Evolution not limited to life on earth?

[Luc Turpin]

https://phys.org/news/2023-10-scientists-philosophers-nature-evolutionary-law.html

 

[iNow]

4 minutes ago, Luc Turpin said:

https://phys.org/news/2023-10-scientists-philosophers-nature-evolutionary-law.html

 

Will you please summarize the key points in the article and introduce what about it you wish to discuss with others here?

Basically, help readers to follow along here without explicitly clicking the link, while also not treating thread forum as a blog?

[Luc Turpin]

Presented as a general interest piece for those wishing to know more about evolution and its ramifications.

Hope that by posting, I am not treating it as a blog!

Thank you for the guidance; I am new at it.

[iNow]

3 minutes ago, Luc Turpin said:

for those wishing to know more about evolution and its ramifications.

For instance?

[Phi for All]

32 minutes ago, Luc Turpin said:

Presented as a general interest piece for those wishing to know more about evolution and its ramifications.

!

Moderator Note

This isn't a classroom either. It's a discussion forum. What would you like to discuss regarding evolution on other planets? Please design opening posts to encourage an interesting conversation, in this case on a specific aspect of this article.

 

[swansont]

!

Moderator Note

To rephrase/repeat what others have said: we expect a link and an excerpt of the article when posting in News. Discussion should center around that.

Discussion of invariance of scale has been split:

https://www.scienceforums.net/topic/133088-invariance-of-scale-split-from-evolution-not-limited-to-life-on-earth/

 

[Moontanman]

On 12/18/2023 at 8:33 AM, Luc Turpin said:

https://phys.org/news/2023-10-scientists-philosophers-nature-evolutionary-law.html

 

OK, I read the link and all i can say is... DUH! Change occurs over time, what part of that would lead anyone to believe that evolution is confined to the Earth? 

[AIkonoklazt]

Oh boy...

Quote

At a time when evolving AI systems are an increasing concern, a predictive law of information that characterizes how both natural and symbolic systems evolve is especially welcome

...They just HAD to add that stupid bit in, didn't they?

No. AI systems don't evolve and can't EVER evolve, because the process of evolution isn't that of design. There's no teleology involved in evolution, unlike any and all artifacts. Artifacts involve teleology by definition https://www.merriam-webster.com/dictionary/artifact (substitute "alien," "green little men," etc for "human" if you want)

Even so-called "evolutionary algorithms" come with fitness functions that have been programmed by a person https://en.wikipedia.org/wiki/Fitness_function

You can't just kick the can of programming down the road until it disappears into a hand-waving rhetorical background.

[iNow]

Of course AI are not exchanging DNA, but they ARE rejecting flawed forecasts and reinforcing successful ones… I.e. engaging in a form of selection, aka: evolving… and this is happening quite rapidly. 

But your record of having an enormous chip on your shoulder about all things AI remains unblemished. 

[swansont]

2 hours ago, AIkonoklazt said:

Oh boy...

...They just HAD to add that stupid bit in, didn't they?

It's not stupid if you consider that evolution is merely change over time. 

We speak of language evolving, or technology evolving. Of course AI is going to evolve. It just won't necessarily be Darwinian.

2 hours ago, AIkonoklazt said:

No. AI systems don't evolve and can't EVER evolve, because the process of evolution isn't that of design.

It's not at all apparent to me that, even if one restricted this to the narrow definition, that computer code COULD NOT be written to modify itself.

What is the restriction making this impossible?

 

[AIkonoklazt]

2 minutes ago, swansont said:

It's not stupid if you consider that evolution is merely change over time. 

We speak of language evolving, or technology evolving. Of course AI is going to evolve. It just won't necessarily be Darwinian.

It's not at all apparent to me that, even if one restricted this to the narrow definition, that computer code COULD NOT be written to modify itself.

What is the restriction making this impossible?

 

Ah therein lays the rub. Why not use the term "change over time" instead of (especially in "news articles") "evolving" (complete with misleading pictures to boot? https://www.science.org/content/article/artificial-intelligence-evolving-all-itself

"Written to modify itself" makes as much sense as "program something so it doesn't need programming." It's self-contradictory.

[swansont]

3 minutes ago, AIkonoklazt said:

Ah therein lays the rub. Why not use the term "change over time" instead of (especially in "news articles") "evolving" (complete with misleading pictures to boot? https://www.science.org/content/article/artificial-intelligence-evolving-all-itself

Why not use a word that means change over time? Really?

3 minutes ago, AIkonoklazt said:

"Written to modify itself" makes as much sense as "program something so it doesn't need programming." It's self-contradictory.

Where is the contradiction?

[AIkonoklazt]

3 minutes ago, swansont said:

Why not use a word that means change over time? Really?

Really. In fact, nothing wrong with the title of "Artificial intelligence changing all by itself."

4 minutes ago, swansont said:

Where is the contradiction?

Good grief. When you program something, you obviously "needed" to program it.

[swansont]

1 minute ago, AIkonoklazt said:

Good grief. When you program something, you obviously "needed" to program it.

So what? You’re rebutting a claim nobody has made. 

[AIkonoklazt]

6 minutes ago, swansont said:

So what? You’re rebutting a claim nobody has made. 

It is implicit in the claim of "AI evolution," especially with "news articles" like the one I just pointed to.

[StringJunky]

7 hours ago, AIkonoklazt said:

It is implicit in the claim of "AI evolution," especially with "news articles" like the one I just pointed to.

Are you in the Luddite club, do you fear the future? You are coming across like a kid with his fingers in ears and eyes shut, bawling.

Edited December 26, 2023 by StringJunky

[swansont]

13 hours ago, AIkonoklazt said:

It is implicit in the claim of "AI evolution," especially with "news articles" like the one I just pointed to.

Counterpoint: no, it's not. You choose to interpret it that way, which is followed by ranting about how stupid the notion is. But it's your choice. Even in biological evolution, the origin of life is excluded from the theory - that's abiogenesis. So your insistence that a program has to create itself is just performative nonsense.

[AIkonoklazt]

12 hours ago, swansont said:

Counterpoint: no, it's not. You choose to interpret it that way, which is followed by ranting about how stupid the notion is. But it's your choice.

Conterpoint: YES, IT IS. See again the article I pointed to earlier. It claims, specifically:

Quote

AI programs that improve generation after generation without human input.

There isn't such a thing. The "human input" in evolutionary algorithms is the required fitness function.

Even if you change the article's title to "Artificial intelligence is changing all by itself," it's still an untrue statement.

Quote

Even in biological evolution, the origin of life is excluded from the theory - that's abiogenesis. So your insistence that a program has to create itself is just performative nonsense.

The point isn't the origin of the object in question but the origin of an object's purported teleology. Natural evolution never involve teleology (purposeful design), while artifacts inevitably do.

18 hours ago, StringJunky said:

Are you in the Luddite club, do you fear the future? You are coming across like a kid with his fingers in ears and eyes shut, bawling.

Are you someone who absolutely doesn't know how to argue? You're just making personal insults instead of making any argument.

[exchemist]

7 hours ago, AIkonoklazt said:

Conterpoint: YES, IT IS. See again the article I pointed to earlier. It claims, specifically:

There isn't such a thing. The "human input" in evolutionary algorithms is the required fitness function.

Even if you change the article's title to "Artificial intelligence is changing all by itself," it's still an untrue statement.

The point isn't the origin of the object in question but the origin of an object's purported teleology. Natural evolution never involve teleology (purposeful design), while artifacts inevitably do.

Are you someone who absolutely doesn't know how to argue? You're just making personal insults instead of making any argument.

Just trying to understand your point here, is it that what would constitute "improvement" in the program could only be defined by its human users, so the AI would not be able to improve itself without human feedback as to what to change? 

[dimreepr]

10 hours ago, AIkonoklazt said:

Conterpoint: YES, IT IS. See again the article I pointed to earlier. It claims, specifically:

There isn't such a thing. The "human input" in evolutionary algorithms is the required fitness function.

Even if you change the article's title to "Artificial intelligence is changing all by itself," it's still an untrue statement.

The point isn't the origin of the object in question but the origin of an object's purported teleology. Natural evolution never involve teleology (purposeful design), while artifacts inevitably do.

And we're back to the same handwaving and bs arguments you presented in your other thread on the subject; basically it's blind faith in the ineffability of humanity, through some sort of God, as far as I can tell... 

10 hours ago, AIkonoklazt said:

Are you someone who absolutely doesn't know how to argue? You're just making personal insults instead of making any argument.

Well, that's another iron-e-meter bolloxed FFS, I wish they could design one that could evolve past, throwing in the towel; you don't know the difference between an argument and gainsaying, look em up...

Better still:

 

Edited December 27, 2023 by dimreepr

[swansont]

9 hours ago, AIkonoklazt said:

Conterpoint: YES, IT IS. See again the article I pointed to earlier. It claims, specifically:

There isn't such a thing. The "human input" in evolutionary algorithms is the required fitness function.

Even if you change the article's title to "Artificial intelligence is changing all by itself," it's still an untrue statement.

Arguing that a program couldn't include the capability of modifying its code or function is ludicrously ignorant.

 

9 hours ago, AIkonoklazt said:

The point isn't the origin of the object in question but the origin of an object's purported teleology. Natural evolution never involve teleology (purposeful design), while artifacts inevitably do.

The argument is that a program could include the capability of changing its code. Teleology is a straw man.

[Endy0816]

@AIkonoklazt

Don't need a fitness function or similar, though is a saner way to do things.

Any random string of numbers can be run as a program.

[AIkonoklazt]

10 hours ago, Endy0816 said:

@AIkonoklazt

Don't need a fitness function or similar, though is a saner way to do things.

Any random string of numbers can be run as a program.

This random string is still an input.

11 hours ago, swansont said:

Arguing that a program couldn't include the capability of modifying its code or function is ludicrously ignorant.

The capability came from human input. Terms like "self-driving car" or "self-modification / self-production" of any artifact are ignorant misattributions of agency (see the court case I mention below. The judge himself claimed ignorance on the subject)

Quote

The argument is that a program could include the capability of changing its code. Teleology is a straw man.

...That capability came from human beings. Artifacts are not self-actuated entities. Please, let's slow down a moment and take a look at my linked article again, with its headline and (to me) badly chosen title illustration: https://www.science.org/content/article/artificial-intelligence-evolving-all-itself . Which kinds of underlying messages are those two things sending as a pair? I see so many news articles that do the same thing. Public perception is at once being warped and reinforced.

Yes, I am arguing the semantics, and the semantics is important because that's exactly what policy hinges upon much of the time. Take a look at what happened in this Australian court case: https://www.theguardian.com/technology/2021/jul/30/im-sorry-dave-im-afraid-i-invented-that-australian-court-finds-ai-systems-can-be-recognised-under-patent-law (the same person is bringing a whole bunch of lawsuits into courtrooms all over the world, saying in an interview that the reason is he wants AI to be accepted by the public as people) AI legal personality was already banned by UN agency UNESCO in 2021, but another UN agency is gearing up for another around of guidelines without stating whether it would even take the previous precedent into consideration or not. I'm writing my own recommendation to them right now.

Edit: In addition, the South African case https://ipwatchdog.com/2021/07/29/dabus-gets-first-patent-south-africa-formalities-examination/id=136116/

Edited December 28, 2023 by AIkonoklazt
South African patent

[iNow]

46 minutes ago, AIkonoklazt said:

Yes, I am arguing the semantics, and the semantics is important because that's exactly what policy hinges upon much of the time

Nobody here except you has mentioned policy, AI or otherwise. The thread regards whether evolution is an earth only process. 

[Endy0816]

@AIkonoklazt

There are hardware RNG's.

https://en.m.wikipedia.org/wiki/Hardware_random_number_generator

Beyond activation no human needs to be involved