The potential effect of Artificial Intelligence on civilisation - a serious discussion

I asked Chat GPT to write short article on how "acid-mesh" affected parachutes.
Chat GPT "blew it: by suggesting that "acid mesh" was beneficial to parachutes.
The problem is that the "acid mesh" problem came and went before the internet became popular, ergo few references to "acid mesh" available on the 'web.
OTL "acid mesh" was a problem with mosquito netting treated with fire-retardant chemicals. When exposed to certain combinations of heat, humidity and temperature, the mesh turned acidic and ate parachute fabric. The "eaten" fabric soon rotted out, ruining its structural integrity. The fabric failed and a couple of skydivers died during the mid-1980s. The problem was traced to a few bad batches of mesh and the suspected parachutes were grounded. Since only round reserve parachutes used large, meshed panels, sales of round reserves collapsed and by 1992, major skydiving dealers (e.g. Square One) quit selling new round reserves.
The Federal Aviation Administration issued Airworthiness Directives grounding the suspected parachute canopies and a couple of factories published "alternate means of compliance" to test ... maybe wash ... and re-certify old round reserves. I must have tested a thousand suspected canopies, but lost thee tools a decade ago and now no longer want anything to do with round canopies made during the acid-mesh era.
It is difficult for young parachute riggers to find the manufacturers' Service Bulletins, FAA's Airworthiness Directives, "alternate means of compliance" on the internet because the problem came and went before the internet became accessible to the great unwashed masses.
 
riggerrob said:
I asked Chat GPT to write short article on how "acid-mesh" affected parachutes.
Chat GPT "blew it: by suggesting that "acid mesh" was beneficial to parachutes.
The problem is that the "acid mesh" problem came and went before the internet became popular, ergo few references to "acid mesh" available on the 'web.
OTL "acid mesh" was a problem with mosquito netting treated with fire-retardant chemicals. When exposed to certain combinations of heat, humidity and temperature, the mesh turned acidic and ate parachute fabric. The "eaten" fabric soon rotted out, ruining its structural integrity. The fabric failed and a couple of skydivers died during the mid-1980s. The problem was traced to a few bad batches of mesh and the suspected parachutes were grounded. Since only round reserve parachutes used large, meshed panels, sales of round reserves collapsed and by 1992, major skydiving dealers (e.g. Square One) quit selling new round reserves.
The Federal Aviation Administration issued Airworthiness Directives grounding the suspected parachute canopies and a couple of factories published "alternate means of compliance" to test ... maybe wash ... and re-certify old round reserves. I must have tested a thousand suspected canopies, but lost thee tools a decade ago and now no longer want anything to do with round canopies made during the acid-mesh era.
It is difficult for young parachute riggers to find the manufacturers' Service Bulletins, FAA's Airworthiness Directives, "alternate means of compliance" on the internet because the problem came and went before the internet became accessible to the great unwashed masses.
Click to expand...
So what is the point here?
 
GTX said:
riggerrob said:
I asked Chat GPT to write short article on how "acid-mesh" affected parachutes.
Chat GPT "blew it: by suggesting that "acid mesh" was beneficial to parachutes.
The problem is that the "acid mesh" problem came and went before the internet became popular, ergo few references to "acid mesh" available on the 'web.
OTL "acid mesh" was a problem with mosquito netting treated with fire-retardant chemicals. When exposed to certain combinations of heat, humidity and temperature, the mesh turned acidic and ate parachute fabric. The "eaten" fabric soon rotted out, ruining its structural integrity. The fabric failed and a couple of skydivers died during the mid-1980s. The problem was traced to a few bad batches of mesh and the suspected parachutes were grounded. Since only round reserve parachutes used large, meshed panels, sales of round reserves collapsed and by 1992, major skydiving dealers (e.g. Square One) quit selling new round reserves.
The Federal Aviation Administration issued Airworthiness Directives grounding the suspected parachute canopies and a couple of factories published "alternate means of compliance" to test ... maybe wash ... and re-certify old round reserves. I must have tested a thousand suspected canopies, but lost thee tools a decade ago and now no longer want anything to do with round canopies made during the acid-mesh era.
It is difficult for young parachute riggers to find the manufacturers' Service Bulletins, FAA's Airworthiness Directives, "alternate means of compliance" on the internet because the problem came and went before the internet became accessible to the great unwashed masses.
Click to expand...
So what is the point here?
Click to expand...
My point is that ChatGPT is unlikely to understand a problem that came and went before the internet became public.
That is why I tell young parachute riggers that they are no required to repack parachutes older than themselves .... 'cus they are unlikely to find old manuals or Service Bulletins of Airworthiness Directives on line.
 
Last edited:
I remember the problem people had with Vulcan Phalanx and Seawolf being able to engage targets without human intervention.. So they tended initially to have this function turned off.
At the moment AI seems aimed at replicating what a human would do so it can be used instead.
This seems in line with how things have developed so far. Few people bother to write anymore or do calculations in their head.
The risk comes if you are foolish enough to build an AI system which controls its own power supply and communications with the outside world. This would be the Skynet moment.
 
Recently I learned, there are already AI programs for architecture. You can ask the program to design a building in a specific style with specific dimensions and you get a result. It is already being used, as said, just for inspiration… Honestly, if you take a look at newly built housing quartiers in Germany, this might result in more creativity than the typical small cube on top of a big cube which is seen here almost everywhere these days. It cant get worse!

I feel sorry for the architects, but I don’t expect the aesthetics will suffer from that, might be different in some other parts of the world, but there are more countries with terrible architecture (USA, Vietnam…)
 

Attachments

  • wuerfel.jpeg
    wuerfel.jpeg
    108.9 KB · Views: 21
riggerrob said:
I asked Chat GPT to write short article on how "acid-mesh" affected parachutes.
Chat GPT "blew it: by suggesting that "acid mesh" was beneficial to parachutes.
The problem is that the "acid mesh" problem came and went before the internet became popular, ergo few references to "acid mesh" available on the 'web.
OTL "acid mesh" was a problem with mosquito netting treated with fire-retardant chemicals. When exposed to certain combinations of heat, humidity and temperature, the mesh turned acidic and ate parachute fabric. The "eaten" fabric soon rotted out, ruining its structural integrity. The fabric failed and a couple of skydivers died during the mid-1980s. The problem was traced to a few bad batches of mesh and the suspected parachutes were grounded. Since only round reserve parachutes used large, meshed panels, sales of round reserves collapsed and by 1992, major skydiving dealers (e.g. Square One) quit selling new round reserves.
The Federal Aviation Administration issued Airworthiness Directives grounding the suspected parachute canopies and a couple of factories published "alternate means of compliance" to test ... maybe wash ... and re-certify old round reserves. I must have tested a thousand suspected canopies, but lost thee tools a decade ago and now no longer want anything to do with round canopies made during the acid-mesh era.
It is difficult for young parachute riggers to find the manufacturers' Service Bulletins, FAA's Airworthiness Directives, "alternate means of compliance" on the internet because the problem came and went before the internet became accessible to the great unwashed masses.
Click to expand...

Let me get this straight,
I, ruin your chutes
I, beat the life, out of your skydivers
And then you walk here, and you bring me more mesh ?
Brilliant plant, ese
...
You, have one thing wrong. This, is no mesh
...
KABOOOM !!!
 
Last edited:
Justo Miranda said:
GTX said:
vulcanpost.com

Cutting through the hype: ChatGPT is impressive, but it is not AI - and may never be

AI = an interface?
vulcanpost.com
Click to expand...
If that thing has been programmed to be politically correct there will not be much intelligence left in it, if the programmer pretends that the thing has an intelligent behavior after having suffered an ideological brainwashing he must also be quite stupid, but he will receive popularity and official support.
Click to expand...
Define "politically correct."

The only AI I have been fooling with is ChatGPT, which produces somewhat amusing output of mixed reliability and less than stellar quality.
 
Justo Miranda said:
GTX said:
vulcanpost.com

Cutting through the hype: ChatGPT is impressive, but it is not AI - and may never be

AI = an interface?
vulcanpost.com
Click to expand...
If that thing has been programmed to be politically correct there will not be much intelligence left in it, if the programmer pretends that the thing has an intelligent behavior after having suffered an ideological brainwashing he must also be quite stupid, but he will receive popularity and official support.
Click to expand...
The point is not the "political Correctness" as you put it but rather that these so-called AIs are not really Intelligent but rather are simply regurgitating what has already been created by humans within the bounds of their programming. They are undoubtedly 'smart' software (aided by 'dumb' humans :rolleyes: ) but truly 'intelligent' well probably not...

The following are probably some useful primers for this topic:

www.forbes.com

7 Types Of Artificial Intelligence

While artificial intelligence (AI) has become a commonly used and understood term, there is still a degree of obscurity regarding the different types of AI that exist and can exist in the future. Understanding the types of AI can help in better understanding the scope of present-day AI research.
www.forbes.com www.forbes.com

www.simplilearn.com

Artificial Intelligence Tutorial for Beginners

This Artificial Intelligence tutorial✔️helps you to learn artificial intelligence, its types, history, future, along with the applications. Read to know more
www.simplilearn.com www.simplilearn.com
 
I went to 'upper' school with a lot of bright kids, via what we'd now call 'STEM Stream'. Snag is while most were very good at memorising and recalling facts etc, rote-stuff, few bothered to read beyond the course texts, too few seemed interested or creative beyond narrow range...

Me, I was consuming current and back-copies of eg the weekly New Scientist, monthly Scientific American, Science Journal, Practical Electronics etc etc, soon exhausted the school library, routinely read a local library book an hour when I'd access. Summer, I'd get through several books an evening...

I was no genius, I wasn't especially talented, but I had a near-feline curiosity plus a quiet passion for information to help make sense of my crazy, crazy world...

Automation ? I began work as a QA/QC 'Bench Chemist', literally 'lab-coat + test-tube holder'. Technology infiltrated. Turned out I had a real knack for making such work...

As robotics etc arrived, we became more oriented towards caring for their quirks. Somewhat fewer of us were soon doing a dozen times the work.

What was the analogy circulated recently ? Used to be a plough-man would guide his team to cut a straight furrow or three. Now, he's 'riding herd' on a tractor-cab with enough electronics to daunt a jet-pilot...
 
GTX said:
Justo Miranda said:
GTX said:
vulcanpost.com

Cutting through the hype: ChatGPT is impressive, but it is not AI - and may never be

AI = an interface?
vulcanpost.com
Click to expand...
If that thing has been programmed to be politically correct there will not be much intelligence left in it, if the programmer pretends that the thing has an intelligent behavior after having suffered an ideological brainwashing he must also be quite stupid, but he will receive popularity and official support.
Click to expand...
The point is not the "political Correctness" as you put it but rather that these so-called AIs are not really Intelligent but rather are simply regurgitating what has already been created by humans within the bounds of their programming. They are undoubtedly 'smart' software (aided by 'dumb' humans :rolleyes: ) but truly 'intelligent' well probably not...

The following are probably some useful primers for this topic:

www.forbes.com

7 Types Of Artificial Intelligence

While artificial intelligence (AI) has become a commonly used and understood term, there is still a degree of obscurity regarding the different types of AI that exist and can exist in the future. Understanding the types of AI can help in better understanding the scope of present-day AI research.
www.forbes.com www.forbes.com

www.simplilearn.com

Artificial Intelligence Tutorial for Beginners

This Artificial Intelligence tutorial✔️helps you to learn artificial intelligence, its types, history, future, along with the applications. Read to know more
www.simplilearn.com www.simplilearn.com
Click to expand...
In my opinion an artificial intelligence could be a machine capable of reading everything that has been published about the theory of superstrings and branes and telling me in simple language if all that makes any sense in reality or is nothing more than nonsense.

Or a machine that could act as an interface between a quantum computer and a human to solve problems that theoretical physicists cannot.
 
Justo Miranda said:
GTX said:
vulcanpost.com

Cutting through the hype: ChatGPT is impressive, but it is not AI - and may never be

AI = an interface?
vulcanpost.com
Click to expand...
If that thing has been programmed to be politically correct there will not be much intelligence left in it, if the programmer pretends that the thing has an intelligent behavior after having suffered an ideological brainwashing he must also be quite stupid, but he will receive popularity and official support.
Click to expand...

Leaving aside politics, just how would one consistently and non-ideologically write the rules for "political correctness?"
 
Justo Miranda said:
If that thing has been programmed to be politically correct there will not be much intelligence left in it, if the programmer pretends that the thing has an intelligent behavior after having suffered an ideological brainwashing he must also be quite stupid, but he will receive popularity and official support.
Click to expand...

I'm sure there are people out there that would like to take advantage of Chat GPT for the sake of political agendas, as is usually the case when the wealthy purchase something that can influence so many people. Such dangerous power to have, especially when people are allowing themselves to be vulnerable when interacting with it, similar to learning in a classroom. It's very easy to influence a mind in these environments.
 
The interface problem was already brilliantly solved in Star Wars: a small, practical robot that gets along well with machines and an unnecessarily humanoid robot that translates for humans.
 
Creative artificial intelligence or artificial intelligence with memory? That's the question.
 

Attachments

  • 8178753613.jpg
    8178753613.jpg
    287.8 KB · Views: 17
Archibald said:
riggerrob said:
I asked Chat GPT to write short article on how "acid-mesh" affected parachutes.
Chat GPT "blew it: by suggesting that "acid mesh" was beneficial to parachutes.
The problem is that the "acid mesh" problem came and went before the internet became popular, ergo few references to "acid mesh" available on the 'web.
OTL "acid mesh" was a problem with mosquito netting treated with fire-retardant chemicals. When exposed to certain combinations of heat, humidity and temperature, the mesh turned acidic and ate parachute fabric. The "eaten" fabric soon rotted out, ruining its structural integrity. The fabric failed and a couple of skydivers died during the mid-1980s. The problem was traced to a few bad batches of mesh and the suspected parachutes were grounded. Since only round reserve parachutes used large, meshed panels, sales of round reserves collapsed and by 1992, major skydiving dealers (e.g. Square One) quit selling new round reserves.
The Federal Aviation Administration issued Airworthiness Directives grounding the suspected parachute canopies and a couple of factories published "alternate means of compliance" to test ... maybe wash ... and re-certify old round reserves. I must have tested a thousand suspected canopies, but lost thee tools a decade ago and now no longer want anything to do with round canopies made during the acid-mesh era.
It is difficult for young parachute riggers to find the manufacturers' Service Bulletins, FAA's Airworthiness Directives, "alternate means of compliance" on the internet because the problem came and went before the internet became accessible to the great unwashed masses.
Click to expand...

Let me get this straight,
I, ruin your chutes
I, beat the life, out of your skydivers
And then you walk here, and you bring me more mesh ?
Brilliant plant, ese
...
You, have one thing wrong. This, is no mesh
...
KABOOOM !!!
Click to expand...
Struth!
After the acid mesh was discovered, the skydiving industry largely quit making round parachutes with meshed drive-windows. Instead they converted to square (Jalbert) reserves with zero mesh.
So that only left 1 percent mesh in pilot-chutes.
Then Relative Workshop (now United Parachute Technologies) invented a couple of pilot-chutes with zero mesh.
 
Let face it, this AI tech is a double-edged sword,
you can use for good of mankind or worst of Mankind.

We live in interesting Times, like the Chinese curse say.
A time of transition, were old social system is replace by a New System.
like transition of Gothic to Renaissance/Reformation with all conflicts and Wars.
Here is it Cold War Era that replace by something new...

AI Tech will play important role in this transition during 21 century,
Allot Job normal during colt War will become obsolete with AI Tech.
Like assistants in Administration, Law firm and others will be replace by AI Tech.
Something similar will happen with Artist, illustrators and writers.
Imagine the next strike of Writers Guild of America and Hollywood switch to AI script...
although Ai Tech will bring improvement in Technology, like Aerospace or constructions.

But you could use this Technology for malicious means...
Perfect face recognition on Security Camera, precise profile of wanted persons.
Deep Fake of Person to black mail them or falsification of history for those who want total control !


But this not AI fault but fault the people using this technology for good or malicious means,
I think that was Frank Herbert warn us in DUNE about this AI Tech misuse, not about thinking machines...
 
Last edited:
The Guardian hackwriters tried to see if ChatGPT could write good TV plots and comedy lines and whether it could produce acceptable book blurbs for publishers. The results are pretty cringe. It makes you wonder if companies really will use it for more than minor tasks, at least at this stage.

www.theguardian.com

From horrific jokes about dog sex to ruining Succession: AI’s terrible attempt to write TV

We ask ChatGPT to recreate scenes from recent TV classics – and the results are hilariously dreadful. Hollywood execs may want to rethink that plan to have AI start writing our television…
www.theguardian.com www.theguardian.com

www.theguardian.com

‘Beowulf is lit AF’ – could ChatGPT really write good book blurbs?

Some in the book industry have already begun exploring automation of its pitches to readers. We took this functionality for a test drive
www.theguardian.com www.theguardian.com
 
Example of people using ChatGPT to essentially cheat at their work and being caught out:

https://twitter.com/x/status/1662308313525100546
View: https://twitter.com/d_feldman/status/1662308313525100546


This sort of thing shows one of the issues society needs to come to terms with using these so called AIs.

I also find it insightful to consider that people doing such things, especially if they are students, are really just cheating themselves. so you might sneak through on an assignment or similar but at the end of the day, you are the fool missing out since you missed out on actually learning.
 
GTX said:
I also find it insightful to consider that people doing such things, especially if they are students, are really just cheating themselves. so you might sneak through on an assignment or similar but at the end of the day, you are the fool missing out since you missed out on actually learning.
Click to expand...
Yes, but maybe, no. If what they are doing produces valid results, then it's probable that in the future that's how it'll be done. When I was in junior high lo these many decades back, I had a drafting teacher who thought that CAD drafting was cheating, that you needed to learn how to use the T-square and do proper lettering and all. I haven't used those skills since the early 90's. Soon it's likely that the "learning" that legal types have needed to date won't necessarily be needed.
 
Orionblamblam said:
GTX said:
I also find it insightful to consider that people doing such things, especially if they are students, are really just cheating themselves. so you might sneak through on an assignment or similar but at the end of the day, you are the fool missing out since you missed out on actually learning.
Click to expand...
Yes, but maybe, no. If what they are doing produces valid results, then it's probable that in the future that's how it'll be done. When I was in junior high lo these many decades back, I had a drafting teacher who thought that CAD drafting was cheating, that you needed to learn how to use the T-square and do proper lettering and all. I haven't used those skills since the early 90's. Soon it's likely that the "learning" that legal types have needed to date won't necessarily be needed.
Click to expand...

I think your criticism is too superficial. It assumes that the heavy lifting in future will be performed by an AI as opposed to using your brain. I briefly taught a class in cartooning with a friend. Prior to this, I had thought teaching was the easiest job in the world: sit behind a desk, write a few things on the blackboard, hand out assignments and tests, and you were done. Not so. I saw some students struggling and found myself guiding them through things they didn't understand. Their problems were not cookie-cutter simple. Each student needed a different set of instructions to see what the problem was and a little encouragement to solve it.

In high school, I once did a drawing of a rocket and showed it to the drafting teacher. I did not take drafting but he did comment that it was better than what some of his students were handing in.

I was also the letterer on an independently published comic book and did some pencils. It's hard. Always will be. The best art books I had always encouraged freehand drawing on actual paper, so when it came time to put something together, you understood the fundamentals well enough to do a good job.

So whether it's drafting or legal work, you need self-discipline, you need to know how to do research properly. You need to understand it. Not let your brain atrophy. To not let AI become a crutch to lean on. That would be stupid. Developing your mind matters. Learning how to solve complex problems matters. If anyone thinks that being lazy is OK because AI will just spit out the answer, just remember that that kind of thinking leads to Skynet being built. They already have a Terminator prototype called Figure 01. And they're not afraid to use it.
 
edwest4 said:
So whether it's drafting or legal work, you need self-discipline, you need to know how to do research properly. You need to understand it. Not let your brain atrophy. To not let AI become a crutch to lean on.
Click to expand...

I saw the same arguments in school re: calculus. "You'll need to be able to do this to be an engineer!" Then I got out into the world and found that *nobody* used calculus. That's what computers were for.
 
Orionblamblam said:
edwest4 said:
So whether it's drafting or legal work, you need self-discipline, you need to know how to do research properly. You need to understand it. Not let your brain atrophy. To not let AI become a crutch to lean on.
Click to expand...

I saw the same arguments in school re: calculus. "You'll need to be able to do this to be an engineer!" Then I got out into the world and found that *nobody* used calculus. That's what computers were for.
Click to expand...

That's what Skynet is for. Got it...
 

Similar threads

Back
Top Bottom