250 AI-Generated Pages from the 1987 Radio Shack Catalog
November 30, 2022 11:33 PM   Subscribe

This post was deleted for the following reason: Poster's Request -- loup



 
Oh no, help.
posted by loquacious at 11:39 PM on November 30, 2022 [1 favorite]


I’m increasingly of the opinion that the best use for this kind of generative art is for creating 70s pulp sci fi novel covers. Which I mean as a compliment. Truly we live in a golden age.
posted by threecheesetrees at 12:40 AM on December 1, 2022 [3 favorites]


There is almost no design language I know better than the Lotus 123-era Radio Shack ads, and these are not bad!
posted by rhizome at 12:50 AM on December 1, 2022 [2 favorites]


I don't know what that blue thing in the top right picture is, but I kinda want one
posted by Jacen at 12:51 AM on December 1, 2022


The images are cute, but I wonder why no one has figured out how to scrape out the text-area hieroglyphics and replace them with appropriate AI-generated text.
posted by Umami Dearest at 2:23 AM on December 1, 2022 [2 favorites]


For Jacen.
posted by JHarris at 3:08 AM on December 1, 2022 [3 favorites]


Clearly the next step is the Heathkit catalog, but is humanity ready?
posted by mittens at 3:23 AM on December 1, 2022 [3 favorites]


I remember being a tween/young teen back in the early-mid '80s and someone telling me about this thing called Lotus 123. They said it was "a Computer Program that had a manual that was THIS thick!" as they held their fingers about two inches apart. As a part owner with my brothers of a Commodore 64 with a cassette tape drive, I was astonished.

I loved gaping at catalogs like these. Wish I had gotten deeper into computer sciences. After a 20 year career as an art director in advertising, I'm considered a "computer wiz" by many of my family and friends, even though my wizardry amounts to many nights trying to figure out WTF was going wrong so I could finally finish my work and go home. That and my ability to use Google and follow directions.
posted by SoberHighland at 4:25 AM on December 1, 2022 [3 favorites]


I’m increasingly of the opinion that the best use for this kind of generative art is for creating 70s pulp sci fi novel covers.

I was really tempted to post this the other day: Johnny Darrell uses AI Midjourney to generate stills from an alternate version of Disney's TRON created by Alejandro Jodorowsky. I mean - WOW.

Theft.

And then I reconsidered.
posted by JoeZydeco at 5:03 AM on December 1, 2022 [5 favorites]


I've had dreams like this.

Was there ever a My First Sony catalogue? Because that could be fun, too.
posted by RonButNotStupid at 5:44 AM on December 1, 2022


The text is exactly what I see when I try to read things in my dreams.
posted by achrise at 6:37 AM on December 1, 2022 [10 favorites]


This takes place well after the era celebrated in the FPP, but it's set during the holiday season and it's by MeFi favorite Jon Bois, so here's Jon's mini-memoir of working at Radio Shack when it was well within its death spiral. (Previously on the blue; Bois on the blue.)
posted by Halloween Jack at 6:42 AM on December 1, 2022 [1 favorite]


I find myself a little unsettled by how easy it is to make things. I really want to read the article about how they make these. [written by GPT-NeoX]
posted by The Half Language Plant at 6:50 AM on December 1, 2022


AI-generated text is one of my favorite things.
posted by slogger at 7:28 AM on December 1, 2022


The tech reminds me of the New Flesh from David Cronenberg’s Videodrome.
posted by jonp72 at 9:11 AM on December 1, 2022


I realized we're casually posting links to Masto threads now (whoo-hoo!) so has anybody figured out how to properly archive threads if, say, the media storage on an instance gets full and the admin starts to purge?
posted by credulous at 10:15 AM on December 1, 2022 [2 favorites]


When we're really in trouble is when AI can adequately mimic commenters

I think we're there.
posted by The Half Language Plant at 10:16 AM on December 1, 2022


All A.I. Generated Art is the product of Art Theft.

I'm going to disagree with this because I've been seeing this go around, I've been thinking about it a lot and it's not sitting well with me.

For starters, we're talking about vintage Radio Shack ads, here, and if there's any source material that should be considered public domain by default and ok to use as training sources, it's probably vintage and long defunct advertising.being used as a parody like this.

The designers who did that work were paid for it a long time ago, the products aren't even being sold any more, and the company itself doesn't even meaningfully exist. No one is functionally getting ripped off here.

And as far as commercial arts and design work goes in general, well, this has always heavily relied on stock libraries, morgue files stocked with a wide variety of sources, and it wasn't at all uncommon to repurpose and edit elements into new works with or without licensing or copyright permissions, or using stuff that is aged out of copyright entirely.

I've seen dozens and dozens of modern logos being used by companies large and small that are straight ripoffs of previously existing logos, often from companies long defunct or from other countries or marketplaces.

As for the case of using source materials and whole artistic styles from living, working artists, especially to use it to compete with, say, character design or aesthetics for commercial projects? Yeah, this is problematic.

But it's also running right up against the whole concept of creative fair use, rework and manipulation, and if we push on that wall too hard and take a hard line stance that all AI art is theft, we might actually break important and valuable parts of copyright laws as well as concepts of freedom in creative and artistic expression, such as parody.

I'm not sure if there is even really a philosophical, metaphysical or meaningful legal difference between an AI image generation tool or an artist using a morgue file or the entire internet as source material to use for inspiration to create new works.

Artists have been using sources like this for a long, long time in a lot of different ways.

Note that I'm not rooting for or defending these AI projects or whatever large corporations that may be funding them, but I do strongly feel that it's not going to work out in the favor of artists getting paid if we change the definition of fair use to stick it to these companies.


There's also the interesting metaphysical/philosophical discussion to be had about the creative process of writing good prompts. It seems like it might actually open up a whole new creative skill set of being able to be some kind of an AI whisperer or oracle and producing consistently good results by knowing how to put the right words together like some kind of incantation.

I'm personally really fascinated by this and I think I wouldn't hate doing it as a job because I like words and art. I've messed around with some of the free image generators and it's absolutely fascinating because you can get dramatically different results by messing around with descriptive word choices, orders and grammar.

I'm pretty sure some form of this job is going to exist in the future where it's a creative skill that can be taught and practiced.
posted by loquacious at 10:52 AM on December 1, 2022 [3 favorites]


The text is exactly what I see when I try to read things in my dreams.

Weirdly I don't think I have ever tried to read anything in my dreams. If I have I don't remember, and I have a lot of very vivid dreams that I do clearly remember, and I'm pretty adept at lucid dreaming.

I will note that some of these AI images have seem to end up in my dreams or general waking thoughts because so many of them are unsettling, hallucinatory and even deeply weird and/or wrong.

The way some of the free image generators have been obscuring things like faces to prevent or limit faked images of real people is often bizarre nightmare fuel.
posted by loquacious at 11:01 AM on December 1, 2022 [1 favorite]


When we're really in trouble is when AI can adequately mimic commenters

Depends on the commenter. YouTube comments are a good place to play “Bot or Human No Smarter Than A Bot?”
posted by ricochet biscuit at 12:41 PM on December 1, 2022


When we're really in trouble is when AI can adequately mimic commenters

I mean, it can win at Diplomacy, so I mean ... now we just need an AI that repairs fractured friendships and marriages for the rest of the players.
posted by credulous at 1:10 PM on December 1, 2022


Mod note: Hi all, deleted a few comments that are contributing to an unnecessary derail. Please take a look at our community guidelines and consider if you are participating in accordance with those guidelines.
posted by travelingthyme (staff) at 7:45 PM on December 1, 2022


Make capitalism great again! 🥳
posted by beesbees at 10:08 PM on December 1, 2022


1. I am not a Dude
2. Although a potentially 'interesting' use of A.I Generative images, it doesn't sidestep the fact that tools like Stable Diffusion are created and maintained via Art Theft of others work on a Massive Industrial Scale
3. The Uncanny Valley quality of all these Generative images seems (currently) to be a Feature not a Bug.
posted by Faintdreams at 1:00 AM on December 2, 2022 [1 favorite]


I certainly did not mean to misgender you. I was using "dude" in a snarky "person on the internet" way, which I regret, and will not use in the future on Metafilter. My sincere apologies.

I stand by my statement that merely repeating the words "AI art is art theft" in every AI image generation thread is spammy and not a useful way to contribute to a discussion.

I think this comment by tychotesla from a previous thread does a good job of laying out the complex feelings many artists have about the issues with AI art. loquacious' comment above as well.
posted by gwint at 10:05 AM on December 2, 2022


« Older *waves* new Masto thread   |   Bookmarklet to fix ads on Twitter Newer »


This thread has been archived and is closed to new comments