AI Safety

zapper · 2024-10-02 20:38:16

@golinux I now read what you said about stupidity being the driving force of humanity, but there is one other thing, egos/stubbornness also plays a huge part too..

Probably a larger part actually.

Last edited by zapper (2024-10-02 20:38:37)

golinux · 2024-10-02 21:09:35

@zapper . . . My definition of "stupid" is just more inclusive than yours . . .

Devarch · 2024-11-03 12:11:37

Some thoughts.

Oncle's AIs are against humanity. Only for control and fascism.
You can not avoid ai in this world.
If necessary, one can run ai models locally including uncensored.
The best way to run them: sandboxed/isolated, zero trust, recheck.
Local models are small and stupid comparing to oncle's models. But it's possible to use many of them, one per task.
Better to avoid to become addicted to ai.

Just IMHO.

trinidad · 2024-11-03 15:37:10

“Just as one spoils the stomach by overfeeding and thereby impairs the whole body, so can one overload and choke the mind by giving it too much nourishment. For the more one reads the fewer are the traces left of what one has read; the mind is like a tablet that has been written over and over. Hence it is impossible to reflect; and it is only by reflection that one can assimilate what one has read. If one reads straight ahead without pondering over it later, what has been read does not take root, but is for the most part lost.” a little Schopenhauer

AI will certainly eat itself because it can't naturally forget, and that will set human progress back hundreds of years in the end. There is no kindness in ambition. I wish I owned a coal mine.

TC

zapper · 2024-11-03 15:38:40

@golinux your definition of stupid includes narcissistic and arrogant?

golinux · 2024-11-03 16:17:00

@zapper . . . Of course.

zapper · 2024-11-06 02:34:42

@golinux alrighty got it. Yeah, stubbornness and arrogance is good to put in that category of stupid.

Since people who are arrogant usually know almost nothing.

I don't think I know as much as I used to, I will say that.

I am a dot in this universe. Why should I have a big head anymore...

It did so well for me in my first 35 years

(sarcasm)

Last edited by zapper (2024-11-06 02:35:19)

stargate-sg1-cheyenne-mtn · 2024-11-06 10:02:31

re: tiny dot

https://archive.org/details/the-tiny-dot

re: cosmos

Universe Size Comparison | Cosmic Eye (Original HD)

https://www.youtube.com/watch?v=8Are9dDbW24

fanderal · 2024-11-06 22:16:36

stargate-sg1-cheyenne-mtn wrote:

re: tiny dot

Thanks for the link. Overly simplistic yet an accurate depiction of why we're seeing AI's unchecked growth. Social networks and MSM infused with Bernays' methods of manufactured consent makes me wonder how much of what we see is noticed or understood by the general public.

Recent AI news

NO FAKES – A Dream for Lawyers, a Nightmare for Everyone Else
August 19, 2024
https://www.eff.org/deeplinks/2024/08/n … ryone-else

"Because it [NO FAKES Act of 2024] is a federal intellectual property right, Section 230 protections—a crucial liability shield for platforms and anyone else that hosts or shares user-generated content—will not apply. And that legal risk begins the moment a person gets a notice that the content is unlawful, even if they didn't create the replica and have no way to confirm whether or not it was authorized, or have any way to verify the claim. NO FAKES thereby creates a classic “hecklers’ veto”: anyone can use a specious accusation to get speech they don’t like taken down."

Billionaire Larry Ellison says a vast AI-fueled surveillance system can ensure 'citizens will be on their best behavior'
Sep 15, 2024
https://www.businessinsider.com/larry-e … 024-9?op=1

Microsoft, BlackRock form group to raise $100 billion for AI data centers
Sep 19, 2024
https://www.networkworld.com/article/35 … nters.html

Creeps Flocking To Telegram To Generate Nude Images And Videos With Easily Available AI Bots: Report
"Probe revealed 50+ nudify AI bots subscribed by over 4 million monthly users!"
Oct 17, 2024
https://in.mashable.com/tech/83883/cree … ots-report

Boy, 14, killed himself after AI chatbot he was in love with sent him eerie message
23 October 2024
https://www.dailymail.co.uk/news/articl … ssage.html

Devarch · 2024-11-10 20:40:42

Apocalypse Now

Claude AI to process secret government data through new Palantir deal

https://arstechnica.com/ai/2024/11/safe … -new-deal/

stargate-sg1-cheyenne-mtn · 2024-11-10 23:02:27

reminded of the television series "Person of Interest"

https://en.wikipedia.org/wiki/Person_of_Interest_(TV_series)

yeti · 2024-11-19 21:34:50

This may not be a safety topic for us, but someone else may see that differently:

FORTUNE
Elon Musk’s AI turns on him, labels him ‘one of the most significant spreaders of misinformation on X’
https://fortune.com/2024/11/14/grok-mus … -spreader/https://fortune.com/2024/11/14/grok-mus … -spreader/

;-D

greenjeans · 2024-11-19 23:07:49

How do you define "intelligence"? For surely that has more bearing on this topic than anything else.

My own father was literally a rocket scientist who helped put the first men on the moon, a mathematical prodigy of epic proportion, but in reality he was something of a savant, he couldn't figure out how to change the oil on the car, nor even use the simplest of handtools to build things that I could already do at 10 years old.

I know a guy that absolutely nobody would think was intelligent, but he can solve mechanical/fabrication problems in epic fashion, give him a new problem to solve that there isn't even an established method of solving yet, and in a few minutes he will have the simplest, most elegant, fastest and most secure solution there is. And it's just effortless for him, I learned a great deal working with him, mainly that I myself, as so many "smart" people are, was guilty constantly of waaaaay overthinking things.

blackhole · 2024-11-20 13:55:49

yeti wrote:

Levelled?
Like "only a good one with an AI can stop a bad one abusing AI"?
I do not buy that!

JWM-Kit has a valid point and argues from a realist perspective - ideology and moralising won't stop the development of GAI by "Big Tech". As with splitting the atom, you can't put this "AI" thing back in the box and just forget about it and/or ban it. You cannot go and smash the printing presses as the Luddites did and even if you did, the presses will be remade (as they were).

It will be used and is being used (MS Copilot), in the next phase of the ever expanding global surveillance/spying/profiling/telemetry infrastructure which permeates today's web and computing platforms. Without free and open alternatives, this "AI" will be squarely in the hands of "Big Tech" only, who will use that as leverage to more firmly cement the monopolies it has been building throughout the web and within "open source" for the last 15 to 20 years.

It is also being used in defense, "Project Maven" for example. The problem is already far larger than most realise...

Last edited by blackhole (2024-11-20 13:56:24)

stargate-sg1-cheyenne-mtn · 2024-11-20 15:42:30

tick, tick, tick...

https://en.wikipedia.org/wiki/Doomsday_Clock

https://thebulletin.org/doomsday-clock/current-time/

yeti · 2024-11-22 23:27:49

Robert Miles AI Safety
Using Dangerous AI, But Safely?
https://www.youtube.com/watch?v=0pgEMWy70Qk

yeti · 2024-12-06 20:26:13

Annoying counts as danger ...

Fran Blanche
YouTube Now Has A Brain - And It Is Insane!
https://www.youtube.com/watch?v=DteWHExa04I

delgado · 2024-12-07 20:43:04

ChatGPT Just Solved Chess

https://www.youtube.com/watch?v=rSCNW1OCk_M

yeti · 2024-12-21 20:08:56

Cool Worlds
Will A.I. Replace Scientists?
https://www.youtube.com/watch?v=vnl9Xf3wwU0

pcalvert · 2024-12-21 21:47:19

Here's a good example of a threat posed by AI that is not theoretical:

CoPilot+ PCs, Windows Recall, Keylogging Privacy Invasions!
https://rumble.com/v4zmoxu-copilot-pcs-windows-recall-keylogging-privacy-invasions.html

Last edited by pcalvert (2024-12-22 15:36:32)

stargate-sg1-cheyenne-mtn · 2024-12-22 07:51:03

@pcalvert, see this thread:

https://dev1galaxy.org/viewtopic.php?id=6484

and specifically this post by ralph.ronnquist:

https://dev1galaxy.org/viewtopic.php?pid=48861#p48861

for more information on the _full_link_displayed_preferred_procedure_

it is really quite simple once you get used to it
(also, fortunately the thread is found via search easily since there are not many topic texts containing "hyperbola")

pcalvert · 2024-12-22 15:38:35

@stargate-sg1-cheyenne-mtn

Thanks for the tip.

The officially official Devuan Forum!

#51 2024-10-02 20:38:16

Re: AI Safety

#52 2024-10-02 21:09:35

Re: AI Safety

#53 2024-11-03 12:11:37

Re: AI Safety

#54 2024-11-03 15:37:10

Re: AI Safety

#55 2024-11-03 15:38:40

Re: AI Safety

#56 2024-11-03 16:17:00

Re: AI Safety

#57 2024-11-06 02:34:42

Re: AI Safety

#58 2024-11-06 10:02:31

Re: AI Safety

#59 2024-11-06 22:16:36

Re: AI Safety

#60 2024-11-10 20:40:42

Re: AI Safety

#61 2024-11-10 23:02:27

Re: AI Safety

#62 2024-11-19 21:34:50

Re: AI Safety

#63 2024-11-19 23:07:49

Re: AI Safety

#64 2024-11-20 13:55:49

Re: AI Safety

#65 2024-11-20 15:42:30

Re: AI Safety

#66 2024-11-22 23:27:49

Re: AI Safety

#67 2024-12-06 20:26:13

Re: AI Safety

#68 2024-12-07 20:43:04

Re: AI Safety

#69 2024-12-21 20:08:56

Re: AI Safety

#70 2024-12-21 21:47:19

Re: AI Safety

#71 2024-12-22 07:51:03

Re: AI Safety

#72 2024-12-22 15:38:35

Re: AI Safety

Board footer