I recommend all of his talks for inspiration and fun. He's a hybrid of an ingenious physicist and a very talented actor and comedian - reminds me of Robert De Niro and Al Pacino; it's something about their common New York accent.
Richard Feynman, the Nobel, 'It's a pain in the neck.'
Richard Feynman - The Character of Physical Law - Part 7 Seeking New Laws (full version)
Feynman on Scientific Method. (Related closely to General Intelligence principles in AGI)
Richard Feynman - The Distinction of Past and Future.
BTW, a POV to irreversability of mine -- it's because of the bad communication between particles. The smallest particles/details, whatever their nature is, are correlated too much and are harder to be controlled by an external structure than to do what's already "encoded". It's hard to cause an action without causing by-effects which are uncontrollable or undesirable, but also unavoidable - that's related to the "quantum uncertainty".
The smallest particles are what drives causation, and they are supposed to have the smallest equivalents of computing power and memory compared to bigger particles; that's why their "behavior" is the simplest, have the shortest scope and seems "chaotic".
Feynman 'Fun to Imagine' 1: Jiggling Atoms
... All of the series ...
Feynman 'Fun to Imagine' 4: Magnets (and 'Why?' questions...)
Richard Feynman - The Character of Physical Law - Part 6 Probability and Uncertainty (full version)
Thanks to Alexander for sharing a link to some of Feynman's talks.
7 коментара:
The whole point of Feynman's explanation was that the fundamental forces were (mostly) reversible. So it's nothing to do with the smallest particles/whatever, on the opposite - in the second part of "The Distinction of Past and Future" he shows irreversibility to be a macroscopic phenomenon, namely "the general accidents of life".
Do you think of the Universe as a cellular automaton? My first reading of your "bad communication" was "they loose information", then I figured by "bad" you maybe meant "limited" as in only within the cellular neighborhood per turn. What's your idea exactly?
Otherwise, I totally share your admiration for Richard Feynman. His way of trying to visually understand things is quite inspirational. BTW, here's a video that caught my attention a while ago: Richard Feynman - Ways of Thinking (Part One of Two). Part one is enough. There he describes his "mental counting device", how it differs from the one of a colleague of his, and what they can/cannot do while they are counting. :D Just if you're curious.
Hi Georgi, it's nice to see you. :)
My most exact idea is in my teenage theory original works:
http://eim.hit.bg/razum
With some translations and additional notes and slides:
http://research.twenkid.com/agi_english/
I believe Universe is a computer, however I have/had not a particular preference for the type of a computer (unlike Zuse or Wolfram), I think it can be implemented anyhow and have an effect as if it was a CA or a random-access-memory computer. To me it rather seems as a mixture of both, the accesses by some "central" processor/s are done in between "clocks" (frozen state for them), so it's undetectable by the internal parts of the Universe, which detect things after clock.
E.g. some forces (data) are "instantaneous", others need time - sort of clocks, the data are retransmitted between the "particles" - the "processing units".
My explanation of time-slow down in high speeds is that the processing units are missing synchronization signals from each-other.
"Bad" has many meanings :), such as there's no explicit memory for history, and there are too much entangled cross-correlations - each particle impacts each other.
I watched that video as well - yeah, it's/he's funny. :) It's about consciousness, it's diverse, but pretty slow and clumsy.
> Hi Georgi, it's nice to see you. :)
Oh, where were my manners, hello there. :)
> My most exact idea is in my teenage theory original works:
>
> http://eim.hit.bg/razum
>
> With some translations and additional notes and slides:
>
> http://research.twenkid.com/agi_english/
Unfortunately, I was able to only skim through your material. Thanks for the translations though — Smetacho-Bulgarian is too hard for me sometimes. :D
> I believe Universe is a computer, however I have/had not a particular preference for the type of a computer (unlike Zuse or Wolfram), I think it can be implemented anyhow and have an effect as if it was a CA or a random-access-memory computer. To me it rather seems as a mixture of both, the accesses by some "central" processor/s are done in between "clocks" (frozen state for them), so it's undetectable by the internal parts of the Universe, which detect things after clock.
Not having a preference is fair enough if you only care about computability (imagine the Universe turns out to be a formal grammar, hoho). Of course, if everything we can ever detect from the inside are your supposed states "after clock" then the "true" implementation of the Universe will forever remain just a speculation.
Having it implemented directly as a CA allows in my view for simpler metaphysics — we can handwave it as some inevitable simple process which behaves like a bunch of cells interacting, no need to wonder how your Cosmic Megacomputer got built in the first place.
Apropos, I'm just commenting on the topic here, I don't believe the Universe should be necessarily digital.
> E.g. some forces (data) are "instantaneous", others need time - sort of clocks, the data are retransmitted between the "particles" - the "processing units".
>
> My explanation of time-slow down in high speeds is that the processing units are missing synchronization signals from each-other.
This signal missing should be "by design", you know, time dilation is not a defect, it's a feature. :) You have to figure in what kind of architecture the consistent miss of clock signals would result in the 1/sqrt(1−v²/c²)-fold slowdown of time relative to the "stationary" observers. It's always pretty to add some formula to your theory. ^^
Consider also whether signal missing is the best way to achieve that effect. It's curious that your particular explanation requires running in continuous time (by "time" I mean the "time outside time", heh), not just a succession of states. Turns out your theory is somewhat analog. :D So, if we don't experience "outside time" then the Universe doesn't need to hurry and can take its "time" to compute everything digitally.
> "Bad" has many meanings :), such as there's no explicit memory for history, and there are too much entangled cross-correlations - each particle impacts each other.
Aha, so I understood the exact opposite of what you meant. I thought it was bad because there was too little communication, not too much of it. :)
> I watched that video as well - yeah, it's/he's funny. :) It's about consciousness, it's diverse, but pretty slow and clumsy.
Well, at least it's for free. :D
Thanks for your comments, Georgi,
Georgi>Thanks for the translations though — Smetacho-Bulgarian is too hard for me sometimes. :D
Todor:
Sorry, but I think you overplay here :), a lot of words from the Yunashki Dialect were used only in Part 2 of the foundation works (it's most wordy), in others they are clearly explained, if I'm not mistaken.
Georgi>Having it implemented directly as a CA allows in my view for simpler metaphysics — we can handwave it as some inevitable simple process which behaves like a bunch of cells interacting, no need to wonder how your Cosmic Megacomputer got built in the first place.
Yes, but outside our internal time it doesn't matter how "long" it takes for a computation and it's unknown which is simpler there. In some conditions it's simpler to have a general purpose "CPU" with a uniform memory cells, which mimick cells in CA, rather than a zillion of cells, which are more complex than a memory cell.
Georgi> Apropos, I'm just commenting on the topic here, I don't believe the Universe should be necessarily digital.
Todor:
OK, but talking about simplicity, I'll just remind you that "analog" could be infinitely complex and impossilble to define.
On the other hand digital design could be finite, and they can even imitate infinity with finite resources, so that the internal observer is confused. You can just be limited by the design not to be capable to reach to the "end", so you can't check what's there.
Regarding time sold down:
Georgi> This signal missing should be "by design", you know, time dilation is not a defect, it's a feature. :)
Todor:
Yes - the sync signal is transported one step in a coordinate space at each clock (could be Plank's constant or whatever, and the clocks number is huge).
When the transmitter of the signal is changing its coordinates, the sync signals are travelling more steps or they are missing the receiver - they arrive on past coordinates of the receiver and it misses some of the signals, it's not there at the moment of evaluation.
When the objects are stationary to each other they receive the maximum possible amount of sync signals and their time is synchronized.
When they move, some of the sync clocks are missed and their relative time is desynced, proportional to the missed number of signals.
Continues...
Georgi>It's curious that your particular explanation requires running in continuous time (by "time" I mean the "time outside time", heh), not just a succession of states. Turns out your theory is somewhat analog. :D
Todor:
I don't think so, I meant number of clocks missed, that's completely digital. The states could be just: "sending signals". Next state is "sending another signal". I guess you lack background in digital systems design, that's why you say so.
I'll explain:
In every digital device there are one or usually many clocks - this is a very fundamental concept and in my opinion it has a deep metaphysical foundation (I had speculated about that in my theory) and relation to "time".
Clocks in CPUs etc., periodic signals which travel through the system, are "time" for the virtual universe of a computer system, it's namely for synchronization of different parts of the system, which are supposed to work in parallel in the same time, while the physical time - the lower level clock - is not the same. The clock in computers sets "time" and "contemporary".
Electric signals travel different times to different parts of the schematics, while sometimes these different parts have to work together in order to produce result for the following module.
That's why you need to set "windows" in the lower level time and use lower resolution time and mechanisms to delay the signals deliberately (store in registers, collect potential in capacitors etc.) in order to cross events which would otherwise hapeen in differnet phsysical times.
Thank you for the replies.
Todor> Sorry, but I think you overplay here :),
Guilty. :) My reading almost doesn't count. It's highly possible that I've even missed your main point. On that account I was thinking of quitting the discussion and leaving the last word to you but one thing wouldn't let me rest: :)
Todor> When the transmitter of the signal is changing its coordinates, the sync signals are travelling more steps or they are missing the receiver - they arrive on past coordinates of the receiver and it misses some of the signals, it's not there at the moment of evaluation.
What do you imagine happens when the receiver is travelling towards the transmitter. From my (humble) understanding of special relativity there shouldn't be any difference in dilation with different directions of travel, if speeds are the same.
I assume that external space somehow corresponds to our internal one. This shouldn't be necessary but everywhere you talk about movement and coordinates in external space.
Now that I'm at it, I can't help but comment some more. :D
Todor> Yes, but outside our internal time it doesn't matter how "long" it takes for a computation and it's unknown which is simpler there. In some conditions it's simpler to have a general purpose "CPU" with a uniform memory cells, which mimick cells in CA, rather than a zillion of cells, which are more complex than a memory cell.
OK, now I can see what you have in mind. I don't know how much more complex automaton cells have to be than memory cells but they surely are uniform, that's their point. In fact, I was thinking of them as not having any internal structure at all, which maybe strengthens your case by probably requiring some outside mechanism, like a "CPU" or a read/write head, to implement the update rule, but even then we still have the rules which govern the insides of this mechanism to explain.
Cells locally interacting in parallel also don't have trouble with being infinitely many, but I guess you're not a fan of this. :)
Todor> OK, but talking about simplicity, I'll just remind you that "analog" could be infinitely complex and impossilble to define.
It can be said that the Universe just happens as it is, definitions are our problem entirely.
Todor> I don't think so, I meant number of clocks missed, that's completely digital. The states could be just: "sending signals". Next state is "sending another signal". I guess you lack background in digital systems design, that's why you say so.
Correct, I don't have that background. Must be a misuse of terminology on my side, I took "analog" to be a synonym for "continuous" in all cases. I meant that the time intervals for signal travels were possibly non-integer, like they supposedly were in our world, from a classical viewpoint. Anyway, I completely disregarded the possibility of them consisting out of numerous minuscule steps.
I'll also backtrack a little on my previous statement that the true implementation will remain "just" a speculation. Ugh, of course, we can rate different speculations by their simplicity, explanatory power, whatever. In some sense this is all we do.
GeorgiWhat do you imagine happens when the receiver is travelling towards the transmitter. From my (humble) understanding of special relativity there shouldn't be any difference in dilation with different directions of travel, if speeds are the same.
Todor:
Yeah, I thought of this case also. :)
It could be in the difference of the speed of sync signal compared to the speed of the transmitter and receiver.
I guess it may be dependent on limited mutually exclusive computing resources which are needed for both operations - recording coordinate translations/transformations in "real time" (maximum speed/smallest time step in Universe) and synchronizing between different items/particles at each step.
It can happen that way:
- Translation of the transmitter/receiver is of higher priority than synchronization, and is done first.
- If a particle/item is moving, motion (translation the content in memory) takes computational "slots" and less time sync signals left, so time gets different.
In that interpretation the direction doesn't matter, translation in any direction to a receiver/transmitter requires the same computational resources.
Regarding the true implementation - I do care mostly about models which would "work"/predict, and it would be nice if such models assist in understanding general intelligence.
Post a Comment