131–140 of 155 posts
Earlier quoted context omitted.
A DALL-E generated IKEA style manual for a nonexistent piece of furniture would be absolutely hilarious.
Introducing the Fiibean wepa Taki: https://labs.openai.com/s/95bli4C4CKZ73DMigFNdDKQ4
The image shown is nothing has some concepts in common with Stålenhag, but it has a lot more depth. I can't find any Stålenhag which evokes the same way.
In Stålenhag's work, or what of it is easily accessible online, there are some recurring gimmicks, which are absent in the DALL-E work, or present in a disguised form.
DALL-E doesn't juxtapose any mint condition 1970's Volvos or Saabs against a background showing a menacing superstructure from the distant future. It doesn't invoke any cheesy mechanical monsters or characters inspired by popular Sci-Fi. For instance, "
Stålenhag lays out everything plainly. In about a millisecond, you see what you're supposed to see: the easily recognizable elements like ordinary people or retro objects, against fantasy objects, set in some nice lighting: atmospheric effects like haze, sunset or whatever that completes the mood.
In the DALL-E there are some old objects, but they are not instantly seen, and they are in old condition. For instance what looks like the wreck of an airliner appears in the rubble in the foreground, but you might not see it right away. That makes the scene plausible. It is more of a spooky glimpse at something that could actually someday be, somewhere.
It has the DALL-E weird elements in it. Like why do the cables on what looks like a telephone pole seem to connect to elements in the distant background; DALL-E seems to have perpetrated an https://en.wikipedia.org/wiki/Impossible_object there, but it's not wham-in-your-face; not everyone will spot it right away.
There is what looks like a channel of water through the debris, but what looks like a reflection in it is so weirdly wrong that you can't make up your mind whether it's a weird local reflection or a glimpse through an aperture in the bottom of a floating sky structure, where a distant reflection is shining up from some ocean below.
Stålenhag's work doesn't have any of these sorts of demented optical labyrinths in store for the viewer. He's like a Norman Rockwel*, targeting the nostalgia of that viewer who grew up on a diet of 70's, 80's and 90's anime and sci-fi, and is longing for the 20th century.
--
* E.g. check out the cheesy fromage called "The Long Way Home". Rockwell plus Storm Troopers. This just screams "Hey, you look transparently like a Generation X geek; would you be interested in buying a limited edition print of me?"
Earlier quoted context omitted.
>ML revolutionized translation a long time ago and the demand and pay for translators went down. This is incorrect. Translators employed: 2005 - 29240 2021 - 52170 Yearly salary: 2005 - 38300 2021 - 52170 Source - BLS ( https://www.bls.gov/oes/current/oes_nat.htm ) Employment growth was ~79% on a US population growth of ~13.5%. 38300 2005 $ equal 52854 2021$. So pretty much demand for translators exploded while real…
Translation is not exclusively a salaried job. Most translators in the fields I'm familiar with (games and manga) are freelance contractors, and most are getting paid by the word/page, not a yearly salary. ML has depressed the rates those people earn. It's been good for the demand for editors though since they need to pay someone to fix the DeepL output up so they can release it.
Earlier quoted context omitted.
This is how I feel as well. I'm a software engineer by trade, but I'm really much more interested in writing music and stories. I always imagined myself retiring the day I have saved up enough money so I could focus on art and creativity. My life certainly has felt much less meaningful since I saw what DALL-E 2 can produce. And my main concern is not that of professional artists' financial situations. I'm mainly worr…
Computer-generated content may pose an essentially insurmountable challenge to human practitioners of most art forms, but not all of them. The forms of art that are at risk of being overtaken by ML are those in which individual works of art can be faithfully represented in some storage format (which may be anything from a video file to a piece of paper with things printed onto it), which is used to transmit the work…
Earlier quoted context omitted.
Are you copying one or more of those works? If so, yes. If you aren't, then no. How do you know? Well as an artist, after thousands of hours of copying, and thousands of hours of drawing from imagination, you know which activity you are doing. As an untrained person, you might not know, but the quality/value is so low no one cares.
But drawing from imagination is just copying from an abstraction existing in your brain. And that abstraction comes from inputs from the outside that you saw in the past, where would it come from if not that? In the Obama example it comes from copyrighted inputs. Without the photos in the magazines and newspapers I would not have idea how Obama looks. Any draw I do, it's going to be derivative of those copyrighted ph…
Publishers license text to print in books. Music producers license samples to use in songs. Artists license photos and textures to use in illustrations and scenes. _License the source material used in training sets_. _Then_ go out and industrialize art, everyone wins!
Earlier quoted context omitted.
I highly doubt that’s even theoretically possible, given the vast range of variety and quality of music created by humans.
Feel like people would have said the same with images 6 months ago tbh. (I mean sure, you could get a fake photograph of a person, but the text control that clip has unlocked is just wild…)
You could maybe make the argument for distinguishing between a human and a machine playing a piece of music.
But if I write a song that just goes A, D, C, G on a loop (plenty of songs like that), how would you determine that a machine song that goes A, D, G, C is not human written?
Earlier quoted context omitted.
Translation is not exclusively a salaried job. Most translators in the fields I'm familiar with (games and manga) are freelance contractors, and most are getting paid by the word/page, not a yearly salary. ML has depressed the rates those people earn. It's been good for the demand for editors though since they need to pay someone to fix the DeepL output up so they can release it.
Unfortunately for that line of reasoning, the hourly rate according to the BLS has also been pretty much constant in real $.
Earlier quoted context omitted.
Unfortunately for that line of reasoning, the hourly rate according to the BLS has also been pretty much constant in real $.
https://www.taipeitimes.com/News/feat/archives/2022/08/09/20...
Earlier quoted context omitted.
Former professional translator here (Japanese-to-English, 1986 to 2005). I have very mixed feelings about MT. On the one hand, as you and others point out, rates for human translation have been hurt, and the quality of a lot of MT-enabled translation visible to the public is mediocre to poor. On the other hand, MT is enabling communication among people that was not possible or practical before. A small start-up is ab…
Good points! As an example, as a hobby, I am interested in things like anime video compression and such. Two of the most essential sites for me in terms of resources and information have been almost entirely Chinese. And while they appreciate and respect the international users, at the end of the day, they can’t translate each and every single forum post into English or any other language that someone may need. And a…
What it comes down to is that under capitalism, a lot of art like movies, tv, video games, certain more polished professional tools, etc, requires an artifical concept of ownership to ensure that people are able to spend time creating them, because choosing to do too much uncompensatable labor is effectively choosing to starve to death.
But maybe, like me, you see the possibility of a future where we can find a way to compensate art that isn't locked into the grim realities of capitalism. Well, cool, but then political strategy has to be considered. As is, ML seems to be a technology that funnels money away from compensating creators and towards large corporations that can invest in ML and their shareholders. And in our practical reality, more money buys more speech, more speech tends to lead to more power. Thus, self interest rules the day, with the wealthiest uninterested in changing the fundamental economics systems and working to stop it.
That's why I am concerned about ML and copyright. It would be no problem in an ideal world, but in our world, it makes the status quo worse in a way that prevents progress towards that ideal.
Earlier quoted context omitted.
Well, it probably didn't memorize any single of them. One thing that has changed dramatically from back when I studied AI (a whole 5ish years ago, sigh...), to my shock, is that the number of training epochs has dropped dramatically. Sometimes down to just one. If it's only ever looked at your image and updated its weights once , could that really be enough to copy it? Only if virtually everything in your image is al…
> political press photographers aren't hired for originality! Yes they almost definitely are, what else would they be hired for? You wouldn't watch the same news everyday (even if it feels that way) > the number of training epochs has dropped dramatically So? Lets imagine a documented nurodiverse person (lets use a photographic memory) "snapsohts" a famous "works" and then sits in a room, after 5 years training, to p…
Craftmansship. They have subtle, but probably very good, intuitions about lighting, facial expressions etc. and also a lot of concrete knowledge about these things. All of which an algorithm can pick up on by examining thousands of images once each.
Once or many matters because if you just examine each data point one, there can be no overfitting in the traditional sense. And that's what shocks me about the "one epoch is all you need" fruits.