The anti-AI thread

[Kaido]

So one of the things to realize is that the communication method embeds itself in the final artform, otherwise known as:

"The Medium is the Message"

What does "the medium is the message" mean?

“The medium is the message” was coined by media theorist Marshall McLuhan in his book, Understanding Media. It’s also the name of the first chapter of that book. I’ve written a summary of Understanding Media, and include it in my list of best media studies books, but here I explain what it...

kadavy.net

Which is why things like anime are so hard to translate into say, a live-action movie & retain the same effect:

https://www.imdb.com/title/tt1098327/

NVIDIA has introduced DLSS 5, which adds an AI overlay to existing games to improve the quality. Brilliant idea...in theory! The problem is:

1. The Medium is the Message
2. Games are created by Artists
3. Artists fine-tune their art for the constraints of the technology & their skill at the time

The issue is that once people are familiar with something, you can't take away from that. The tech demos for DLSS largely change the character's visual identities so much that it's largely considered "AI slop". The general take has been:

"Generative uncanny valley"

So now you've got an AI's interpretation of the art, which essentially tries to guess at the art direction to try to make it better, which is different than say ray-tracing, upscaling, new game engines, higher FPS, etc.

https://www.tiktok.com/video/7618347796789251358

"It's Re-Rendering the Game!" – It Turns Out Game Artists Don't Love DLSS 5, Despite Nvidia's Claims

www.ign.com

And now for the memes:










haha

https://x.com/k1rallik/status/2034282630921715949

 

[jpiniero]

MSN

Bezos is looking to raise up to 100 billion to buy manufacturing companies and AI them.

Yeah Bezos. It won't end well.

 

[Steltek]

MSN

Bezos is looking to raise up to 100 billion to buy manufacturing companies and AI them.

Yeah Bezos. It won't end well.

Well, Skynet WILL need a manufacturing arm, won't it?

 

[Kaido]

MSN

Bezos is looking to raise up to 100 billion to buy manufacturing companies and AI them.

Yeah Bezos. It won't end well.

Half my clients are in manufacturing & half are in food. I started getting more into robotics last year. All of my robotics contractors use AI. But it's all 100% to build specific, repeatable scenarios for machine work, vision systems, etc. It has to:

1. Solve a particular problem
2. Be reliable repeatable every single time

Like, self-driving is complex, but you're going forrward/stopping/going backwards & turning With stuff like food robotics, there are sooo many issues:

* Sanitary operation
* Food safety
* Rust
* Cannot ever go hayware around humans, especially with knives
* Can operate with wet surfaces
* Can hold Jello
* etc.

AI chatbots at drive-thrus are getting better, but there's also only a finite number of available options to choose from on the menu. As far as general-purpose robots go...I dunno! I mean, I never thought I'd see the level of voice, research, and image technology that I've seen today! But I definitely feel like robotics will be a MUCH harder problem to solve AND make a profit from!

 

[mikeymikec]

Derp:

PwC will say goodbye to staff who aren't convinced about AI

: Professional services giant did not read its own report on lackluster benefits

www.theregister.com

 

[lxskllr]

Derp:

PwC will say goodbye to staff who aren't convinced about AI

: Professional services giant did not read its own report on lackluster benefits

www.theregister.com

Let's ask around and see what other businesses think of ai, and see how they're doing.
The results show that ai is largely a waste of time.
We're going all in on ai!

Make it make sense...

 

[mikeymikec]

Let's ask around and see what other businesses think of ai, and see how they're doing.
The results show that ai is largely a waste of time.
We're going all in on ai!

Make it make sense...

To be fair though, Slashdot has a poll going which is basically split 50/50 between positive and negative experiences:

Slashdot Poll | By the end of 2026, how ...

(though it would be sadly funny if the ~50% of voters were AI bots)

Logically therefore AI can't be completely useless and therefore if applied correctly it can be useful.

Personally I think it's inevitable that it will eventually live up to the hype, but my concern is at what cost to humanity. Aside from widespread job losses and unemployment levels (which would be serious enough to produce an entire strain of problems aside from AI-specific issues), I don't think there's any powerful people in this world that I would trust to maintain/improve AI responsibly, and I think as a result we're going to find the fastest and derpiest route to a Horizon Zero Dawn or Skynet type scenario because we have some real fucking tools making the decisions.

I think if any alien species are watching us, they're taking bets on how we're going to fuck our civilisation into oblivion inside of ~50 years (at least, the element that will play the biggest role, e.g. end-stage capitalism, right-wingers gone ballistic, climate change, AI).