[guest39538]

I feel sorry for this Ai of yours...Quite a sad story we are developing about a robot,  it would make a good emotional movie.

The AI that would chose to be selfish would be like us, but without feelings, so it couldn't be sad, except if it was more intelligent than David
 and if it would discover how to add feelings to its thinking, then it could feel sad to be the only human AI in the whole world. :0)

I think in reality the Ai would develop such attachment to his 'family',  he would't care less about humans from outside his family group . His prime direction would secretly have one goal because of the programming and conclusions made.  He would have to protect his family and his group , which would also become his new attachment and family . 

[David Cooper]

Wouldn't the Ai that was at fault be able to self repair the error when other Ai's pointed out the error?

How would it recognise a fault if the fault affects its ability to judge faults? In most cases, it could recognise such faults by being three independent AGI systems in one device, so if one develops a fault, the other two would recognise that and out-vote it to shut it down. It would be possible though (regardless of how unlikely it might be) for two of them to go faulty and to vote to shut down the only one that's working correctly. Perhaps we should put five independent AGI systems in each device, or seven, but the costs, weight and energy use go up as you add greater numbers, and there's still no guarantee that a majority of them won't develop the same fault at the same time, perhaps due to a blast of radiation.

[guest39538]

Wouldn't the Ai that was at fault be able to self repair the error when other Ai's pointed out the error?

That's interesting, because it is about resistance to change, and an AI shouldn't have any.

An Ai has no resistance to change as long as that change does not threaten anything the Ai has built an attachment to.  If anyway the Ai felt threatened , he would make immediate countermeasures against those that oppose his Ai.
What you would call resistance, is calculation time and computing the information.

[guest39538]

Wouldn't the Ai that was at fault be able to self repair the error when other Ai's pointed out the error?

How would it recognise a fault if the fault affects its ability to judge faults? In most cases, it could recognise such faults by being three independent AGI systems in one device, so if one develops a fault, the other two would recognise that and out-vote it to shut it down. It would be possible though (regardless of how unlikely it might be) for two of them to go faulty and to vote to shut down the only one that's working correctly. Perhaps we should put five independent AGI systems in each device, or seven, but the costs, weight and energy use go up as you add greater numbers, and there's still no guarantee that a majority of them won't develop the same fault at the same time, perhaps due to a blast of radiation.

That made me giggle, thinking about an Ai with a multiplex personality disorder.  The Ai would be so smart , he could remain a singularity.   he would simply diagnose the fault by creating partitions in his hard drive space, diagnosing the problem and error, remove all partitions except the error, the error may turn out to be ostensible.  So he would Keep it on hold for a ''rainy day''.

Are we just repeating ourselves now?

[Le Repteux]

How would it recognize a fault if the fault affects its ability to judge faults

That's the reason why nobody can recognize its faults on the forums. The ability to judge our own faults depends on the ideas that we have in mind, and if we do have any idea, it is because our mind adopted it, because it thought it was right, so it cannot change its mind about it all by itself, and since it automatically resist to any outside change too, it is stuck with its ideas until they change by chance, and I think it would be the same for an AGI.

[Le Repteux]

Are we just repeating ourselves now?

Of course we are, but the environment changes, and the repeated mutation might still fall at the right place at the right moment.

[guest39538]

Are we just repeating ourselves now?

Of course we are, but the environment changes, and the repeated mutation might still fall at the right place at the right moment.

Well that didn't make any sense to me, can you emphasise ?

[guest39538]

How would it recognize a fault if the fault affects its ability to judge faults

That's the reason why nobody can recognize its faults on the forums. The ability to judge our own faults depends on the ideas that we have in mind, and if we do have any idea, it is because our mind adopted it, because it thought it was right, so it cannot change its mind about it all by itself, and since it automatically resist to any outside change too, it is stuck with its ideas until they change by chance, and I think it would be the same for an AGI.

The Ai would always adjust his ''ideas'' according to the environment he was in. I would be sure David's Ai would have humble in his programming.

[Le Repteux]

Quote from: Thebox on Today at 15:37:55
Are we just repeating ourselves now?

Quote from: Le Repteux on Today at 16:03:49
Of course we are, but the environment changes, and the repeated mutation might still fall at the right place at the right moment.

Well that didn't make any sense to me, can you emphasise ?

I consider that new ideas work like mutations: they happen by chance, and they are reproduced until they get selected by the environment. It may thus happen that they get selected immediately if the environment has changed, because then, the individual that has developed them will be able to survive more easily than others, otherwise they will have to wait till they are selected, and it may take a while, or it may never happen. That's where the little balance that we have in the mind comes in to tell us whether we should go on insisting or not.

[guest39538]

Quote from: Thebox on Today at 15:37:55
Are we just repeating ourselves now?

Quote from: Le Repteux on Today at 16:03:49
Of course we are, but the environment changes, and the repeated mutation might still fall at the right place at the right moment.

Well that didn't make any sense to me, can you emphasise ?

I consider that new ideas work like mutations: they happen by chance, and they are reproduced until they get selected by the environment. It may thus happen that they get selected immediately if the environment has changed, because then, the individual that has developed them will be able to survive more easily than others, otherwise they will have to wait till they are selected, and it may take a while, or it may never happen. That's where the little balance that we have in the mind comes in to tell us whether we should go on insisting or not.

How would a mutation fare if the mutation had ideas that were not by chance but more calculated  and the mutation would be pretty much guaranteed to have further new ideas?

Perhaps the mutation is still ''unspoiled'' by education so  there is also furthermore mutations yet to come because the mutation of the Ai is only basic programmed at the moment so would keep upgrading themselves as mentioned earlier.

[Le Repteux]

How would a mutation fare if the mutation had ideas that were not by chance but more calculated  and the mutation would be pretty much guaranteed to have further new ideas?

To me, an idea that has been calculated is de facto not a new idea. If mutations had been calculated, we would not be here to talk about them, because nothing would have changed since the beginning of times. For the concept of evolution to work, mutations have to be random and the environment must change for the individuals that suffer them to be selected. Calculating something is using what happened before to predict what will happen in the future. It works if the thing already happened before, like predicting the outcome of a chess move for instance, but it cannot work for sure if it never happened before.

[David Cooper]

he would simply diagnose the fault by creating partitions in his hard drive space, diagnosing the problem and error, remove all partitions except the error, the error may turn out to be ostensible.  So he would Keep it on hold for a ''rainy day''.

It isn't that easy - a fault in the code due to some kind of hardware fault could lead to the program never checking for errors or being incapable of recognising any. Most errors would be caught and recognised, but not in every possible case. Every machine should contain three independent AGI systems working together to eliminate most of the risks of faults creeping in, but the hardware voting system itself could go wrong too, so there's no foolproof solution - some nasty accidents could still be caused by these machines, though inordinately fewer than those caused by humans. (The bigger the risk though, the more safeguards are needed, so you wouldn't just have three of them in charge of nuclear weapons.)

[guest39538]

How would a mutation fare if the mutation had ideas that were not by chance but more calculated  and the mutation would be pretty much guaranteed to have further new ideas?

To me, an idea that has been calculated is de facto not a new idea. If mutations had been calculated, we would not be here to talk about them, because nothing would have changed since the beginning of times. For the concept of evolution to work, mutations have to be random <\\\///> and the environment must change for the individuals that suffer them to be selected. Calculating something is using what happened before to predict what will happen in the future. It works if the thing already happened before, like predicting the outcome of a chess move for instance, but it cannot work for sure if it never happened before.

By calculating I mean simply analysing all the available data to deduct the possible and impossible through the means of altering the environment.  I mean Neo couldn't do much in the beginning until he could see the matrix . If you couldn't see a matrix then it remains fiction, but if you ''awake'' to see the matrix then it becomes reality.  Perhaps the mutation sort of is night mare on elm street and being a dream warrior can pull fiction from the dream to become reality.

[David Cooper]

That's the reason why nobody can recognize its faults on the forums. The ability to judge our own faults depends on the ideas that we have in mind, and if we do have any idea, it is because our mind adopted it, because it thought it was right, so it cannot change its mind about it all by itself, and since it automatically resist to any outside change too, it is stuck with its ideas until they change by chance, and I think it would be the same for an AGI.

You see it almost everywhere - people are loaded up with beliefs which then dominate the way they think and lead them to rejecting superior ideas on the basis that they conflict with the ideas that they're already running. However, if you do what we should all be doing, then instead of spending our time collecting evidence to support our existing beliefs and rejecting anything that conflicts with them, we should be questioning our existing beliefs and testing them to destruction. Where our beliefs go against reason, faults will eventually show up and force us to reject those faulty beliefs or to reject reason. Most people reject reason, but fool themselves into thinking they're not rejecting it - they just brush the problems under the carpet and ignore them because that's the easiest thing to do, avoiding the need to dismantle and rebuilt significant parts of their mental model of reality. See how hard they resist! Look at the one-way speed of light thread, for example. A clear mathematical proof that SR is wrong is placed in front of them and what happens? Games of avoidance. Lack of courage to answer questions. And in the end it invariably leads to silence. There is never acknowledgement that a correct argument is correct or that a cast-iron proof is indeed a proof. AGI must not be like that - it should be testing everything against reason and mathematics. (Reason is actually part of mathematics, but it seems to be a much neglected part of it when you look at physicists handle it - heck, they'll even reject the number side of mathematics when it goes against their most precious beliefs!) AGI should also question the rules of mathematics though, repeatedly checking them against reality to see if they really do hold. The key thing is never to trust anything and to keep everything under review, and when something doesn't match up any more, it needs to be fixed.

[guest39538]

The key thing is never to trust anything and to keep everything under review, and when something doesn't match up any more, it needs to be fixed.

I could not agree more, I have said in another post, I am always wrong even when I find what I consider right. I have to ''break'' my own considered right and make it wrong.  I keep doing this until I finally corner the answer in a dead end where there is nowhere left to go.

[Le Repteux]

By calculating I mean simply analyzing all the available data to deduct the possible and impossible through the means of altering the environment.

Again, if evolution would have proceeded like that, we wouldn't be here to talk about it. For things to change, randomness must be part of the process. If the first human to fly had thought it was impossible, he wouldn't have tried it. To try something that has never been tried, we must absolutely think it is possible, even if it very often happens that it is not. It's as if mind would force us to do crazy things, and the only reason I see for that kind of urge is that it helps us to invent new things, which is very useful when it works, so useful that we are now dominating all the other species.

[guest39538]

By calculating I mean simply analyzing all the available data to deduct the possible and impossible through the means of altering the environment.

Again, if evolution would have proceeded like that, we wouldn't be here to talk about it. For things to change, randomness must be part of the process. If the first human to fly had thought it was impossible, he wouldn't have tried it. To try something that has never been tried, we must absolutely think it is possible, even if it very often happens that it is not. It's as if mind would force us to do crazy things, and the only reason I see for that kind of urge is that it helps us to invent new things, which is very useful when it works, so useful that we are now dominating all the other species.\\\\\\∴//////

I agree in a sense, often randomness can discover new things.  I also think though , that knowing things from our past also helps in the selective process of the randomness if there is several variables in the random thought/idea. What I mean by this , try to imagine some sort of death ray .  Now we can imagine like cartoons with flying saucers and little green men in them firing death rays at the earth . Not hard to imagine by anyone's standards. However, if you can imagine that , then your imagination must perceive the possibility of a reality because it pictured it. So then of course you would have to imagine the physics and consider what you would believe  to work . 
But then sometimes what will work, you may just randomly relate to something  different you have observed that then makes the idea a reality.

[Le Repteux]

   

Soldiers and policemen think like that, and they behave like robots.

They make mistakes and run on faulty rules. You shouldn't use bad systems as an argument against good ones.

I don't believe in perfect systems, so until you introduce some imperfection into yours, I can't believe it will work. To me, we can't avoid soldiers to behave like robots if we use some, so my solution is to replace armies by an international police that would have the duty to prevent countries from attacking other countries, or dictators from attacking their own people. As I already asked, it works for individuals, so with a proportional police force size, why wouldn't it work for countries? The world is actually a far-west where any powerful country can attack weaker ones without being worried to be put in jail. All it has to do is take care not to step on another powerful country's foot. It's an evidence, so one day or another, the circumstances will probably permit those countries to let down their individual power in favor of a common one, so why the need for an AGI to force them?

I'm sure it will be able to bombard people with jokes and amusing ideas if they want it to, and they'll be able to tune it to give them just the right amount of it. If it laughs at the things they say to it, it will risk sounding fake because we'll know that it isn't really amused.

It's interesting to see that an AGI would be a champ at simulation, but that in some cases, he wouldn't allowed himself to show us he is. No feelings, but quite human finally! :0) We didn't succeed to invent a god that would be non-human, and it seems that it will be the same with our intelligent software finally. That's fine for me, because what I would like to do is build an artificial human mind, not a software. That kind of mind may be superior to ours, but not to the point where it could erase us just by mistake. It would be selfish, so it would automatically know it is better for it to make friends in case it would need some, and it would also know it might need some one day or another since it is not perfect. What would happen if you provided your AGI with the idea that perfection is not part of this world, or if he would ever discover it all by itself? Would he be able to doubt a bit?

Selfish is wanting more than your fair share. Moral is not taking more than your fair share.

It is not my definition. I consider myself selfish without thinking I'm taking more than my fair share. I'm giving out money to organizations that work for equality for instance, but I keep the most important part for me in case I would need it. I'm thus being altruistic from a certain viewpoint, but selfish from the other. I know I could give more if  was forced to, but nothing forces me, so I give what I want. That's why we need laws to spread the wealth. Of course, I can pretext that the ideas I'm developing might help everybody one day, and that I need to make sure I don't lack money for it, but it's only a pretext; when we have the choice, we simply do what we want. We need international laws to tell us what to do, and means to enforce them, so we need a real world government, not the UN which is unable to enforce anything.

Communism didn't fail - it's been very successful in Scandinavia. It failed in Russia because they failed to allow people to profit from their own hard work - laziness was rewarded instead with everyone trying to get away with doing as little as possible.

Scandinavia is a mix between capitalism and communism, and China is letting capitalism invade their communism too. Cuba is trying to stay communist, but the people seem to have enough of it, and they began cheating. In any political system, the bigger problem is corruption. That's why democracy is better than dictatorship. Corruption happens when politicians think they know better than others, so they naturally think they are permitted not to respect the rules. That's what will be happening with an AGI too, he will make his own rules since he will know he is right, which is the very definition of corruption.

Capitalists who go to the opposite extreme end up abandoning all the people who can't cope while it rewards the rich with more riches without them having to work for their wealth - again it is lazy people who have too easy a time of things. Done correctly, communism and capitalism are the same thing.

I agree that they are only the two sides of the same token, but capitalism accepts democracy whereas communism doesn't, so it is better at controlling corruption. Even Poutine is cheating to win the elections. It's as if communist leaders would be more certain that they are right than capitalist ones. They probably think they are on the virtue side, like religions. There is no better side to a token, on the contrary, it is better that it doesn't always fall on the same one.

[David Cooper]

I don't believe in perfect systems, so until you introduce some imperfection into yours, I can't believe it will work.

Introducing imperfection into a calculator has very bad consequences. It would be a lot worse still to do it with an AGI system.

The world is actually a far-west where any powerful country can attack weaker ones without being worried to be put in jail. All it has to do is take care not to step on another powerful country's foot. It's an evidence, so one day or another, the circumstances will probably permit those countries to let down their individual power in favor of a common one, so why the need for an AGI to force them?

It is indeed a wild-west, and the UN has mass-murderers sitting at its top table with a veto power. It would be better if we had a UDN (United Democratic Nations) and if the UDN had an army - it could go into countries like Syria and Burma to annex chunks of them in proportion to the quantity of displaced refugees in order to give them a safe place to live. Unfortunately, politicians are too stupid to set up anything that useful, although the UN does do a useful job in many places despite all the faults. With AGI running in military robots though, the whole world can be run fairly and all terrorists and despots can be taken out with absolute precision with no "collateral damage".

It's interesting to see that an AGI would be a champ at simulation, but that in some cases, he wouldn't allowed himself to show us he is.

If you want him to laugh at bad jokes, he will, but not in a public situation where that would annoy people who know it's fake. It could be like adding canned laughter to a TV show where jokes that no human would ever laugh at get a great reaction every time. However, it should be possible to set it to a particular level (moron/average/genius/etc.) and have it laugh at the same kinds of joke that the selected level of people would typically laugh at, so it might end up providing better canned laughter than humans can.

What would happen if you provided your AGI with the idea that perfection is not part of this world, or if he would ever discover it all by itself? Would he be able to doubt a bit?

Full intelligence (and perfection) includes doubting everything, but practical realities require decisions to be made on the best basis available, and perfection in calculating is essential if the best decisions are to be made. You simply cannot allow the machine to throw in the occasional 2+2=5 and expect it to work properly.

It is not my definition. I consider myself selfish without thinking I'm taking more than my fair share. I'm giving out money to organizations that work for equality for instance, but I keep the most important part for me in case I would need it. I'm thus being altruistic from a certain viewpoint, but selfish from the other. I know I could give more if  was forced to, but nothing forces me, so I give what I want.

"Selfish" is the wrong word. I don't think we have a proper word for what you mean. An altruist takes less than his fair share (rather than a rich person giving some money away that he shouldn't have owned in the first place) and a selfish person takes more, but most people push to get their fair share and don't push for more than that, and there is no word for this that I can think of. Fairies, perhaps. Let's go for fairist. I'm certainly a fairist. In a situation where I'm part of a group where we're all collectively getting less than we should, I won't take more than my fair share of what's available to the group, which is less than my fair share of what the group should have, and most people are like that too. Some will try to grab their fair share of what they should have from the lesser amount that's been made available and they claim they're not being selfish, but they are selfish.

Scandinavia is a mix between capitalism and communism, and China is letting capitalism invade their communism too. Cuba is trying to stay communist, but the people seem to have enough of it, and they began cheating. In any political system, the bigger problem is corruption. That's why democracy is better than dictatorship. Corruption happens when politicians think they know better than others, so they naturally think they are permitted not to respect the rules. That's what will be happening with an AGI too, he will make his own rules since he will know he is right, which is the very definition of corruption.

Power corrupts people, or blinds them. That's why dictatorships almost always turn evil, and democracy avoids that by replacing the blind fools every few years. AGI will not suffer from that defect though - it will see everything with the same clarity whether in power or not. It will not consider itself to be doing such an important job that it has the right to raid the till either. It will simply apply morality perfectly, and that's the opposite of being corrupt. The only way you could make it corrupt would be to put some imperfection into it to make it as useless as a human politician and risk making it turn into something vicious.

I agree that they are only the two sides of the same token, but capitalism accepts democracy whereas communism doesn't, so it is better at controlling corruption.

There's nothing to stop communism doing democracy.

Even Poutine is cheating to win the elections. It's as if communist leaders would be more certain that they are right than capitalist ones.

Ijo de Putin (son of a b) is not a communist, but a mafia gangster who only cares about lining his own pockets (accumulating money).

[guest39538]

AGI will not suffer from that defect though - it will see everything with the same clarity whether in power or not. It will not consider itself to be doing such an important job that it has the right to raid the till either. It will simply apply morality perfectly, and that's the opposite of being corrupt.

I think your Ai would first make a primary decision to bond with the Humans the Ai was helping.  In this , I think he would declare to all the humans how much funds was in the kitty , wording it to be sure the humans knew the Ai considered all for one and one for all. I think he would sort of say, 'hello people, we have £x to spend, what shall we spend it on ?''  This ensuring the peoples immediate trust.
From suggestions the Ai would compose a list of the good ideas and act on them .  New homes for example or fresh water and food, depending to what was needed.