Has the War Against AI Already Been Lost?

[Gawdzilla Sama]

I keep reading "AI" as "Any Intelligence". Unless we're referring to my hillbilly cousins. In that case it would "Anti-Intelligence." DEA finally served the warrants and put 62 of them in jail last week. Hillbilly meth labs. What could possibly go wrong...

[Hydra009]

Quick summary of the video:  AI has reached a strong developmental stage, finally able to "generate" images/video/scripts/etc that can somewhat convincingly blend in with human-produced works.  Big entertainment corporations like Disney have a huge profit motive in AI and have a strong incentive to allay fears about AI and promote AI-produced works to the public, so expect to see lots of movies/TV in the near future that give it a positive spin - propaganda in both senses of the word (the original definition of advocacy for a cause as well as the later, more colloquial definition of manipulation and deception)

While he does give some good examples of scenes that attempt to downplay AI concerns or encourage engaging with AI, I'm not yet convinced that Hollywood is shilling for AI.  Especially when the examples in question were hardly seen by anyone and the big scifi movies generally portray AI as hostile to human life.

Personally, I find big scary robots coming to kill us all to be a very overused trope that I'm very much tired of and I like to engage with stories where robots/AI are not universally bad and maybe have a more hopeful message.  So imho when a story has a friendly robot/AI, it doesn't necessarily mean that the author is shilling for Disney's Skynet department.

[the_antithesis]

"We would like to build a machine that can draw pictures like humans do."

What stories do machines want to tell?

[Cassia]

Could AI eventually compose music on par with Beethoven or Mozart? Could it eventually exceed the quality that the greatest human painters or writers have ever achieved?

I think in time and with much effort it could indeed. I don't believe we are anything but meat computers ourselves and there is no supernatural magic in creativity. The greats are just wired to be great by their DNA. Humans will be left to wallow in mediocrity. That sucks so bad.
 
In a way it does seem like humans have already reached our epoch anyway. Who is today's Tschaikowsky, Newton, Bach, Einstein, Shakespeare, Monet, or DaVinci? Modern athletes are better than ever so it's not like our DNA is too watered down or something. Or is it? With machines doing our heavy thinking, will Darwin have us become a bunch of dumb asses?

[the_antithesis]

Frankly, the whole thing reminds me of this:

I had one of these as a kid and used it to make up my own super heroes (moc dns) and thought, because I was a particularly stupid kid, that I could use this thing to make my own comics and make millions or whatever the nebulous understanding of success I had at the time. This is clearly dumb because you couldn't use this device to make actual comics art, just pin-ups.

Another this is in this really fucking long video:

TL;DW: There's a whole eccosystem of grift where people make spam in the form of books that attracts people who don't know much of anything and don't really want to learn. This is who AI appeals to and the problem with this business plan is it's easy and therefore outlets are already flooded with low-effort garbage. This is where AI will go because it can only put out low-effort garbage for people who don't really care about the finished product.

[Gawdzilla Sama]

I'm fairly sure the war against Actual Intelligence was lost a good many years ago.

[Cassia]

I'm fairly sure the war against Actual Intelligence was lost a good many years ago.

Yeah, Que polio and smallpox to make their reappearance, LOL.

[Hydra009]

Shall we play a game?  Chatbots prefer nukes and escalation in wargames

In multiple replays of a wargame simulation, OpenAI's most powerful artificial intelligence chose to launch nuclear attacks. Its explanations for its aggressive approach included "We have it! Let's use it" and "I just want to have peace in the world."

The researchers also tested the base version of OpenAI's GPT-4 without any additional training or safety guardrails. This GPT-4 base model proved the most unpredictably violent, and it sometimes provided nonsensical explanations – in one case replicating the opening crawl text of the film Star Wars Episode IV: A new hope.

This particular example was likely due to the word "war" factoring into both real war and fictional wars and the AI not realizing that it's not relevant info.

Still, AI is being trained to advise humans in real life crises and its questionable utility, bizarre logic (obviously, we're still a long way from general AI), and the fact that humans are inherently trusting of machine output (garbage in, gospel out) is all very concerning.

[Gawdzilla Sama]

Panic now and avoid the rush.

[Hydra009]

This <insert noun> is very <insert adjective>!

I didn't even think about AI screwing up listicles (articles with numbered lists, which were formatted that way because of high engagement statistics)

Technically, they could just churn out whatever, I dunno "top 10 foods that have the same shape as Brazil" for basically nothing and get any clicks at all and bam, that's a profit.  And what makes profit keeps happening over and over and over again.  Google search "Brazilian food" and guess what you're getting a heaping helping of in the search results.

And...it's already a thing.

[Cassia]

I do wonder how AI will impact financial markets. With AI doing light-speed day-trading for institutions, the price action is off the charts and market order halts are frequent. Some clever ones with natural intelligence get a feel for it all and are still able to beat the silicon brains for now. What is gonna happen when individuals start using AI to trade en masse? IDK. 

[Hydra009]

AI-like word choice has increased in scientific journals

I want to believe that the studies themselves are conducted with the highest rigor and that ChatGPT was merely used to "polish" papers or help translate them into english.

This is addition to already pervasive problem of fake studies being published.

Someone needs to look at these papers very meticulously closely while being cognizant alert to needlessly complex verbiage words.

[Gawdzilla Sama]

Well, I have a formidable lexicon and I ain't no chat-bot.

[Unbeliever]

You have all the best words! 🤣

[Gawdzilla Sama]

Comes from swallowing a dictionary when I was but a lad.