The problem of designing an AI to understand what it sees

@kellyjay said

Evolution can only take what is there and change it, the changes we can see are very small and do not add up to a new organ or over time a system.

False; A series of small changes, whether specifically via evolution or just generally, eventually can make a large change.
How can this logically NOT be true?
If that's false then a continuous slow increase or a long series of incremental increases, thus series of small changes, in water level of a water reservoir can never make the large change of filling the whole reservoir!
Exactly what part of the concept of a series of small changes adding up to a large change do you have a problem with or don't understand?

@humy said
False; A series of small changes, whether specifically via evolution or just generally, eventually can make a large change.
How can this logically NOT be true?
If that's false then a continuous slow increase or a long series of incremental increases, thus series of small changes, in water level of a water reservoir can never make the large change of filling the whole reservoi ...[text shortened]... series of small changes adding up to a large change do you have a problem with or don't understand?

How can it not be logically valid, seriously? Do you think life is a computer game in that you always get good functional changes, and once you have them, it remains true you’d get to keep them? Not only do the changes have to be good and functionally useful, but they must also build upon a previously good one! Small changes to develop a heart through time means all the changes required for a heart much stay intact throughout the process. You ever looked at the math in what you just said, or do you take it for granted that is true? If you admit that we are talking about a finite period, this limits the number of changes that can occur in a period. How many evolutionary changes do you think were required from a single cell to a human? If you factor in the percentages of known harmful mutations compared to good ones, then attempt to figure the odds on specific changes building upon other changes, do we see numbers guaranteeing success, or we looking at something else entirely, like less than a snowball chance in hell.

This isn't even taking into account environment changes, or food supply shortages, and so on that could wipe out life.

@kellyjay said
Abiogenesis and evolution are two different things, for abiogenesis to be true many things had to have occurred and that is no different for you and AI before we get to altering code. The hardware has to be created, the software you use has be loaded on the computer to write the code you are going to use to create what you call AI. I can write scripts in PERL and did in my o ...[text shortened]... n faith and science, we can validate in science you are only making claims when it comes to fossils.

Download one of the Linux distributions, you get all this stuff for free without the pain of trying to set it up on Windows.

@kellyjay said
Not only do the changes have to be good and functionally useful, but they must also build upon a previously good one!

Even if that was always true, so what?
Actually, its false.

Small changes to develop a heart through time means all the changes required for a heart much stay intact throughout the process.

What do you mean by changes "much stay intact throughout the process"? "intact"? What would a change that "isn't intact" mean? What are you talking about?

If you admit that we are talking about a finite period, this limits the number of changes that can occur in a period.

Right. So what?

How many evolutionary changes do you think were required from a single cell to a human?

Many. So what?

If you factor in the percentages of known harmful mutations compared to good ones

Why? The harmful ones are irrelevant here as they are simply weeded out by evolution. There are plenty of beneficial ones and its the absolute number of beneficial ones that counts and not the harmful-to-beneficial ratio that counts because, just as I just said, the harmful ones are irrelevant here as they are simply weeded out by evolution.

then attempt to figure the odds on specific changes building upon other changes, do we see numbers guaranteeing success

"Success" of doing what, exactly? There is no 'goal' or 'purpose' or 'ultimate end' for evolution to have "success" of 'achieving' and if you think there is supposed to be then you are confused. Evolution isn't 'trying' to achieve some particular end result.

This isn't even taking into account environment changes, or food supply shortages, and so on that could wipe out life.

Environment changes have often wiped out most of life short of all life. What about it? These disastrous events that cause mass extinctions actually STIMULATE evolution by opening up new niches in their aftermath for new species to evolve to fill so, if anything, this HELPS the evolution process. Without such disastrous events we would probably never come to exist.

@deepthought said
Download one of the Linux distributions, you get all this stuff for free without the pain of trying to set it up on Windows.

Thank you for that, if it were not for that being a work computer I'd would. I'm not going to add to it something like this without it be approved.

@humy said
Even if that was always true, so what?
Actually, its false.

Small changes to develop a heart through time means all the changes required for a heart much stay intact throughout the process.

What do you mean by changes "much stay intact throughout the process"? "intact"? What would a change that "isn't intact" mean? What are you talking about?
[quote] If yo ...[text shortened]... s HELPS the evolution process. Without such disastrous events we would probably never come to exist.

I think you have turned off your brain when it comes to this topic. Your telling me that overtime for the development of a heart DNA doesn't require that all of the necessary alterations to produce a heart don't need to occur and remain in place so a heart can arise?

@kellyjay said
Your telling me that overtime for the development of a heart DNA doesn't require that all of the necessary alterations to produce a heart don't need to occur and remain in place so a heart can arise?

No, I am NOT "telling" you that "doesn't require" that.
Where did you get that from?
I said nothing even vaguely like that and in fact implied the exact opposite because I clearly implied "all of the necessary alterations to produce a heart" DO (and via evolution, obviously) "need to occur and remain in place so a heart can arise" and that that's exactly what has happened (via evolution).
And why would the genetic alterations to "produce a heart" NOT "remain in place"? (if that's what you are implying here? If not, what did you mean by that? ) Are you saying NO mutation can be passed on to the next generation? Because, if so, I can show you good evidence to the contrary. Explain what you meant to say...

@humy said
No, I am NOT "telling" you that "doesn't require" that.
Where did you get that from?
I said nothing even vaguely like that and in fact implied the exact opposite because I clearly implied "all of the necessary alterations to produce a heart" DO (and via evolution, obviously) "need to occur and remain in place so a heart can arise" and that that's exactly what has happened (v ...[text shortened]... tion? Because, if so, I can show you good evidence to the contrary. Explain what you meant to say...

KellyJay: "Not only do the changes have to be good and functionally useful, but they must also build upon a previously good one!"


humy: "Even if that was always true, so what?
Actually, its false. "

I said they had to, you said it was false, and now you are in denial? Let me know when you have your story straight, I'm not sure what your views are at the moment.

@kellyjay said
How can it not be logically valid, seriously? Do you think life is a computer game in that you always get good functional changes, and once you have them, it remains true you’d get to keep them? Not only do the changes have to be good and functionally useful, but they must also build upon a previously good one! Small changes to develop a heart through time means all the chan ...[text shortened]... king into account environment changes, or food supply shortages, and so on that could wipe out life.

"If you admit that we are talking about a finite period, this limits the number of changes that can occur in a period."

But 4.5 billion years is an incomprehensibly long period of time. Mammals only appeared sometime in the last 10 million years, and Homo erectus ~ 2 million, a small sliver of which comprises recorded history. Also consider that every adult person needs somewhere between 50-70 billion cell divisions per day just to maintain homeostasis. Humans are relatively new to this world and yet we've probably had around 350,000 ancestral generations as a species. Despite the exquisite fidelity of the DNA replication machinery, there is ~ one mistake made every cell division and a fertile male produces 1800 million sperm. That's a lot of opportunity.

"How many evolutionary changes do you think were required from a single cell to a human?"

Your snowball is a good analogy. Two small snowballs rolling down a hill acquire new snowflakes as they go. One snowball stochastically outcompetes the other by accumulating more snowflakes, so it keeps rolling while the other peters out.

@wildgrass said
"If you admit that we are talking about a finite period, this limits the number of changes that can occur in a period."

But 4.5 billion years is an incomprehensibly long period of time. Mammals only appeared sometime in the last 10 million years, and Homo erectus ~ 2 million, a small sliver of which comprises recorded history. Also consider that every adult person needs ...[text shortened]... tcompetes the other by accumulating more snowflakes, so it keeps rolling while the other peters out.

4.5 billion years is a finite period, what it is to us may be incomprehensible, but that doesn't change the fact it has a beginning and end. During that time period, how much time are we even talking about concerning the window of opportunities for life? There will be no chances for life if all the necessary conditions for it are not present at the same time! If one is not present, that would be a show stopper then all of that time would not count; it wouldn’t matter.

Keep in mind just because a second goes by that doesn't mean there was the possibility that something could occur that could start life! Only during those times that everything is ready would it even be possible for an attempt at success, an attempt is not a guarantee there would now be life. It isn’t clear it could happen in the natural world if everything was in play. Billions of years would not matter if all the necessary ingredients for life were not present, circular reasoning might get you there, but don’t we require something more substantial for verification? That Chemist said science and faith have different standards; in science, we should be able to show evidence, what claim we know should be repeatable, and right now we know chemical reactions will continue till they reach their conclusion, that we know now does not look suitable for life coming out of the natural world!

@wildgrass said
"If you admit that we are talking about a finite period, this limits the number of changes that can occur in a period."

But 4.5 billion years is an incomprehensibly long period of time. Mammals only appeared sometime in the last 10 million years, and Homo erectus ~ 2 million, a small sliver of which comprises recorded history. Also consider that every adult person needs ...[text shortened]... tcompetes the other by accumulating more snowflakes, so it keeps rolling while the other peters out.

You are not answering my question, from a single cell, how many mutations are required to go from q one cell life, to a human being? If it requires a lot more time than the period of time they were first seen in, that then becomes a problem, wouldn't it?

@humy said
Even if that was always true, so what?
Actually, its false.

Small changes to develop a heart through time means all the changes required for a heart much stay intact throughout the process.

What do you mean by changes "much stay intact throughout the process"? "intact"? What would a change that "isn't intact" mean? What are you talking about?
[quote] If yo ...[text shortened]... s HELPS the evolution process. Without such disastrous events we would probably never come to exist.

"What do you mean by changes "much stay intact throughout the process"? "intact"? What would a change that "isn't intact" mean? What are you talking about? "

Just because a positive mutation occurs, that doesn't mean something else cannot happen that voids its contribution. When adding up only good mutations without taking into account all the life ending bad ones is stacking the deck in an unrealistic way.

@kellyjay said
4.5 billion years is a finite period, what it is to us may be incomprehensible, but that doesn't change the fact it has a beginning and end. During that time period, how much time are we even talking about concerning the window of opportunities for life? There will be no chances for life if all the necessary conditions for it are not present at the same time! If one is not ...[text shortened]... their conclusion, that we know now does not look suitable for life coming out of the natural world!

"circular reasoning might get you there"

we have a fossil record.

@kellyjay said

Just because a positive mutation occurs, that doesn't mean something else cannot happen that voids its contribution.

Such as what, exactly?
Give me just one example of this "something else" that is observed to or at least assumed to sometimes happen to "voids its contribution" of a "positive mutation" ...

@wildgrass said
"circular reasoning might get you there"

we have a fossil record.

Yes, and it tells you what?