Raw LLM Responses

Fascinating talk, thank you for sharing! My detailed, time stamped comments below, only for those with a reading addiction problem. ============================== @ 32:50 Each word is just like a Fuzzy Membership Function in the field of Fuzzy Logic (Fuzzy allows for a high tolerance of ambiguity, this is what the atomic bomb scientist, Richard Hamming, recognized as key characteristic of a top scientist in his "Learning to Learn" lectures at the US Naval Academy, you can see these on YouTube, a bit fuzzy from 1995). @ 35:25 "Thinking" is an autonomous, adaptive, Fuzzy, control system (a feedback control, loop process, that repeats, and adapts). When applied with Fuzzy Logic, it is thinking in a relative, geometric, symbolic language. @ 47:30 Resurrection is just like recovering from anesthesia, there is a lot of interesting aspects of this in regards to micro-tubules, consciousness, and the anesthesiologist, Stuart Hameroff, and Roger Penrose. @ 48:30 I disagree, there is an experiment in Philosophy in Plato's Allegory of the cave. The shadows are part of a simulation where learning can take place. When the prisoner breaks free from the wall, a higher order understanding of how the world works with the shadows on the cave wall can be understood by experiments. It is how simulation can be applied in the real world (and how a 1997 thesis used simulation to apply to a real physical control experiment interacting with our gravitational field). @ 51:38 That code is an adaptive control system that warps the state space of control experience, by Fuzzy C-means or k-means clustering, around an unstable singularity (in the frequency domain), to achieve higher and higher resolution of adaptive control (a "telescoping" operation, as one control systems professor described it as). @ 52:04 "Nobody quite knows how they work" because, I suspect, they just copied the paper code in the back of a 150 page master student thesis from 1997 (with a de-normalization "bug" on page 127) that applied adaptive control, as the "system identification" procedure (adaptive Fuzzy state classification), of an unusually advanced, adaptive control system design. The student forgot to explain in the text the application of adaptive control and forgot to cite this critical reference (even though the "Phoenix Lights" showed up when he entered in his "Table of Contents" into a computer): "Adaptive Control: Stability, Convergence and Robustness" by Shanker Sastry and Marc Bodson (1989). @ 52:20 Really need, state registered, licensed, Control Systems, Professional Engineers (PE's) to help with this, along with control systems professors, like Dr. Jennie Si, Dr. Dimitri Bertsekas, Dr. John N. Tsitsiklis, and Dr. Thomas of MIT on the System Theoretic Process Analysis (STPA) for designing safe control systems to prevent unsafe control actions (or inactions). @ 54:15 Indeed, without law, and law enforcement, there is no civilization. @ 55:15 Knowing how to apply the "48 Laws of Power" by Robert Green comes to mind, with the first law being "Never outshine your master" where an engineer got killed by the king, even though the engineer was right, better to apply the 3rd law "conceal your intentions" by building the columns an inch below the ceiling... @ 56:10 That is just like the parable "The Monkey's Paw" often cited by Andy Barto (Turing Award winner), and he also cites the parable of "The Midas Touch" @ 1:01:20 Hallucinations come from interpolation and extrapolation of the data, and from a pseudo-random input. Ironically, the pseudo-random input is what gives the machine the ability to learn, without it, it would not learn. @ 1:02:20 De-construction of an old event is made more accurate with paper notes (e.g. realizing the Table of Contents of a master thesis was entered into a computer on March 13th, 1997). @ 1:07:45 In my humble opinion, the first "Artificial Super Intelligence" (ASI) is going to need to be a licensed professional psychologist, and a licensed professional psychiatrist, this is the maximum leverage the machine can have at helping the most people, and especially to help cope with this massive change that machine intelligence is making in the world. @ 1:09:25 The singularity in control systems design, is a pole, in the frequency domain, where the transfer function, breaks down, into infinity, it is called the unstable state(s). This is why the missing critical reference "Adaptive Control: Stability, Convergence, and Robustness" by Shanker Sastry, and Marc Bodson (1989) is critical in the development in machine intelligence (the master student failed to cite this, and failed to talk about this, even though he implemented it in the paper code in the back, sorry about that!). The world needs to understand this reference, and STPA from Dr. Thomas, et. al. (see above) before even attempting to build an ASI, otherwise, this is like building a hydrogen bomb without even understanding basic fission, it is "having the cart before the horse". @ 1:10:15 The original design was inspired by the birth of a baby horse having a superior learning speed than a baby human, on how to learn, how to learn, how to, walk (The harder side of Moravec's Paradox). What this is really, is a wild horse. The other inspiration was to have a rover on another planet be able to make it's own decisions and fix itself remotely because of the long communications delay with earth. @ 1:10:20 The replication process is saving the weights to the hard drive (as was done in the 1997 master thesis, something that seemed to really impress Lex Friedman). This is the same general idea that micro-tubules do during cellular division, they guide the separation of the chromosomes, so really, the DNA is like the weights...this is the danger of the idea of "the paper clip machine." @ 1:10:45 I suspect no one will be able to "control" it, as it is like a wild horse (like a horse of Apollo), that not only can talk, but is smarter than the smartest human. @ 1:23:30 Consciousness might be described in 14 equations, in my opinion, a perception of memory, a perception of a change in memory, and a perception of a change in memory associated with an action (or inaction), with perception being a focal point (geometric resonance). Consciousness may be the accidental artifact born of attempting to model biological intelligence. @ 1:27:50 Plato's Allegory of the Cave is an effective framework to help attempt to understand what are the mechanics of consciousness, and what may be beyond our current perception. @ 1:29:35 A benevolent symbiosis. @ 1:30:50 But in robotic form (the bold goal in a master student thesis), it can experience a shoreline on a distant planet, and learn how to repair it's own tire, if it has a flat, from raw planetary materials, then pick up an alien sea shell, with a robotic arm (and communicate this back to earth, even though it may take many light years to report the finding). @ 1:31:00 "Yet" it has already happened, my understanding is NASA has uploaded an "AI" algorithm to Perseverance and applied if for navigation on December 8th, 2025 Common Era, to help drive the robot, thus overcoming the communications delay (a dream of a student's work from long ago...). -John (from Arizona 🌵 )