ChatGPT and AI - Maybe we and Second Life are Safe? (news)

[CaithLynnSayes]

Slightly off topic but i feel it might fit here. I often use a browser extension called Natural Reader. It recently got an update with AI voices and when i use it, even here on the forums, it reads so much more as an actual person would. it's gotten so good it sometimes tricks me into thinking i'm listening to an actual person.

I use this because i understand better when i'm listening instead of reading. Also English isn't my first language.

I just thought i'd mention this as it is somewhat in the spirit of this topic.

[Kuva Nishi]

On 3/31/2023 at 1:43 PM, animats said:

Vicuna: I'm sorry, but "hdi mk" is not a common chat abbreviation. Without more context or information, I cannot provide a meaning for this abbreviation. It is possible that it is specific to a particular group or community, or that it is a typo or misspelling. If you provide more information or context, I may be able to assist you better.

Might get better results with GPT-4, which has a bigger training set. I'm trying out Vicuna, which is a smaller model, because training it with SL-specific knowledge might be affordable.

Or better yet, we could just teach people to type properly rather than train a computer to decode laziness.

[Love Zhaoying]

26 minutes ago, CaithLynnSayes said:

Slightly off topic but i feel it might fit here. I often use a browser extension called Natural Reader. It recently got an update with AI voices and when i use it, even here on the forums, it reads so much more as an actual person would. it's gotten so good it sometimes tricks me into thinking i'm listening to an actual person.

I use this because i understand better when i'm listening instead of reading. Also English isn't my first language.

I just thought i'd mention this as it is somewhat in the spirit of this topic.

Now, if they could use better voices instead of the awful ones used for TikTok videos.

[CaithLynnSayes]

39 minutes ago, Love Zhaoying said:

Now, if they could use better voices instead of the awful ones used for TikTok videos.

The new Plus voices NR has are really good. I think you can preview them for free. I have a paid subscription, so i don't know.

[EliseAnne85]

On 3/29/2023 at 11:48 PM, Love Zhaoying said:

Heard on the news this morning.

I thought it was interesting that Steve Wozniak was highlighted in the story I heard.

I guess if these experts are right, and the AI development is "backed off" / slowed down a bit, our Second Life maybe be safe from the AI takeover, and RL jobs may be safe a little while longer. Science Fiction and debates (literal debates, high school / college, etc.) used to talk about this when I was younger, I'm just a little surprised it is happening now.

https://www.bloomberg.com/news/articles/2023-03-29/ai-leaders-urge-labs-to-stop-training-the-most-advanced-models

Elon Musk called for a slowdown in AI as well recently.  I listened to a little of his interview.  He said if humans can program AI to kill chimpanzees, that's what AI will do.  Hopefully, it never comes to this, of course.  Humans programming AI for barbarism.  

The US has talked recently in the last few years about restoring the upper and middle classes.  The upper middle classes could be restored via taking use of 0% loans our government was offering to anyone who wanted to start a "green energy business".  So, what is/are those "green energy" that people may have taken 0% interest loans, AI that removes jobs and jobs remove cars on the road and need for excessive consumerism = green energy?  It's scary to think "green energy" may be removing most jobs altogether.  

[Istelathis]

1 hour ago, EliseAnne85 said:

Elon Musk called for a slowdown in AI as well recently. 

Yeah, but then he turns around and discusses building his own AI.  Everything that comes out of that guy's mouth is questionable, he probably wants everyone else to slow down on AI so he can build his own, then catch up to what is available right now, and shut everyone else out of business as he surpasses them. 

 

 

Edited May 4, 2023 by Istelathis

[Bree Giffen]

I don't think anyone can afford to 'slow down' on AI because someone else is not going to slow down and that other person is going to be generations ahead of anyone who decided to slow down. Pretty sure Russian, China, Iran and anyone else who can afford it will continue to fully explore AI. It's almost like a nuclear arms race. 

[EliseAnne85]

1 hour ago, Istelathis said:

Yeah, but then he turns around and discusses building his own AI.  Everything that comes out of that guy's mouth is questionable, he probably wants everyone else to slow down on AI so he can build his own, then catch up to what is available right now, and shut everyone else out of business as he surpasses them. 

 

 

Oh, yeah, that's true.  Yes, he did.   I read Mr. Musk works at least 120 hours a week as a CEO of I don't know how many businesses.  He needs to retire from some things and find his focus on a few things.  He seems a bit manic at times, though I don't know if he has an actual chemical imbalance or is a workaholic.  But, you are right.  He did an about face after he said AI needs to slow down for 6 months.  

[Coffee Pancake]

5 hours ago, EliseAnne85 said:

He said if humans can program AI to kill chimpanzees, that's what AI will do.

Yes, save the killing of primates for humans, specifically the humans who work for Elon Musk

https://www.dazeddigital.com/science-tech/article/55461/1/elon-musk-neuralink-brain-chips-trails-killing-monkeys

If you want to see what murderous AI looks like, drive a Tesla on auto pilot near children.

5 hours ago, EliseAnne85 said:

Humans programming AI for barbarism.  

We don't program AI. That's not how AI learn.

AI learn from training data. If the training data is racist, then we get racist AI. If the training data is nothing but triangles, we get an AI thats good at triangles.

1 hour ago, EliseAnne85 said:

I read Mr. Musk works at least 120 hours a week as a CEO of I don't know how many businesses.

No he doesn't. he just doesn't have any friends, hobbies and deliberately wants nothing to do with his children.

3 hours ago, Istelathis said:

Yeah, but then he turns around and discusses building his own AI.  Everything that comes out of that guy's mouth is questionable, he probably wants everyone else to slow down on AI so he can build his own, then catch up to what is available right now, and shut everyone else out of business as he surpasses them. 

Everyone got the jump on him while he was off pumping crypto, and now he wants people to wait while his employees catch up with the new hotness.

Edited May 5, 2023 by Coffee Pancake

[Evah Baxton]

If quantum ever leaves the emulator, every encrypted connection will be broken, every crypto transaction will be cracked, and we won't even have the power to label AI as AI. It will rise above us faster than we can blink.

[Aiyumei]

27 minutes ago, Evah Baxton said:

If quantum ever leaves the emulator, every encrypted connection will be broken, every crypto transaction will be cracked, and we won't even have the power to label AI as AI. It will rise above us faster than we can blink.

What makes you think every encrypted connection or transaction can't be cracked at the moment? All you need is to have the corresponding encryption key or private api key and you have full access to the corresponding transaction or connection. In terms for connection it's even easier because all these VPNs out there on the market currently track every single thing you do on the web while using them and they can locate you at a moment's notice if needed. 

 

3 hours ago, Coffee Pancake said:

If you want to see what murderous AI looks like, drive a Tesla on auto pilot near children.

A while back there was outrage around Mercedes because their autopilot was instructed to run over pedestrians in order to save the driver should bad situation arise.

[Love Zhaoying]

3 hours ago, Coffee Pancake said:

AI learn from training data. If the training data is racist, then we get racist AI. If the training data is nothing but triangles, we get an AI thats good at triangles.

I hope they don't use History books to train AI (or certain works of "fiction")!

[Bent Luik]

On may 27th 2047 computers will take control over all electronic devices, including military drones.
Only ai controled avatars will be allowed on Second Life.
So enjoy the next 24 years.

 

 

Edited May 5, 2023 by Bent Luik

[Love Zhaoying]

2 hours ago, Bent Luik said:

On may 27th 2047 computers will take control over all electronic devices, including military drones.
Only ai controled avatars will be allowed on Second Life.
So enjoy the next 24 years.

 

 

"Bots - taking over Second Life since 2003. Our motto is, 'One of these days, we'll get AI!' " 

Your move, stoopid hoo-mons!

 

[Ceka Cianci]

2 hours ago, Bent Luik said:

On may 27th 2047 computers will take control over all electronic devices, including military drones.
Only ai controled avatars will be allowed on Second Life.
So enjoy the next 24 years.

 

 

People should really think about that, when there is talk about having electronic devices implanted into themselves in the future.. hehehe

[Love Zhaoying]

3 minutes ago, Ceka Cianci said:

People should really think about that, when there is talk about having electronic devices implanted into themselves in the future.. hehehe

..some people already "implant" devices into themselves!

Edited May 5, 2023 by Love Zhaoying
Meant, in a "naughty way".

[Ceka Cianci]

3 minutes ago, Love Zhaoying said:

..some people already "implant" devices into themselves!

They will be the first to be assimilated.

[Love Zhaoying]

16 minutes ago, Ceka Cianci said:

They will be the first to be assimilated.

One of the first Cyperpunk novels (I believe by William Gibson but I forget the title, it could have been in Neuromancer) had the protagonist sit on a "data interface" probe. 

Described in the story as quite scary and painful-looking. ETA: I assume it was supposed to be humorous, it was quite funny to me, at least.

Edited May 5, 2023 by Love Zhaoying

[EliseAnne85]

11 hours ago, Coffee Pancake said:

We don't program AI. That's not how AI learn.

AI learn from training data. If the training data is racist, then we get racist AI. If the training data is nothing but triangles, we get an AI thats good at triangles.

12 hours ago, EliseAnne85 said:

I thought he said "programmed", now I'm wondering what word he used.  It's been a helluva of a Spring where I live and we got an avalanche of pollen, so I've been on antihistamines that make me drowsy.  But, that's what I thought he said.

Just wanted to say I hope they train AI better so that *women* can open jars and bottles and stuff.  Some of these products being made are too difficult for women to open.  It's insane!  Some, I have to try and take a knife and cut them open.  Ridiculous, just ridiculous!  

Also, imagining trained pizza making.  So, every pizza will be trained to be exactly alike.  Makes food sound a little bland.  The passion and the finishing touch of making food will be gone.  The human eye has the ability to see when the food looks just right, and then it's done.

Edited May 5, 2023 by EliseAnne85

[Istelathis]

10 hours ago, Coffee Pancake said:

Everyone got the jump on him while he was off pumping crypto, and now he wants people to wait while his employees catch up with the new hotness.

I just wrote a few paragraphs detailing my feelings toward Elon, and decided to delete them.  The guy drives me kind of bonkers, which result in me leaving behind long rants about him 😋 But essentially yeah, this is most likely just another one of his predictable 4d chess moves that people for some reason, think are absolutely genius, where he gets to sprint forward of everyone else, at our literal expense, and claim his trophy.  

Regardless, despite Elon's faux concern over AI, there is no slowing it down, we can all host it on our own computers now.  Which is good, imho, this technology should not belong to only the select few and come at an outrageous expense for everyone.  Hopefully it stays this way.

Edited May 5, 2023 by Istelathis

[Anaryane]

14 hours ago, Istelathis said:

Ouais, mais ensuite il se retourne et discute de la construction de sa propre IA. Tout ce qui sort de la bouche de ce type est discutable, il veut probablement que tout le monde ralentisse l'IA afin de pouvoir construire la sienne, puis rattraper ce qui est disponible en ce moment et exclure tout le monde des affaires à mesure qu'il les dépasse. 

 

We are talking about a person who dismissed the developers on the size of their work (coding), and not on the quality of the code in question... In addition to having parasitized many companies that work for him, and to create visibly toxic climates. Technically, this guy doesn't know how to do anything other than put his marbles in what brings in money and use his image for marketing, even if it means talking bull*****. This guy would have the means to do much more altruistic and humanistic things, but no... In short, my opinion is very fixed on him. 

Edited May 5, 2023 by Anaryane

[Love Zhaoying]

20 minutes ago, Istelathis said:

another one of his predictable 4d chess moves

In Star Trek TOS, the Vulcans played the fancy 3-D chess.  I'd like to think the Romulans would be bad at it (being the violent, less smart cousins to Vulcans).  Musk is a Romulan.

[EliseAnne85]

On a different note, I tried the Sloyd AI 3D object generator last night.  The objects look like sculpty objects we've already had for years.  Too bad we didn't have it ten or so years ago.  

I have no need for sculpty objects.

But, it's super easy if anyone wants to use it.  You just click on the object you want and then you can change it if you want to.  

 

Edited May 5, 2023 by EliseAnne85

[Rowan Amore]

AI leaders warn the technology poses 'risk of extinction' like pandemics and nuclear war - ABC News (go.com)

[Rachel1206]

 

NVidea presented this at Computex 2023, Taipei, where the player can interact with NPC. Quite remarkable.