AI Versus "Soul" A Study in Forum Tolerance For the Mundane

[TJMcG125 8]

I got my own router/modem and have saved a lot of rental money that I was paying for the one Comcast provided me. Does Verizon allow that?

I never asked, but the monthly charge for the 5g+ service and my bundled unlimited data phone is less per month than I previously paid for just Starlink. And it's faster and has fewer interruptions, until this major failure. I access the tower that is close to us. Can another router do that? They told me yesterday that we are very close to getting fiber in the area, but we are in a more remote area, so I don't expect to have that available to us for a long time.

 

[smoke665]

@AlanKlein dont know about Verizon, but T-Mobile supplies the router. The only charge is for the WIFI mesh to cover a wider area. I ended up taking it back, didn't work, didn't need it anyway.

 

[snowbear128 59]

We got an extender from T-Mo when we got the phones, a number of years ago. The WiFi router has always been provided by the service (VZ, Comcast, now Fidium).

My other router is a Bosch fixed base with the plunge cut adapter.

 

[TJMcG125 8]

We got an extender from T-Mo when we got the phones, a number of years ago. The WiFi router has always been provided by the service (VZ, Comcast, now Fidium).

My other router is a Bosch fixed base with the plunge cut adapter.

Verizon gave me a free extender, but we don't use WiFi, the house has data ports and it's all ethernet.Their router has two ports, where Starlink had only one, so I already had a switch. I can turn WiFi on and off in the phone app.

 

[smoke665]

@TJMcG our whole house is wired with Cat 5, tv, & phone lines to a termination panel in a downstairs hallway closet. Unfortunately the best signal for the T-mo is in an upstairs window across the room from a port. I could reroute a cable, but the wifi has worked so well that I havent bothered with it. Being untethered for computers, printers, & scanners is liberating

 

[v1photos]

On another site there is a long, ongoing discussion about the devils and angels of AI in photography with at least two posters claiming that "AI has no soul." I for one think nothing has a soul, but just for fun, I took a picture of mine, a throwaway from a trip to Mesa Arch in Utah, and tried my mightiest to describe it to Nano Banana, to see how close it would come. I am not a professional "prompt engineer" (yes, I saw that term used), but I did my best. Here is what I said: "
"I am looking at Mesa Arch in Canyonlands National Park I am on the east side curve of the arch, looking at the west side. The sun is bright, and the underside of the arch is glowing orange. Over on the West side, a man is standing, apparently holding a camera, and his head is tipped down slightly as if viewing the camera. He is wearing a gray hoodie jacket with the hood over his head and blue jeans. There is a tiny bit of a pale pink shirt showing at the bottom of the jacket. Through the arch, you see just a glimpse of the canyon below."

I know these AI discussions can get boring, but humor me... Here are the two pictures. Which of these has "soul"?

View attachment 293809View attachment 293810

Soul isn’t in the pixels or the prompt, it’s in the intention behind the image and the meaning the viewer brings to it.

 

[Rickbb77 16]

AI images aren't created; they are compiled from images that we humans put up on the interweb. When you prompt an AI to make an image it simply searches the database for the most likely components and assembles them.

Just a composite, a collage if you will, of images that match the programming, (your prompt).

Does it have soul, not any more than random snapshots taken with a point and shoot.

But the image you had made, the guy's face looks off to me, light is all weird on it.

 

[smoke665]

AI images aren't created; they are compiled from images that we humans put up on the interweb. When you prompt an AI to make an image it simply searches the database for the most likely components and assembles them.

Not really, there's alot of misunderstanding as to what AI really is. 8n practice images are used as training to compile Diffusion models which have been trained on billions of image-text pairs to understand the relationships between words, shapes, colors, and styles. The process involves converting a text prompt into a numerical format, then reversing a "noise" process to create a new, coherent image from scratch. The following video explains the process.

 

[Rickbb77 16]

But they have just added a layer of programming to take the database images they are "trained" on to apply these filtering process to make it appear to be a new image.

In my mind it's akin to pasting up a collage and then squinting at it to "blur" it enough to "see" something different that is not really there. Yes, it's much more complex than just pasting up what is already there, they do have to "noise" filter it up to prevent it from being a straight up copy. But I'd argue that it's not "creating" anything new, new looking yes, but not new from scratch.

 

[smoke665]

But I'd argue that it's not "creating" anything new, new looking yes, but not new from scratch.

No, watch the video. All the traing does is provide a mathematical relationship between text and image that AI uses to construct new images from text.