Roger E. Avedon

Karma: 3

Roger E. Avedon 12 Apr 2026 22:15 UTC
1 point
2
in reply to: Richard_Kennaway’s comment on: The Ethics of AI-Assisted Creative Work
I’m sure we’ll agree it’s a complex social process. Gemini says there are many examples of photographers that were dismissed as doing technical or even chemical experiments during the early period who are now revered as great artists. Gustave Le Gray used different negatives of sea and sky to create compositions with balanced exposure in the 1850s.

This is a very beautiful piece of art:

https://static.the-independent.com/s3fs-public/thumbnails/image/2016/02/19/01/5927595.jpg

Roger E. Avedon 12 Apr 2026 22:04 UTC
1 point
0
in reply to: Richard_Kennaway’s comment on: The Ethics of AI-Assisted Creative Work
My intention with Ms. Smith was to distinguish disclosure from non-disclosure, where non-disclosure exploits reasonable expectation that any creative output is the work of a human, and disclosure no longer exploits that expectation, qua Dennett.

To be clear, I’m not advocating for any kind of acceptance of prompt-only content creation. I don’t think I’m advocating for any particular acceptance of AI-assistance in general, only for a possible framework for how to think about the ethics that attach to it.

Roger E. Avedon 12 Apr 2026 21:53 UTC
1 point
0
in reply to: Richard_Kennaway’s comment on: The Ethics of AI-Assisted Creative Work
Just did a little quick research (using Gemini!) and although photography was immediately embraced by the public, it took around 50 years for the fine arts establishment to begin accepting it for curation, awards, etc. The related anecdote here is that the Paris Salon — the world’s greatest art exhibit across the 18th and 19th Centuries — included photography in 1859, but segregated it from fine arts as part of industrial and mechanical crafts in a completely separate building. The first Guggenheim Fellowship for a photographer isn’t until 1937.

Given the 2026 rules for the Oscars, it would be somewhat ironic if the last bastion of complete rejection of AI-assistance in content creation was SF writing.

Roger E. Avedon 12 Apr 2026 21:10 UTC
1 point
0
in reply to: Richard_Kennaway’s comment on: The Ethics of AI-Assisted Creative Work
Could you expand on your comment? Do you mean that whether AI-assistance is exploitive even when disclosed depends critically upon the detail in the disclosure? So, for your extension of the example, if Ms. Smith is offering her purely prompt-based robopaintings for sale that would be exploitive (independent of offered price?) but not if she provided a detailed explanation of what she did versus the robot? Of are you saying that, per your broader view of AI, Ms. Smith is exploiting herself?

I think I tried to address both quite legitimate concerns in the essay, so I’m eager to hear where you think I’m confused or mistaken. Thanks for the comment, regardless.

Roger E. Avedon 12 Apr 2026 5:07 UTC
1 point
0
on: The Ethics of AI-Assisted Creative Work
Worth flagging for readers: Audrey Henson’s reporting at The Drey Dossier on the Shy Girl case, which landed after this essay was substantially complete — “91 Percent Human” (March 22) and “The Shy Girl AI Scandal Is Way Worse Than You Think” (April 2). The editor’s note at top addresses it briefly.

The Ethics of AI-Assisted Creative Work

Roger E. Avedon12 Apr 2026 4:09 UTC

−1 points

11 comments17 min readLW link

Roger E. Avedon 12 Sep 2022 2:35 UTC
1 point
0
on: Fermi Paradox: Iron Age Milky Way
The Industrial Revolution seems largely the result of a particular positive feedback singularity, that the steam engine, which converts iron, coal and water into mechanical energy that can be used to pump water, was concocted in England, where high quality veins of coal and iron ore happened to be located deep underground beneath the prevailing water table. The steam engine pumping out mines, with steadily increasing power and efficiency, facilitated access to more iron ore and coal, which could be more efficiently transported by steam locomotives riding iron rails and by steamships, to foundries fired by coal to make more iron used to make steam engines, trains, rails and steamships. And off we went to the stars.
Once you get steam engines, you get mechanical engineering as everyone races to figure out how to do even more with less. Feedback control theory comes directly from the steam engine too (James Clerk Maxwell analyzing Watt’s Flyball Governor). The alien planetary autocrat surely wants better steam tech to out-monument the last guy. So steam plus self-interest probably gets you space travel independent of democracy or free markets.
One can’t prove a negative, of course, but on an iron-nickel planet in the habitable zone with carbon-based life using water as a solvent, steam engines do seem inevitable once you get large scale adoption of iron tools. There’s an interesting paper or three to be written about Drake’s f_c for the mechanics and thermodynamics of alternate steam engine triples (metal, fuel, fluid) potentially available on classes of exoplanet by formation and composition.

Roger E. Avedon 27 Jun 2022 22:40 UTC
1 point
0
in reply to: Mo Putera’s comment on: A Quick Ontology of Agreement
Excellent and thank-you! I’d somehow forgotten about Nerst and would have linked to his work directly. I think the additional value Hahn’s ontology brings to erisology is an explicitly positive gradient, in a hill-climbing sense. For any disagreement, Hahn’s ontology allows the parties to accept some level of agreement (Where are we on the agreement landscape?) and have an objective target for improvement, assuming good faith on everyone’s part. I’m inclined to try to communicate it to Nert based upon your linking the two!

A Quick Ontology of Agreement

Roger E. Avedon26 Jun 2022 17:39 UTC

5 points

2 comments2 min readLW link

Roger E. Avedon 13 Jun 2022 16:12 UTC
2 points
0
on: [Book review] Gödel, Escher, Bach: an in-depth explainer
Why is there no Section IV? It’s a delicious irony of reading a book review of Gödel, Escher, Bach on LessWrong that upon noticing that the Latin-numbered sections jump from 3 to 5 (III, V) I’m left wondering if that’s simply a silly mistake or if the omission of section IV is some profound meta-self-referential in-joke on the part of the reviewer that I’m just not clever enough to see.

Roger E. Avedon

The Ethics of AI-As­sisted Creative Work

A Quick On­tol­ogy of Agreement

The Ethics of AI-Assisted Creative Work

A Quick Ontology of Agreement