Ryan CordellFeb 12, 2023

I think @jeffjarvis takes the right lessons from media history into the ChatGPT/LLM debates in his response to Ted Chiang

Looking back at historical technological shifts, we see folks worried about potential harms—of, say, the printing press. But our lesson isn’t "look how silly they were!" because in every case they were partly right—the press did contribute to more misinformation, even violence

https://medium.com/whither-news/journalism-is-lossy-compression-86380f0bdb50

Journalism is Lossy Compression - Whither news? - Medium

There has been much praise in human chat — Twitter — about Ted Chiang’s New Yorker piece on machine chat — ChatGPT. Because New Yorker; because Ted Chiang. He makes a clever comparison between lossy…

Whither news?
Show threadRyan CordellFeb 12, 2023
but as @jeffjarvis also shows, those same histories show it takes time & experimentation to understand what a new technology will actually change & what guardrails we need to create around it—there are zero cases where we fully understood a new medium & could manage it from the outset, because a medium’s eventual shape is completely opaque in the moment of its creation—a primary reason to study these histories is to learn to evaluate our own shifts *slightly* more efficiently
Show threadRyan Cordell
our historical lenses are curved in such a way that we look back at "printing" as a singular event co-terminus with the positive developments we associate with the technology, but @jeffjarvis notes there were at 100 years of significant turmoil as people figured all that out—that all gets collapsed in the rear-view—I appreciate holding both the dominant effects of print (stability, fixity) in tandem with the messy realities that inevitably complicate that narrative
Feb 12, 2023 at 5:40pmWeb
Show threadRyan CordellFeb 12, 2023
I’m also sitting with @afamiglietti79’s post on this same topic today, because I tend to agree that the real dangers of ChatGPT-integrated-search are likely not the big headline-grabbing mistakes, but the subtle mistakes & dark corners it might open up—& those real harms are something we can’t waive away—we have to be developing ways to advocate, fix, & even yes sometimes resist these effects https://afamiglietti.org/uncategorized/the-articulation-where-llms-could-do-harm/
Show threadRyan CordellFeb 13, 2023

As if on cue, this episode of the One Year podcast appears in my feed, describing American musicians’ concerted protest in 1942 against new recording technologies that were hurting the live music economy—why hire bands for your bar, for example, when you can just get a jukebox? Why hire live bands for the radio when you can spin records instead?

How does this relate to my thread? Well again, these musicians weren’t wrong—the new recording technologies were absolutely reshaping their music economy & having real, often devastating impacts on musicians—their protest wasn’t misguided or silly. And in retrospect this debate gets lost because, well, recorded music persisted & eventually the music economy adapted—& our historical lenses are curved as I was arguing above.

But it can be instructive to take this historical moment of uncertainty & rupture seriously—what might this history suggest about how we should think about/manage/teach/&c. AI art, writing, &c.?

https://slate.com/podcasts/one-year/s4/1942/e3/recording-ban-1942-james-c-petrillo-the-american-federation-of-musicians-and-the-creation-of-bebop

Show threadAndy FamigliettiFeb 13, 2023
@ryancordell Andrew Ross had a piece in American Quarterly back in 2006(!) "Technology and Below-the-line Labor" that was good. I had some not entirely finished thoughts a few weeks ago here: https://afamiglietti.org/uncategorized/were-going-tinker-with-the-contours-of-ip-when-we-need-to-do-automated-luxury-communism-again/
Show threadDerek BruffFeb 12, 2023
@ryancordell @afamiglietti79 That’s a great piece. Thanks for suggesting it.
Show threadloucoveyFeb 12, 2023
@ryancordell @jeffjarvis There is a simple computer science term for this phenomenon" GIGO (garbage in garbage out)
Show threadRyan CordellFeb 12, 2023
@Loucovey @jeffjarvis I see that term quite a lot in more recent versions of these debates, and it’s often a good descriptor of what’s happening. I do get frustrated when that description of a system gets invoked as if it explains or defends bad systems that have actually been implemented & are having real-world effects on people.
Show threadMiaFeb 12, 2023
@ryancordell @jeffjarvis I wonder what folk in the future will call this 'incunabula' period of generative AI? Incunable AI, hmm
Show threadJeff JarvisFeb 12, 2023

@mia

I argue in my upcoming book that it's only 1480 in Gutenberg years. The net, the web, and AI are all incunabular.

cc: @ryancordell

Show threadRyan CordellFeb 12, 2023
@mia @jeffjarvis I tell my students every semester that we are in the incunable period of the computer age, at least as a widely-available medium. All we’re creating is experimental & much of it will look absurd to people in a few hundred years’ time. But some of these ideas will stick. Evaluating which is which may be impossible for folks stuck in this moment
Show threadFrancis IrvingFeb 13, 2023

@ryancordell @jeffjarvis yes this! Love the reference to inventing essays and novels.

I think it took more than 100 years. Was more like 400 years until the public library and universal literacy were invented, both vital to making the book and the newspaper reach their full potential of good for society.

We probably need to be quicker this time:
https://www.flourish.org/2015/08/making-our-information-society-safe-and-fair/

Making our information society safe and fair : Francis Irving