Will the complexity pendulum swing back?

[Agamon]

Hey, AI slop is a lazy way to dismiss what AI can already do, much less where it's headed in the next few years.

Sorry, I'll use more 25 cent words to describe AI detritus next time. lol

 

[Morrus]

Hey, AI slop is a lazy way to dismiss what AI can already do, much less where it's headed in the next few years.

Incorrect. AI slop is an excellent way to describe the massive onslaught of low quality plagiarised crap filling YouTube, Kindle, search results, and, unfortunately, RPG books. It is slop in every sense of the word.

Maybe AI will one day end up with ethical and good quality uses. Right now, it’s slop. Plagiarised slop.

 

[kermit4karate]

Sorry, I'll use more 25 cent words to describe AI detritus next time. lol

Thank you. Might I suggest "AI frippery" or "AI toilet water"?

 

[Snarf Zagyg]

Incorrect. AI slop is an excellent way to describe the massive onslaught of low quality plagiarised crap filling YouTube, Kindle, search results, and, unfortunately, RPG books. It is slop in every sense of the word.

Maybe AI will one day end up with ethical and good quality uses. Right now, it’s slop. Plagiarised slop.

Speaking of ethical and good-quality uses, we are now dealing with a bunch of AI-generated invoices being sent out from a municipality's official email account after it was hacked.

Truly, a brave new scammy world.

 

[kermit4karate]

Incorrect. AI slop is an excellent way to describe the massive onslaught of low quality plagiarised crap filling YouTube, Kindle, search results, and, unfortunately, RPG books. It is slop in every sense of the word.

Maybe AI will one day end up with ethical and good quality uses. Right now, it’s slop. Plagiarised slop.

I appreciate the touchy nature of the subject, but I very much disagree. It hasn't been that for months, if not at least a year. It's advancing well beyond the slop stage and is already such an accurate simulation of a real person that I don't believe most people can even tell when they're interacting with an AI.

Ha! How many people are wondering right now, "Hmm..is this frog guy an actual human??"

 

[kermit4karate]

Speaking of ethical and good-quality uses, we are now dealing with a bunch of AI-generated invoices being sent out from a municipality's official email account after it was hacked.

Truly, a brave new scammy world.

For sure. The threats from AI are growing exponentially each day. Anyone who runs a website knows firsthand how much their WAFs, CDNs and load balancers are getting hammered by AI bots now.

Plus, we're seeing new forms of malware emerge that are so dynamic and "intelligent" that they may soon be impossible to detect and stop, except by other AI-enhanced tools.

Miracles and nightmares.... Coming soon.

 

[Agamon]

For sure. The threats from AI are growing exponentially each day. Anyone who runs a website knows firsthand how much their WAFs, CDNs and load balancers are getting hammered by AI bots now.

Plus, we're seeing new forms of malware emerge that are so dynamic and "intelligent" that they may soon be impossible to detect and stop, except by other AI-enhanced tools.

Miracles and nightmares.... Coming soon.

Eh, LLMs still are just sophisticated regurgitation machines. They have no logic whatsoever, to the point that an Atari 2600 with 2kb memory can destroy it in a chess match.

 

[Morrus]

I appreciate the touchy nature of the subject, but I very much disagree. It hasn't been that for months, if not at least a year.

Incorrect. Have a look at YouTube or the Amazon Kindle section sometime. It’s all slop. Fake trailer slop. Fake look-a-like book slop. Crappy incorrect search result summaries. This is now, not a year ago.

 

[kermit4karate]

Eh, LLMs still are just sophisticated regurgitation machines. They have no logic whatsoever, to the point that an Atari 2600 with 2kb memory can destroy it in a chess match.

That's a gross oversimplification. They do regurgitate info, but LLMs are regurgitating based on many billions of data points.

Their predictive ability isn't like a video game where an NPC pulls from 100 different possible responses to an inquiry. They're pulling the most apt next word in their response from a billion similar inquiries.

At what point is the dataset of possible responses so large that it effectively simulates a human's free choice?

Answer: that point is neeeeear.

Is it still a simulation? Yes. Is the simulation becoming so good that it might as well be considered a pretty good facsimile of a human? It's getting there, like it or not.

 

[Micah Sweet]

No.

Have you ever seen the 1E DMG?

There’s no RPG book in existence that even comes close to it in crunch and complexity.

The 1e DMG is literally one of my top five favorite RPG books.