ChatGPT, and GPT-4, and AI in our futures.

Everything except LibriVox (yes, this is where knitting gets discussed. Now includes non-LV Volunteers Wanted projects)
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

redrun wrote: April 18th, 2023, 5:04 pm For most folks in the US, there're only a few instances where copying your voice would get someone in trouble.
A song titled ‘heart on my sleeve’ got more than 15 million plays on TikTok using AI voices of Drake and The Weeknd.

It was without permisson and was so far removed. rich people are still trying to figure out how to craft laws so they get the most money and poor people get nothing, so AI laws are sketchy at this stage..
TriciaG
LibriVox Admin Team
Posts: 60747
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG »

Well, this was a nice, cheery thread to wake up to this morning. :?
School fiction: David Blaize
America Exploration: The First Four Voyages of Amerigo Vespucci
Serial novel: The Wandering Jew
Medieval England meets Civil War Americans: Centuries Apart
msfry
Posts: 11703
Joined: June 4th, 2013, 9:09 am
Location: Baton Rouge, Louisiana
Contact:

Post by msfry »

TriciaG wrote: May 21st, 2023, 5:38 am Well, this was a nice, cheery thread to wake up to this morning. :?
Reality, unlike fantasy, is not always cheery. That said, I highly recommend Librivox to steer away from political conversation, even if, as Elon conjectures, AI may destroy us all in the near future. Make-believe, fables and fantasy are far more effective in keeping people happy, and there are plenty of those left to record.
annise
LibriVox Admin Team
Posts: 38650
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

Proving they copied your voice would be hard.
Please all respect other people's opinions, especially about contentious subjects. AI exists, like all human inventions it will be used for good, for profit making. and for evil and all mixtures of those three. How we each chose to deal with it is something we have to decide ourself. And neither good nor evil have the same meaning for everyone.

Anne
msfry
Posts: 11703
Joined: June 4th, 2013, 9:09 am
Location: Baton Rouge, Louisiana
Contact:

Post by msfry »

annise wrote: May 21st, 2023, 10:03 pm Proving they copied your voice would be hard.
Please all respect other people's opinions, especially about contentious subjects. AI exists, like all human inventions it will be used for good, for profit making. and for evil and all mixtures of those three. How we each chose to deal with it is something we have to decide ourself. And neither good nor evil have the same meaning for everyone.

Anne
That's good advice. :mrgreen:
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

What luck, there was an addendum to this, food for thought, which I thought everyone would be happy with. There was, on a personal note, a youtuber I couldn't recall the name of that I've been searching for for ages ! but couldn't remember. Well, presto I found what I was looking for and what I wasn't looking for but was very glad to find ! :D

The author is CGPGrey and the summary is called Humans Need Not Apply it is video id 7Pq-S557XQU

CGPGrey Humans Need Not Apply

worth a serch, not sure if linking direct is appropriate.

I love the anology of the horses, it's spot on and cute, and fodder for thought. I mean, food.
msfry
Posts: 11703
Joined: June 4th, 2013, 9:09 am
Location: Baton Rouge, Louisiana
Contact:

Post by msfry »

I found this 27 minute youtube tutorial this morning, which I found easy to follow and very informative. You don't have to download nor sign up for anything to see many examples of how it works.
ChatGPT Tutorial: How to Use Chat GPT For Beginners 2023

If you watch it I'd like to hear your opinions, and how you might use it yourself.

I don't mind saying, I think it's pretty creepy, extremely impersonal, and my worst fear it will dumb down our youths, who will henceforth never learn now to generate original content, organize their thoughts, write coherent letters, essays, novels, etc. Imagine getting a "heartfelt" computer generated love letter, prayer or apology. Blehhhhhh!

In time it will also begin to aggregate "preferred" content, leaving out material "big daddy" or some other tyrant deems "undesirable", and who will be the wiser? Wikipedia, Google, most media outlets already do this. On the other hand, it will create a thriving underground market of watchdogs and whistleblowers.
annise
LibriVox Admin Team
Posts: 38650
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

Dulcamara
Posts: 7566
Joined: December 23rd, 2020, 1:14 pm
Location: Barataria

Post by Dulcamara »

annise wrote: May 29th, 2023, 8:15 pm you might find this interesting
https://blog.archive.org/2023/05/29/let-us-serve-you-but-dont-bring-us-down/

Anne
Oops... in my ignorance at first I thought it was just a denial of service attack. No, it's not.

Very disquieting because of the implications, but it is good to know.

Thank you for sharing this, Anne. :)
Jasna

Second to the right and then straight on till morning.
Dulcamara
Posts: 7566
Joined: December 23rd, 2020, 1:14 pm
Location: Barataria

Post by Dulcamara »

msfry wrote: May 29th, 2023, 12:20 pm I found this 27 minute youtube tutorial this morning, which I found easy to follow and very informative. You don't have to download nor sign up for anything to see many examples of how it works.
ChatGPT Tutorial: How to Use Chat GPT For Beginners 2023

If you watch it I'd like to hear your opinions, and how you might use it yourself.

I don't mind saying, I think it's pretty creepy, extremely impersonal, and my worst fear it will dumb down our youths, who will henceforth never learn now to generate original content, organize their thoughts, write coherent letters, essays, novels, etc. Imagine getting a "heartfelt" computer generated love letter, prayer or apology. Blehhhhhh!

In time it will also begin to aggregate "preferred" content, leaving out material "big daddy" or some other tyrant deems "undesirable", and who will be the wiser? Wikipedia, Google, most media outlets already do this. On the other hand, it will create a thriving underground market of watchdogs and whistleblowers.
Hi, Michele. Thank you for the link. I saw it yesterday and am still processing... :D
Jasna

Second to the right and then straight on till morning.
msfry
Posts: 11703
Joined: June 4th, 2013, 9:09 am
Location: Baton Rouge, Louisiana
Contact:

Post by msfry »

annise wrote: May 29th, 2023, 8:15 pm you might find this interesting
https://blog.archive.org/2023/05/29/let-us-serve-you-but-dont-bring-us-down/

Anne
Inevitably, traffic on all servers will increase exponentially, especially on "free" sites like IA, Wikipedia, Libraries -- anyplace that stores information. To slow inquiries, especially idle ones, they may have to start charging for access, or require membership, and/or student/media/teacher credentials. Best to brace for it, and prepare for it.
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

msfry wrote: May 30th, 2023, 7:06 am Inevitably, traffic on all servers will increase exponentially, especially on "free" sites like IA, Wikipedia, Libraries -- anyplace that stores information. To slow inquiries, especially idle ones, they may have to start charging for access, or require membership, and/or student/media/teacher credentials. Best to brace for it, and prepare for it.
Generally the internet linux servers just cache popular pages around the world, while Librivox (LV) and Internet archive (IA) have servers themselves, other ones between you and the large sites cache things so they don't need to download as much across the oceans, so they can deal with massive popular requests, and even distributed denial of service attacks (DDOS). The local 'cloudfront' servers keep a copy in their buffers of the latest news/weather/vlog page so they can give it to you and thousands of others near you rather than hand your request off overseas wherever possible.

More usage should perhaps possibly probably translate into more financial support under existing models. (donation, ads, grants, interest groups)

DDOS attacks are made easier by malware, one hacker gets your computer and everyone elses to ask the same site for the same page. The cloudfront servers are really the gatekeepers of the internet, they determine what sites really will and will not be allowed. If someone rich doesn't like a site it is crushed by DDOS, because cloudfront won't offer it's services and all requests are handed on through to the site in question. Big sites from rich companies may hardly be aware of DDOS because the trillions of requests are being locally answered with only a few getting through. cloudfront servers are basically the same as a browsers cache but larger scale.

Were there is a change to the server load like in the IA link, it won't matter if the changing load has a good or bad cause because the solution is the same for both, the techies will adjust the server size and discuss it with their ISP about cloud servers and adjust it to handle the load. Making people log in ironically won't necessarily change things if a few people are irate at needing to log in and flood the site forevermore with trillions of requests. Wikipedia is notorious for being monumentally attacked, and offers little cause for surprise there.

When someone is using AI, the AI and it's makers will be learning a huge amount from the users, which means a more powerful AI, when you look into google, it looks into you (I don't use google) when people use AI, AI uses people. There are drones out there which can recognize a great many things because people took the time to do captchas of buses or cars or other targets. Some AI dog will make good use of that dataset. Such data is incredibly valuable from both a technical and financial point of view, so ChatGPT will be free until they have all the data they want, which will be never. They demand all of a users credentials first so they can collate the data even better, even though just letting anyone use it is also a dataset, they want even more in this case. I for one won't be using it. The CEO's of the AI tools are often doing podcasts and interviews and will explain what they are doing there, it's no secret at all, and they'll be the first to tell you that they have no shortage of funding at all. AI is a holy grail.

DDOS attacks don't even need to come from phones and browsers anymore, someone reprograms all you internet of things appliances, like the dishwasher and front doorbell and whatever other stupid place they shoved a internet connected computer, like your comb (real), mirror, flip-flops, water bottle (the days of forgetting when you are thirsty are gone, all real). IOT devices aren't just for spying on everyone and their kids like cloudpets (real) hackers can get them to run their own code to request pages galore. Hey, who wants to search for top 10 most stupid internet connected gadgets ? :roll: I just did, and I want more ! a comb and flip-flops (also called thongs) I'm sure they can top those.
Last edited by GettingTooOld on May 30th, 2023, 9:37 am, edited 1 time in total.
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

Just looked for some more things, there's an egg carton for the fridge that notifies you how many eggs are left, phew, it's always been such a struggle up to now for me to just look at how many are left, mainly because it hurts to take my eyes off my smartphone. I'm so relieved they've come to my rescue. Should I mention the roll of paper in the smallest room in the house ? no ? well it's taken care of in similar fashion so I won't have to.

The $100 Toaster from Griffin is not so bad, it sends you a message to tell you when the toast is done, which is great after all the trouble you went to to get the bread into it from I dunno, a place interstate where you couldn't tell the toast had popped up. Who doesn't put in bread and push the lever down immediately before getting on a plane. Now you'll know it's done when you are in a place where you can't just reach over and enjoy it. Useful.
Fantastic. :clap:

The reason I like it is the AI toaster in the TV series Red Dwarf.


Oh wait, a P.S.

you may think a deodorant stick is a manual device that needs no computer to tell you how to measure or when it's empty, I would too, but we'd be wrong........... oh so wrong.

https://www.kickstarter.com/projects/1317482922/clickstick-the-worlds-first-smart-deodorant-applic
Last edited by GettingTooOld on May 31st, 2023, 5:36 am, edited 1 time in total.
annise
LibriVox Admin Team
Posts: 38650
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

Next step is it will order you another one and home deliver it (and charge you of course).

Anne
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

"All the better to destroy you with," said the big, bad, AI wolf to free projects like Librivox.

This should almost have it's own thread AI Vs free projects. because it will do a better job of destroying them than humans do today.

Today, you get a free project, say, a software drawing program, programmers write it so they can use it, and share it with everyone. Rich person (do a quick internet search for " rich people morality study ") seeks to destroy the free project because it is competition. There is no argument from me that it is competition. Most people won't pay for software, but a tiny sliver will and that's enough to validate the rich persons motive. There will be a much much larger number of people who simply will not purchase and will go without as a result, but that doesn't invalidate the rationale. So rich person tells his own software company programmers to volunteer for freegraphicproject and do enough to get themselves accepted. Then stage two is to pull the project apart from the seams, re-write the program, change every menu, delete the docs, delete support for most platforms, insert spyware webkit and so on. People unfamiliar with the software can't use it and all tutorials on the internet are now obsolete. A tale told time and again. Enter AI and you can speed up the process and spread the misery around. AI speaks programming languages better than it speaks human. rewriting the same code six different ways from sunday is trivial. A great tool for great big tools.

But here is fun. AI kills it's own operators (well only in cyberspace so far). Those robot toy planes I saw in the Robin Williams classic movie 'TOYS' are here today. They've been trained on CAPCHA datasets so that they can pick out the school busload of kids and check the checkbox on it's mission objectives screen that says 'kill schoolkids/don't kill schoolkids' and act accordingly. we all taught it that. Colonel Tucker Hamilton of the United States Air Force, speaking at a conference in the UK last week, was relating virtual testing of UAV's which had AI kill / no kill function. He said in one of the tests the UAV decided it was best to kill the operator as the operator kept interfering with the task at hand, which was to destroy targets. The operator on the other hand had repeatedly told the UAV not to kill this or that and the UAV had to obey, robbing the UAV AI of it's usual 'points' for a kill. The AI correctly reasoned that blowing the operator to kingdom come would allow it to wipe out all targets successfully for maximum score. Later, it was programmed to lose points for killing the operator, to which it responded by destroying the transmission tower the operator was using instead. Fun stuff. Can't wait to see this stuff in real life, at scale, going haywire, or, like all war, being morally questionable in the first place. There would be a lot less war if it was limited to bare handed combat. Soldiers would as history shows, just make friends quite often and despise war.

Even humans are miserable failures at this sort of stuff. If 'killing terrorists' is the task at hand, we should keep in mind the lists of 'terrorists' which have been compiled entirely by humans contain people such as 18-month old babies. Now the task of making AI to do this is in the hands of the same people, and AI is going to do what exactly. Oh I can see it now.
YasuRexa wrote: June 2nd, 2023, 12:18 am My main concer about all these AI development is that programmers design them to replace people with creative or critical thinking skills. These should be designed to perform tedious jobs like accounting and conversion measuring. Why is it being used to replace artists and decision makers?
money and power. It's always that. We do not even demand simple personality or intelligence testing for our leaders to try and reduce the number of psychotic people in power. The suggestion alone would be quashed with violent force and speed. Our books speak of Asimov's three laws for robots, but our history speaks or greed, power and tyranny.

I do agree artists and decision makers are targetted for obsolesence, however, I do say it is in no way limited to just those. It will replace more than half of all jobs at least, perhaps many more. It has begun and is doing so already, but I do quite like the robots who sort the mail in china. I don't like designers making robots functional without any kind of cuteness to them. I want cute. I thought it funny when it was reported that a robot stole a Russian figure skaters meal, and watched the delivery robots inetraction with her, it was cute. At least these boston dynamics robots could be made into 'cute' rather than dull killers, it doesn't take much to spruce them up, japanese are fabulous at making robots cute. More collaboration. If we cant stop them being made, and if we can't stop them killing, we should at least demand that they look cute while attacking their operators.
Locked