claim that machine consciousness has been created. No declaration that the mystery of the human Observer has been solved.
Just an unusually ambitious question, some preliminary experimental findings, and the code to let others investigate.
What if the next major advance in AI comes not from building a bigger brain, but from learning how to grow a smaller one?
Details of Ure’s research (intended for the public domain) are available on his AI research site, VOHA: The Emergence and Persistence of Behavioral Identity in Stochastic Learning Agents – Hidden Guild.
And yes, today’s Peoplenomics also includes our 37-page ChartPack and a life-useful briefing of compressed world news.
More for Subscribers ||| Master Index 2018 to Present ||| Master Index 2001 thru 2017 ||| Missing out? SUBSCRIBE NOW!!! |||