An example I saw myself on Spotify was a song originally by Little Feat. The artist was listed only as “Feat” and both the sound and the presented album art were a little off. I’m pretty sure it was an AI ripoff.
I’d like to toss in an analogy with the evolution of music and instruments:
Analog instruments, analog cameras: everything is manual, only the sounds the instruments themselves can produce are present, except when post processed in a darkroom or a recording studio.
Electric instruments, digital cameras: different sounds and different image characteristics are easy to come by, but musicians and photographers still play/operate their respective instruments.
Sampling, AI: musicians sample the music of others to lay under their songs, AI trains on the photos of others to generate artificial images. In neither case is it necessary for the artist to play/operate an instrument.
While most of us have probably enjoyed some songs based on samples, I personally have more respect and appreciation for songs played by real musicians on real instruments. While it’s early days for AI, I similarly don’t expect to be wowed by AI-generated images.
As a life long hip-hop fan, this is a real slap in the face to sampling as an artform. Ouch. Sure there are some big fat ugly samples out there, but there are so many where you’d never know the original unless you were told.
I couple of years ago I was trying to understand (and, to a certain extent, reproduce) the guitar effects on U2’s Achtung Baby album, only to learn that a lot of what sounds like distorted electric guitar is actually a Korg A3 workstation keyboard. Go figure.
Sorry for painting with a big fat brush in a hurried post. I meant no offense to hip-hop, and there are some hip-hop artists that I really enjoy (I won’t sidetrack the thread by getting into which ones).
For sure, sampling is an art form, and done well results in some excellent music. It’s a different art form than playing an instrument, but it’s still done artistically by humans. AI is getting into territory where “done artistically by humans” ceases to apply.
Here’s an Adam Gibbs video from a couple of months ago, where he is removing unwanted elements from a photo. He uses an AI function to get rid of a large item; it generates an assortment of proposed replacements for the area occupied by the item, and he just has to choose which one he likes best. AFAIK, he was just demonstrating the feature and doesn’t use it routinely.
For sure I will, after snow shoveling (we finally got winter in late March!).
Younger users have less complex social lives on average relative to their parents and grandparents
I wonder when was the last time Doctorow talked to a teenager.
In any case, what he calls “dark corners” (platforms managed by communities)
precede “social media” by decades. Some of you here may have used BBS or Usenet, and may still be using IRC or some of those arcane web forums coded up in PHP with webdesigns from the late paleolithic era. Discourse is just a mature version (and it is great), but not conceptually different.
are not “dark” at all in most cases, but open to all well-meaning participants.
Yes, a whole generation that grew up with “social media” is rediscovering these platforms, just like people who grew up with mobile phones discover these strange tools called “cameras” that allow you to take better photos. But neither went away at any point, and had a continuous user base that is thriving.
I believe it doesn’t apply since pixls.us is still indexed by search engines and can easily be read by everyone even without an account or invitation. We are as open or even more open than facebook for that matter.
Younger in this case doesn’t mean only teenager, it also means young adults and it’s a pretty well observed phenomenon that younger people have far less social lives and friends than their parents or grandparents did. Not only due to social media but also the slow eradication of all institutions and communities that slowly got replaced by the internet and eventually shut down.
I am not sure it that this is the consensus by any means. This detailed time-use based study suggests that after you account for covariates, the age trend vanishes (and, interestingly, also that interaction shifts towards family as people get older).
I am not aware of any attempt to quantify the complexity of social interactions with data. But I am skeptical of Doctorow’s claims because it rhymes with the usual diatribes of civilization declining (each generation is worse off/more decadent than their elders etc, and the person making the claim usually happens to be an elder ).
But social interactions happen on the internet, too.
There is also a trend to limit the outside experience of kids by the parents. Up to the 80ies or so nearly nobody was driven to sports practice, dance lessons or to school. They walked or biked and met others on the way. Today there is a traffic jam in front of schools and the highest risk for a kid at a school in Germany is to be run over by another kids parents car. Permanent supervision is an additional problem, being always geolocated by your parent and having mom on speed dial can’t be very healthy for the development.
But I am not that pessimistic, it is an adaption to new times.
I know of groups of kids doing homework together all at their own homes on video call, motivating each other and having fun in between.
I know of girls dropping their phone in their friends bag and going of with their forbidden boy friend. Picking the phone up later at the friends home where they have of course been all the time and only had forgotten to enable the ringer again.
And there is a thriving market for the third phone, the second (perhaps dead) one is for giving up to the teacher in class.
Groups of friends are in permanent conversation from waking up to well past bed time. It would be interesting to see if these friendships outlast school better, not much of my contacts from school survived going to university in another city and my parents moving also.
Does anybody here remember Yahoo! in its original form as a curated list of web sites? Perhaps we’ll go into that direction again and build something like a federated web directory.
Yes, but much more finely structured and weighted. To use this existing list one would have to know and trust this community. So not very useful for the ones who don’t know us.
But if enough other sites and people trust us on our field of expertise one could trust us without knowing us. The original Page ranking algorithm was great in finding such structures and it made Google a great experience and filthy rich. But then came SEO…
I like the Fediverse and Mastodon, such a system could be implemented there. But there are already spammers putting up Mastodon instances and sysadmins defederating them as fast as they can. So the SEOs will find a way.