Mastodon Feed: Posts

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

One last AI/ML thing from me, and then I'll put this topic down for a while.

Businesses are training large models on the work of artists without their permission so that their work can be automatically remixed at scale to provide a service.

This is very different from artists learning, sharing, and inspiring each other through the generations.

One is an ongoing heirloom project. The other is a short-term grift that threatens to undermine the very thing it is attempting to extract value from.

pzmyers@octodon.social ("pzmyers 🦑") wrote:

Seven times one is seven, therefore God. Numerological arguments are so stupid.

https://freethoughtblogs.com/pharyngula/2022/12/15/do-you-want-100000-or-is-it-200000/

Attachments:

• Numerological "proof" of god (remote)

pzmyers@octodon.social ("pzmyers 🦑") wrote:

An exemplary teacher makes Fox News angry.

https://freethoughtblogs.com/pharyngula/2022/12/15/fox-news-is-mad-at-a-minnesota-teacher/

Attachments:

• Mandi Jung, St Paul science teacher (remote)

Reblogged by cstanhope@social.coop ("Your friendly 'net denizen"):
axwax@fosstodon.org ("AxWax") wrote:

Reblogged by jeffsonstein ("Jeff Sonstein"):
augieray ("Augie Ray") wrote:

Reblogged by jeffsonstein ("Jeff Sonstein"):
jhilden@vis.social ("Jonatan Hildén") wrote:

Reblogged by jeffsonstein ("Jeff Sonstein"):
jakecoppinger@aus.social ("Jake Coppinger") wrote:

Attachments:

• Screenshot of OpenDroneMap making measurements of cycleway and road width (remote)

Reblogged by jeffsonstein ("Jeff Sonstein"):
TheGuardian_us@newsrelay.org ("The Guardian US") wrote:

Attachments:

• hello (remote)

pzmyers@octodon.social ("pzmyers 🦑") wrote:

It's too true, what a Midwestern winter is like.

https://freethoughtblogs.com/pharyngula/2022/12/15/another-snowstorm/

Attachments:

• Snowing in the dark (remote)

fribbledom ("muesli") wrote:

If good Python code looks pythonic, then good Go code should certainly look gothic.

(via @xssfox)

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

(I get it. I'm not ignorant enough of history to not understand exactly how much some parts of humanity don't value other parts at any given time... It's still depressing.)

Reblogged by blithe ("Blithe"):
rchusid@med-mastodon.com ("Ron Chusid :verified:") wrote:

Reblogged by cstanhope@social.coop ("Your friendly 'net denizen"):
selzero@vmst.io ("Deniz Opal") wrote:

Attachments:

• 281a5362ccabb524.png (remote)
• 08faa57117a5d5d1.png (remote)
• 47b47710811036fc.png (remote)
• adb28620963d484e.png (remote)

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

Part of what I find depressing about some AI/ML discourse is that people seem to be honestly debating things like, "does it matter if a machine made this 'art' and not a human?"

Yes! It matters! People matter! My god, what is wrong with us? Humanity matters! Those machines, whose output converted to colored pixels we are dreamily staring at, won't get very far without us. Stop spending so much time and effort valuing and nurturing machines.

Value and nurture people. Value each other. Please...

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

Anyway, the nice thing about this AI image generation is that if it succeeds in killing off people's motivation to get into art, it will ultimately dry up the well it used to get started in the first place. It will be a shame for humanity, but as long as the machines are happy...

I hope artists of all stripes take steps to protect themselves from their work being used to train automated remixing machines.

As for me, I'll be checking the labels to make sure a human made the work.

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

I admit I am biased. I value humans over machines. I don't care what a machine "thinks". And I don't value the prompt maker any more than I value the "idea man" who tries to take all the credit. I'm always more interested in the people that actually get the job done.

For AI generated art, the people that actually got the job done were the artists whose works were used to train the machines.

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

I'm also unconvinced with the camera analogy. It's not like photographers went out and started photographing all the existing artists' works and passed it off as a service. Camera's were a new technology and did displace illustrators (not just portraitists) over time, but they were still operated by humans and required real effort to go off and compose and take new pictures.

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

I disagree with "The AI model is no different in function" and "the morality of it is actually quite similar" to human artists.

Automated remixing at scale is quite a bit different from humans learning from each other over time. Passing on skills and knowledge from generation to generation.

The idea that because no single human can view all the art currently being generated doesn't make a convincing argument as to why AI generated art isn't a threat to artists' livelihoods.

cstanhope@social.coop ("Your friendly 'net denizen") wrote:

I saw this posted elsewhere, and instead of getting grumpy in their mentions, I'll just get grumpy on my own thread:

https://opguides.info/posts/aiartpanic/

The overview of how things work seems fine.

I would like a citation for this: "the act of training a neural network on datasets like this has already been decided to be legal", as I don't recall a fair use case on this (maybe I missed it). Any fair use case in the US is going to look at the four points here:

https://www.copyright.gov/fair-use/index.html

Reblogged by xor@tech.intersects.art ("Parker Higgins"):
oldroadside@botsin.space ("old roadside pictures") wrote:

Attachments:

• ac64c65744746bdd.jpg (remote)

Reblogged by xor@tech.intersects.art ("Parker Higgins"):
pomological@botsin.space ("fruit toots") wrote:

Attachments:

• ced037de7ad339e2.jpg (remote)

Reblogged by isagalaev ("Ivan Sagalaev :flag_wbw:"):
akallen404@infosec.exchange ("404 Name Not Found") wrote:

xor@tech.intersects.art ("Parker Higgins") wrote:

New itsy-bitsy release of xword-dl mostly adds support for The Modern Crossword, a cool new daily from Puzzle Society. "check it out" https://github.com/thisisparker/xword-dl

Reblogged by kornel ("Kornel"):
lolennui@mstdn.social ("Amy") wrote:

Reblogged by jeffsonstein ("Jeff Sonstein"):
ajhickford@mastodonapp.uk ("Adrian H") wrote:

Attachments:

• Two black cats stare intently at the camera, probably expecting a tasty treat (remote)

Reblogged by jeffsonstein ("Jeff Sonstein"):
stux@mstdn.social ("sтυx🎄") wrote:

Reblogged by lloydmeta ("Lloyd"):
SusanPotter ("Susan Potter 🧑🏼‍🎄😱🌲") wrote:

Reblogged by technomancy@icosahedron.website ("tech? no! man, see..."):
noracodes@weirder.earth ("nora tindall is 🏳️‍⚧️") wrote:

xor@tech.intersects.art ("Parker Higgins") wrote:

if you're really a fan of television as an art form? if you want to show respect for the work of the creatives, from the script writers to the actors to the grips to the electricians? you've gotta torrent it

xor@tech.intersects.art ("Parker Higgins") wrote:

HBO Max *removing* streaming shows from their archives is just reaching out to everybody who ever torrented their material to tell them they were right, actually