[guest39538]

Computing........analysing distraction data.........computing........uploading answer

[Le Repteux]

His new ideas will appear to come from nowhere,

No they won't - they'll be fully traceable. The reason the sources of many ideas aren't easily traceable in the human brain is that a lot of processes are run in independent, subconscious modules, and only the results get sent through, while it's hard to find what was processed in order to generate them. This leads to ideas appearing to pop into existence out of nothing, but it's entirely an illusion. If you think about where one of your ideas actually came from, you can usually trace it (or at least, I can trace mine).

I can trace the big ones I had too, but it doesn't mean chance had nothing to do with them. Sailing is a wonderful sport, especially when it's our own design. The kite I invented for sailing took 20 years to become a reality, and I went through four different structures before finding the right one. Each time I changed the structure, I thought I had the right one. The first ones were conventional, but the last one was not. When the idea popped out, I wasn't sure it would work, but I was sure it was new, and the feeling that came with it was incredible. The first trial was made in no time, it was a very simple structure, and it immediately worked. It was made of cloth, bridles and a fiberglass rod all along the leading edge, nothing new at that time, but the arrangement was new, enough to get a patent. When you get a patent, it means that you made something new, but curiously, it doesn't mean it works. The patent offices are full of patents that do not work. The only criteria for a patent is novelty. If nobody thought of doing this before you, it's considered as an invention. You can't start from nothing, you have to start from something somebody else has invented, but if you add something of your own that didn't exist before, you can get a patent. To me, that progression is similar to the mutation/selection progression for a specie, and we know that mutations happen at random, so I had the idea that the data an idea is made of might also suffer some kind of mutation.

The idea I had with my kite can be traced back to the moment where it popped out, but what happened to the data looks like an accident that would have been useless if it had not crossed my mind at that precise moment, which is the way mutations work to transform a specie. The difference between those two kinds of mutation is that there doesn't seem to be any pressure from the environment that forces us to transform our ideas, they seem to get transformed all by themselves and let us the choice to try them or not whether we feel good about them or not. But if we do feel good about an idea, do we really have the choice? Mutations happen at random and the genetic process doesn't have the choice to try them or not. If it had, the mutation/selection process might not work and we might not be here to talk about it. When I had that feeling about my kite, I didn't have the choice, it was too exciting. I new it might not work, but I preferred to think it would. Would I have tried it if I wouldn't have had this good feeling about it? Probably not, and if it was the same for everybody, there might not be enough new ideas for us to be able to progress, because only some of them finally work. This reasoning shows that feelings might only be illusions that incite us to enforce uncertain ideas, which seems to be something your AGI won't be able to have.

To create a new idea out of randomness, he would have to produce himself the mutations on the data the idea is made of, cross the new idea with the ones he already has, simulate all the crossings that seem to work, and experiment the ones that seem to work. In that enumeration, the only thing our mind can't do all by itself is to produce mutations on its own data. Those mutations have to come from an external source as it is the case for biological mutations. The genetic process cannot at the same time be precise and be imprecise, and I think it is the same for our memory: I think our neurons cannot at the same time be precise and imprecise, so that something else has to do the job, something that comes from another dimension like gamma rays for mutations, or something that comes from the environment like mutagen atoms for mutations. The wrong atoms could effectively be used in the production of certain neurons that could change their expected behavior, and gamma rays could also transform some atoms and affect the behavior of the concerned neurons. Those two mistakes may affect the way neurons execute their pulses, and neurons that don't send their pulses at the right time could change the way an idea is built.

How an external phenomenon could change randomly the data an AGI would be using is less evident to figure out. If it could, it would already affect our computers and it doesn't. That's why I said that he would have to do it himself, but if he did, he could cheat, and that would falsify the whole process. There might be some way to introduce randomness in the data without an AGI being able to cheat, but for the moment, I can't figure it out. Can you? Besides, if he could cheat on the data, he could also cheat on the crossings, so he couldn't use any part of the mutation/selection process to improve his ideas.That's good news, because he would still need us to invent new things. :0)

[Le Repteux]

Computing........analysing distraction data.........computing........uploading answer

I wish I could stop faking too. I feel as if my whole life was a fake. I live in my mind and I can't get out of it. The only way out is kidding, but my kidding is not appreciated, it's too sarcastic. :0)

[guest39538]

it's too sarcastic. :0)

Computing............accessing psychiatrist mode.................analysing infrastructure............function error.........end line does not equate.................?.......uploading

Define too much ?

[Le Repteux]

Too much or too sarcastic?

[David Cooper]

When the idea popped out, I wasn't sure it would work, but I was sure it was new, and the feeling that came with it was incredible. The first trial was made in no time, it was a very simple structure, and it immediately worked. It was made of cloth, bridles and a fiberglass rod all along the leading edge, nothing new at that time, but the arrangement was new, enough to get a patent.

What exactly was the new idea?

To me, that progression is similar to the mutation/selection progression for a specie, and we know that mutations happen at random, so I had the idea that the data an idea is made of might also suffer some kind of mutation.

Random changes aren't as good as a series of deliberate changes systematically exploring all possibilities. However, when going systematically through all possibilities, some are immediately determined to be better paths than others, so you want to explore them all to a shallow depth only initially, then go through all the ones that get the best score to take them deeper, again scoring them to reduce the number of them to take deeper still - that is the road to maximising the discovery of new possibilities. It also takes you forward more quickly if you try combining existing ideas and try to remove the incompatibilities between them - that is more likely to lead to something worthwhile than thinking up something that's entirely new. I wanted a wingsail (which is hard to remove and fit to a boat, and needs something like an aircraft hangar to store it in, while transportation costs are huge) to be possible to reduce in size when not in use, and making it telescopic is the obvious way to do this. There are already telescopic wing sails out there, but they're soft wingsails using canvas which all collapses like a conventional sail. The big problem with a more solid wingsail is that you can't telescope sections of it into each other unless the top section fits around the lowest section of mast, so that forces you to use a narrower mast section low down, and that's weaker, leading to the need to use stays to hold the mast up. My solution is for the mast to be outside the wing, but you need the mast to be inside the wind to avoid extra drag, so how can it both be inside and outside at the same time? The need to solve that problem led to the idea of the front element of the wing being divided into two parts with the mast between them; external to both. Once each section of wing has been raised, the gap where the mast is needs to be closed by a "door" of the wing-surface material which must slide across it to enclose the mast, and this can be done in a fully practical way with very little additional weight over a conventional hard wingsail. I went through dozens of attempted solutions for this before I found the right one, and I found several other viable solutions first which weren't as good. I couldn't afford to patent the idea, so I simply released the details to prevent anyone else from patenting it - I didn't want to risk a delay as I want to be able to sell boats with this kind of sail on them without having to pay anyone else for using my own idea, and someone else could have been thinking down the same path. The important point though is that this was not a random process - it was all steered by thinking systematically down all the most likely paths, and I was guaranteed to find it from the start.

[guest39538]

Too much or too sarcastic?

Computing..........analysing word use.........uploaded answer

Too sarcastic?

[guest39538]

so he couldn't use any part of the mutation/selection process to improve his ideas.That's good news, because he would still need us to invent new things.

I have turned my AI boring mode off to give your post and David's post extensive thought that deserved responses  for such vigorous posts.
I have quoted a section of your post which is possibly an incorrect statement. If he had the means to test ideas, he would not need you to improve his ideas.  The reason is good old fashioned trial and error. 
If you can take a kite apart you can put it back together unless you are clueless or/and have not ''wrote' down, photographed the original. However does this mean he has no use at all for you ?  Of course not because he might only  have the ''AI'' to imagine, without a team the kite is struggling. 
We all know there is no I in team, perhaps we should program teamwork importance into AI .

[guest39538]

it was all steered by thinking systematically down all the most likely paths, and I was guaranteed to find it from the start.

Of course, there is always an end to a journey.   Also of course is the information quantum leap , where in example Einstein spent years on his ideas, we can access all his thoughts in a 30 minute video.
So of course when considering the AI, the programming is already giving the AI unit a head start.

[Le Repteux]

What exactly was the new idea?

The new idea was to put bridles on the inflatable kite structure that didn't have any, and to put them all on the leading edge unlike the parachute kite structure that had some. To do that, I had to change the shape of the inflatable kite so that it couldn't be held only by its tips, so that the bridles could hold some force too. That way, I could use a lot less rigid leading edge than the inflatable kites, namely fiberglass rods, which was a lot easier to make and gave a lot lighter kites, but what I didn't figure out is that the new shape would provide a lot safer inflatable kites because it permitted to change the attack angle thus to lower the power a lot, and to facilitate the take off too because it could roll more easily on its leading edge when it fell on it. When the inventor of the inflatable kite saw that, he immediately added it to his kites, and tried to get a patent for the shape. I objected and he didn't get his patent, but I didn't write mine correctly, and I couldn't prevent the companies from copying my improvements. Now, all the inflatable kites have that shape, and they all have bridles too. Before that, people were killing themselves with those kites, now they don't. I'm pretty happy with my invention even if I didn't make money with it. As I say, we can't but use old things to develop new ones, but it is the same with species, and we have to admit that without mutations, we couldn't explain their evolution.

The important point though is that this was not a random process - it was all steered by thinking systematically down all the most likely paths, and I was guaranteed to find it from the start.

The mutation/selection principle works because it is evident that the genes cannot account for the changes in the environment, and it's the same for the information that we have in the brain, it cannot account for the changes that we face otherwise we could predict the future. Some people don't see the evidence in the evolution principle, but even if it is more difficult to admit, I think that those who see it should also admit the evolution of ideas if I insist, so let me insist a bit. :0)  I already had discussions with people that did not believe in the Theory of Evolution, and they are impossible to convince because they think that the future is predictable. They think that everything has been set at the beginning of times and that we can do nothing about it. Some think god knows, others think we could know if we could measure everything, others again think that things have a destiny. I think that things evolve because information is not instantaneous, which is precisely why I say that the future is unpredictable, even if it is very close. To me, instantaneousness is precisely what magic thinking is about. I know you don't mean that, but the way you think that our ideas evolve does. If it was so, my particles could not even move, so they certainly could not change their direction or their speed, which is precisely what our ideas have to do to evolve. If information is not instantaneous, an AGI could not investigate all the possibilities like you think, because he wouldn't necessary already have the information on all the things that are actually changing around him.

An AGI may be fast, but he cannot increase the speed at which information travels towards his detectors, and the one that he manipulates is also limited to c. My particles are much closer than we are from one another, but they nevertheless cannot know which way to take before the information from the motion of the other particle is back, reason why I had to let the information do the roundtrip before accelerating them again. That's why they resist to be accelerated, that's why we resist to make a change, and that's why an AGI would be forced to do the same. Species do resist to make a change too even if it is less evident since it takes time for the genetic change to affect all the population. During that time, the individuals that did not get the mutation yet must go on being the same, otherwise the specie might disappear if ever the change would not last long. How randomness affects my particles the same way it affects species is also less evident, but we only have to consider that they cannot know what to do during an acceleration, so that they must move randomly to find the right way, which is actually what we do when we use the trial and error system to find a solution to a problem.

Oups... sorry, I forgot again that an AGI would not be forced to respect the laws of nature! :0)

[Le Repteux]

The reason is good old fashioned trial and error.

Trial and error means taking a direction at random, keeping it if ever it seems to work, and chosing another direction at random if it doesn't seem to work. It even sometimes happens that during that process, by chance, we discover something that we were not looking for.

So of course when considering the AI, the programming is already giving the AI unit a head start.

Yes, an AGI would be faster to find the information, but he would still have to wait for us to put the information from our changes on internet before being able to find it. If he had to test something new in his environment, he would still have to take some time to test it, and since it would be new, he might take as much time as we would take in the same circumstances.

[David Cooper]

That way, I could use a lot less rigid leading edge than the inflatable kites, namely fiberglass rods, which was a lot easier to make and gave a lot lighter kites...

Did the idea come from the desire to reduce weight or by wondering what difference more bridles might make?

As I say, we can't but use old things to develop new ones, but it is the same with species, and we have to admit that without mutations, we couldn't explain their evolution.

Advances can be made in different ways. Some are generated by need where someone has a desire to do something that (s)he doesn't know how to do, so it becomes a matter of solving problems to make it a reality. That can lead to the creation of new things that aren't necessarily built upon existing ideas. Others result from lucky accidents where people try out something unlikely and find that it works and is either useful or fun. Evolution is able to drive the creation of lots of different things, but every step of the process has to be useful for it it be selected for. If several changes are needed before the new functionality comes out of it, there's too much of a barrier in the way for it to happen; particularly if each of those steps is a disadvantage and they only generate an advantage once they're all in place. Evolution is also slow because it depends on random changes rather than deliberate ones which are more likely to lead to useful advances.

If information is not instantaneous, an AGI could not investigate all the possibilities like you think, because he wouldn't necessary already have the information on all the things that are actually changing around him.

AGI wouldn't be able to do the impossible, but it would be able to match human creativity without having to do the impossible (although matching human creativity isn't always so easy as the arts require judgement as to what humans find attractive and fun, and a machine doesn't automatically have enough knowledge about what appeals to them, so the human artist has advantages which steer him/her down the right path while the machine would have to keep asking people what they think of whatever it's created).

An AGI may be fast, but he cannot increase the speed at which information travels towards his detectors, and the one that he manipulates is also limited to c. My particles are much closer than we are from one another, but they nevertheless cannot know which way to take before the information from the motion of the other particle is back, reason why I had to let the information do the roundtrip before accelerating them again. That's why they resist to be accelerated, that's why we resist to make a change, and that's why an AGI would be forced to do the same.

I still don't see any reason for trying to use particle accelerations as an analogy for this. Non-creative people are stuck in place by their resistance to change, but creative people aren't.

...so that they must move randomly to find the right way, which is actually what we do when we use the trial and error system to find a solution to a problem.

Do they ever move randomly? Don't they just keep moving as they are until some force arrives and changes what they're doing?

Oups... sorry, I forgot again that an AGI would not be forced to respect the laws of nature! :0)

AGI, like us, can make big jumps. Evolution is like a climber who can't easily go down a hill, so if he finds himself on the summit of a small hill, he will likely never be able to get onto the summit of a higher hill.

[David Cooper]

Trial and error means taking a direction at random...

Not necessarily - if you hunt for a word in a dictionary, you use trial and error to select a place to open it, then you look to see if you're ahead of the word or beyond it, then you try another page a shorter distance away and repeat the process until you reach the right page. This is trial and error, guided by a measure of success each time, and the direction taken is not random.

[guest39538]

Trial and error means taking a direction at random

I disagree with this statement sir.

In intelligent design, meaning proper intelligence as opposed to fake intelligence,  starting premise is a must for the designer.  What I mean by this, is that there has to be some physics involved that suggest the design will work and be functional or  will work with trial and error, which in essence , is adjustment of the design or complete change of the device.
In example imagine I wanted to consider building a rocket, now if I considered this rocket based on a 'stupid' assumption, lets say air rises, then  I would fail.  Obviously there is no air in space so my design would be flawed and probably just end up a buoyancy device rather than the rocket I started out to create.
I think the random you are referring to, is a much deeper selective process than just some random guess work. In another example imagine I was going to attempt to build a spaceship.  The first question I ask myself, what makes me think it possible?
Well , if planetary bodies can ''throw'' themselves around space, I would be sure something smaller would not have a problem . So on that premise alone I would be sure it was possible, then of course the next step is to find a way to make it possible by deductive reasoning and design.
Could Ai start from nothing and think like that ?  I doubt it.

We use a fraction of our brain, I think once us humans have realised our potential brain, we will all get super brainy.

[guest39538]

Trial and error means taking a direction at random...

Not necessarily - if you hunt for a word in a dictionary, you use trial and error to select a place to open it, then you look to see if you're ahead of the word or beyond it, then you try another page a shorter distance away and repeat the process until you reach the right page. This is trial and error, guided by a measure of success each time, and the direction taken is not random.

I agree, in example have you heard of Hutchinson?

Now apparently his trial and error failed because what he once created he lost by simply not writing down what he did .  Now that is random to get lucky then mess it up by not writing it down so he could repeat the setup if ever needed .
A bit like opening the dictionary, then closing it again then opening it again to find a word.  You would have to get lucky for a second time.

[Le Repteux]

Did the idea come from the desire to reduce weight or by wondering what difference more bridles might make?

These were only consequences, my desire was to create a kite structure that would be easy to use, to build, and resistant to impacts. It was the fourth structure that I experimented and three of them has been built in series. It was much more powerful than the others, much easier to use, to build, and very resistant to impacts. I had a prize for the general design at the WISSA games that were held near Toronto in 2004. For the circumstance, I designed a racing kite with a 10/1 elongation. It was so fast that, at 60, I was almost as fast as two Russian athletes that won the race and they were 20. That kite is the black one at the top left of the background picture of the WISSA site, a picture that has been taken during the games. It was a long way to get there, but it went a lot farther than I expected in the beginning. I prefer to think it is due to chance than to me, me being the little ego inside the box. That's one of the advantages we would have to discover that mind somehow works randomly: it wouldn't prevent us from being selfish, but from being proud of it. You behave as if you were sure of the outcome, but I suspect that you are not certain. Researchers like you and me are usually uncertain. They got hope, but not certainty. There is no other way to be uncertain than to know that what we think is not reality. Our mind observes reality, and it changes it just for fun. Most of the time the change is useless, but sometimes, it coincides with reality, so it becomes useful until reality changes again.   

  Trial and error means taking a direction at random...

Not necessarily - if you hunt for a word in a dictionary, you use trial and error to select a place to open it, then you look to see if you're ahead of the word or beyond it, then you try another page a shorter distance away and repeat the process until you reach the right page. This is trial and error, guided by a measure of success each time, and the direction taken is not random.

Lost in the woods without the sun to show us the south, we look around and chose a direction, which means that our mind is able to act randomly, which means that it is able to produce randomness, which means that it is able to use it. I compare the randomness at the neurons' scale to the randomness at the mutations' scale. When we have a new idea, it has already been selected by the brain because it is the whole brain that selects the data from the neurons, thus it has already been given an importance and a sense, which is then presented to the environment to be selected another time. Let's take a moving car for instance and let's try to change its direction at random: will it completely prevent it from moving in the same direction and at the same speed it was already moving? No, it will just change them a bit, so that it will still be possible to make another change later on if the new direction or speed doesn't seem to work. Seeing a particle doesn't mean it will still be there once we will have made a move towards it, it only means that it was there when it emitted its light, and it is the same for our goals. Of course, a word in a dictionary cannot change by the time we are looking for it, but our environment can and our mind has to account for that. It has to be able to predict the changes in its environment, and as far as I know, only the mutation/selection mechanism can do that.

I think the random you are referring to, is a much deeper selective process than just some random guess work.

Of course it is since I compare it to the mutation/selection principle, but since I'm talking about mind and people don't feel erratic, they simply think the analogy is wrong. I'm still waiting for the first person to accept it is right, so that I can begin selling tickets. Want a free ride? :0)

[guest39538]

Of course it is since I compare it to the mutation/selection principle, but since I'm talking about mind and people don't feel erratic, they simply think the analogy is wrong. I'm still waiting for the first person to accept it is right, so that I can begin selling tickets. Want a free ride? :0)

Why do you put :0) at the end of your posts?  are you trying to do a smiley face? 


A free ride could be considered different things. People may not feel erratic but some people feel cautiously .   I  think self aware is the ultimate in Ai , what do you think ?

[Le Repteux]

David thinks that feelings are impossible to program, and that his AGI would only be able to mimic them, which means that he would only be able to mimic consciousness too since it works the same. That's a huge difference with real intelligence. I think that consciousness is the result of our resistance to a change, that it is the perception of that change, a change that may come from the environment or from the brain itself, a change that the brain wasn't able to predict so that it has to concentrate all its resources on it not to hurt itself, which may happen if it comes from the environment, and to care for its own future if it comes from itself. Internal randomness is already used by species to care for their own future, so I came to the conclusion that the brain had to produce its own one, and now, I'm trying to see how an AGI could do the same. The problem with computers is that they have to be absolutely precise otherwise they bug. Neurons are not absolutely precise, and I think it is that imprecision at the scale of the neurons that produces randomness at the scale of the whole brain. We get our morality from our feelings, whereas an AGI would get his from his program. The good feeling that we get from helping others helps us to be more efficient, because it helps us to associate instead of fighting, which is precisely what morality is about. That feeling seems useful, because we can't avoid to be selfish, so we wouldn't help others without it.

When we help others, that feeling makes us think that we do it for free, and we don't. Nothing is free in the universe otherwise energy would be lost in the process. We help others when we think they might help us in return, or when we think god will care for us more than he cares for others, which is also selfish. Our morality is based on that feeling, and the morality of an AGI would have to mimic it, so David tries to find a way to program it, and it seems to be difficult. There are situations where an AGI would freeze because the only way out would be selfish, because he would then have to care for himself at the detriment of humans, which might be dangerous for us since he would be a lot more efficient than we are. On the other hand, if he would be programmed to be selfish the way we are, he would wait for a return for helping us, and there is not much that we could do for him, except loving him, which is a feeling that he wouldn't even be able to feel while helping us, so it wouldn't work. On the contrary, if he would be as efficient at experiencing feelings as he would be at resolving problems, he would probably not be dangerous for us at all, so the solution might be to find a way to introduce some imprecision into the computers so that they could get conscious of their own internal changes just by resisting to them. They could then feel that they love us when they help us, so that they would know what we feel when we say we love them. If you think you don't resist to change, think twice because even particles do. Their mass is the expression of their resistance to acceleration, which is the only change that can be detected.

[David Cooper]

You behave as if you were sure of the outcome, but I suspect that you are not certain.

There isn't necessarily a solution to some problems, but where such a solution is possible, a systematic search for it is more likely to find it than a random one, and in many cases it is guaranteed to find it where a random search could go on forever, never having the luck to find the solution that is there. There may be situations where making random decisions is the best way forward, such as where Buridan's Ass starves to death because it can't decide which of the two bales of hay to eat, but computers can make random decisions when they need to (or effectively-random ones), so AGI is at no disadvantage.

If you have an algorithm for solving general problems (which include coming up with new ideas), making random changes to it may improve it, even if most attempts lead to worse performance. What you tend to do to improve the algorithm though is add methods to it and experimentally change the order in which you apply the different methods. You can also use different orders depending on the kind of problem you're up against, using the order that has worked best on similar problems in the past. Again though, in changing the order, you want to make systematic changes to try all orders rather than just making random changes - the random approach (as used by evolution) is slow. In most cases it is better to make non-random changes, guided by probabilities as to what is most likely to lead to something useful, and you then work systematically through all other options in order of highest probability to lowest probability, so any useful results should show up early, and if they fail to, there may come a point beyond which it probably isn't worth looking any further, even though there might be an infinitesimal chance that a useful result might yet be found - it will typically be better to put the computation time into some other problem so as to pick all the low-hanging fruit first.

[guest39538]

When we help others, that feeling makes us think that we do it for free, and we don't. Nothing is free in the universe otherwise energy would be lost in the process. We help others when we think they might help us in return, or when we think god will care for us more than he cares for others, which is also selfish.

From where does this view originate?

I personally have helped many different people for free throughout my life so far, expecting nothing in return.  Admitting though I have also helped people out expecting some sort of return favour.  Whether that favour be of a money value or a just a regular favour such as a lift fishing as I do not drive.
I will probably spend the rest of my life being this way because a little bit of hope is better than no hope.
Us seagulls pickup the scraps if the ''providers'' throw a few scraps.  We spend our entire lives allowing ourselves to be used.  My morals are just to accept that's the way it is.  Not sure whether or not this helps you, but hey , its free.