Hi all, So two points, one I think Luke is right to push at Common Crawl, which was for me an impetus for the essay. The rhetoric is telling, but also its relationship to the public domain movement should be troubling. I think that does lead to a comment for Elena. I agree that alternatives exists, but not in as "free" per say or that free should be the objective. What is another objective that, already happening, but not named as such.
Thanks for reading! Happy to see a discussion. Be good, Fen On Fri, 28 Oct 2022 at 06:49, Luke Munn <luke.m...@gmail.com> wrote: > Thanks for sharing Fenwick, it's a very provocative essay. I've also been > thinking quite a bit recently about stable diffusion and what it > represents. I'll just aim at one thought here. > > The critique of typical villains like Google harvesting creative work for > corporate profit has been made by several recently and I won't be disputing > it here. > > More overlooked though is the assumed good of "free culture." Common > Crawl's website says: "Access to data is a good thing, right?" The chairman > of Common Crawl, Elbaz, is firmly of Silicon Valley lineage, preaching > about the virtues of the "open web" on the one hand and raising $25 million > in venture capital funding on the other. Hardly the hard distinction > between FOSS models and free-market capitalism that "free culture" can > signal. > > The sheen also comes off when we look at the actual technical production. > huggingface's stable diffusion model effectively crawls through Common > Crawl, harvesting billions of image-text pairs from its millions of > webpages. The aim is to make a repository of data "available to anyone." > But crucially, it is not their data to make free. > > CC have no rights over the stories, fan pages, emotional confessions, or > images in the millions of pages they scrape. In that sense, CC is not > really sharing so much as colonizing and commodifying. It brings to mind > the empire's framing of Australia (and other lands) as empty, "terra > nullius," effectively erasing the labor of those who had carefully nurtured > and tended to "country" for so many years before the "settlers" arrived. > > What we see then in Common Crawl is a completely impoverished > understanding of the commons. "Sharing" is narrowed until it become a > technical affordance: sharing means harvesting everything (even if it isn't > yours), hosting it on a server, and allowing others to download it without > paying a fee. Common Crawl never grasped what commoning was, and this is > unsurprising because commoning is so alien to the hyperindividualized > late-capitalism that Fenwick mentions. > > Taking up Fenwick's prompt to go beyond technology, then, might mean > putting platforms, web agents, and features aside for a moment, and > thinking socially. If I did this, I might ask who is in my community? Who > do I share things (stories, vulnerable moments, physical labor, financial > help) with? And how do they reciprocate, sharing their thoughts, ideas, > struggles, and resources with me? Even as I write this, I'm aware it sounds > sentimental. And yet this is how commoning - without a formal organisation, > without state funding, without a logo or a seriesA round - actually seems > to work in practice. As Elena highlights so aptly, these are not just > technical but social solutions, imperfect but enduring. > > nga mihi / best, > Luke > > > On Fri, 28 Oct 2022 at 04:49, Fenwick Mckelvey <mckelv...@gmail.com> > wrote: > >> Hi all, >> I have been a lurker on this list for so long, but I am trying something >> different for a change. I have been thinking a lot about Stable Diffusion >> and free culture that I wrote down something. Its a rough draft, but I >> can't help but think its deficits might lead to some constructive >> discussion about the very types of problem I think this list tried to >> address years ago. >> >> Well here goes, if nothing else an end to lurking for a change. >> There is no more free culture >> >> Common Crawl captures the spirit of free culture. In its launch in 2011, >> Lisa Green, director of the nonprofit 501(c)(3) organization, announced “It >> is crucial [in] our information-based society that Web crawl data be open >> and accessible to anyone who desires to utilize it” >> <https://readwrite.com/common_crawl_foundation_announces_5_billion_page_w/>. >> Carl Malamud, a 2022 Internet Archive Hero Award >> <https://blog.archive.org/2022/10/19/2022-internet-archive-hero-award-carl-malamud/>, >> serves on Common Crawl’s board further stressing Common Crawl’s roots in a >> free culture movement embracing the internet as an information commons and >> advocating for the protection of the public domain. >> >> Strange then that Common Crawl is behind our intense moment of financial >> speculation around artificial intelligence, the foundation of the >> highly-concentrated market of AI firms. Common Crawl provides the training >> data for the biggest names in AI today, Open AI’s GPT-3 and Stable >> Diffusion. Its crawls of the internet, its efforts to make the Internet >> available to anyone, have been embraced by an open source AI movement. >> >> I want to focus on this paradox between free culture and commercial AI as >> the later’s success marks the former’s diminish. If all that free culture >> does is feed a new round of venture capitalism, an aggressive automation of >> arts and culture, and a negation of digital life as anything other than >> data, then free culture is no more. Just another cheap data source. >> >> The decline has been a long time coming. Darin Barney, two decades ago, >> warned that the Internet could become a standing reserve of bits, another >> way of describing the mantra that data is the new oil. Barney’s warning was >> to others a site of political struggle, to resist what we once called the >> digital enclosures and to let the Internet become a third-way. Free culture >> and free software were movements that joined technical and political >> innovation. What has becoming striking after the commercialization of free >> software as open source and the depoliticization of piracy is that now >> Common Crawl marks the end of free culture as a viable tactic as well. >> >> >> *What is Open Source AI* >> >> Stable Diffusion and other “open source” AI movement restate the commons >> as a business practice. The roots of the movement might just as well be >> seen in the AI Commons movement >> <https://machineagencies.milieux.ca/ai-commons/>. The AI Commons is a >> movement, part of AI4Good, but one that marked a clear sense of how >> information commons and commercial AI could be completely complimentary. >> The AI Commons marked an important shift in commercial technology firms >> moving from private code to public code, from private data to public data >> that in turn moved these idea from critique to opportunity. >> >> That the commons could be a business model is at the core of the paradox >> marking the demise of free culture. Where once free culture marked a >> rebuttal of information feudalism >> <https://thenewpress.com/books/information-feudalism> or what Rebecca >> Giblin and Cory Doctorow call chokepoint capitalism >> <https://www.penguinrandomhouse.com/books/710957/chokepoint-capitalism-by-cory-doctorow-and-rebecca-giblin/>, >> what I find is that market valuations depend on free culture. From >> Microsoft’s new Autopilot for GitHub that turns GPL and other free licenses >> into training data to write new code to GPT-3 and Stable Diffusion that >> rely on Common Crawls to create its text and image generation tools. >> >> Openness and free culture have allowed and legitimated a new round of AI >> start-ups claiming to be a continuation of the open source movement. Emad >> Mostaque, the founder of Stability AI, explained how openness is behind the >> ethics of Stable Diffusion to the New York Times, “Who is responsible >> for AI? Who is responsible for the output of AI? And how can we make sure >> that it’s open and positive? But I don’t think that happens if the only >> debate that happens is behind closed doors and big technology companies.” >> <https://www.nytimes.com/2022/10/21/podcasts/generative-ai-is-here-who-should-control-it.html> >> By a strict reading of open source and the creative commons these companies >> seem a success story. >> >> Perhaps we might call this moment, optimistically, late-stage capitalist >> Internet as an attempt for scholar to periodize early political approaches >> to computing and to acknowledge a near total imbrication of technology and >> capitalism at work online. The term for me is a response at least to Chris >> Kelty’s reflections of the end of free software and open source as a mode >> of politics. I have always been haunted by this line, “There is no free >> software. And the problem it solved is yet with us” >> <http://peerproduction.net/issues/issue-3-free-software-epistemics/debate/there-is-no-free-software/>. >> Kelty’s conclusion remains an enduring provocation to reimagine, for me at >> least, to reimagine the early pirate politics and free/libre software >> movements undone by their internal politics but kinds of critiques we >> remain in need of. >> >> *After Free Culture? * >> >> The paradox then becomes productive because we need to define what >> specifically about open source AI negates free culture and what needs to be >> done? >> >> The immediate point, one well articulated in the The First Nations >> Principles of OCAP <https://fnigc.ca/ocap-training/>, is that public and >> open can be quite contradictory where openness negates the relations to the >> community and the public. How immediately might companies using the commons >> be held to account, not unlike how the GPL once expected code derived from >> the commons to return to the commons. >> >> There is another easy leap toward solidarity -- such as Google Walkouts >> over Project Maven -- that reflects on AI’s applications. Where the Google >> Walkouts demonstrated a refusal to work on military AI similiar movements >> need to question why AI continues to be what Solon Barocas calls a >> co-opting machine >> <https://www.publicbooks.org/machine-learning-is-a-co-opting-machine/>. >> How is it that the greatest advances in AI seemingly undermine creative >> labour, one of the last fronts in an overall de-skilling of labours >> following Astra Taylor’s precinct observations in the The People’s >> Platform <https://en.wikipedia.org/wiki/The_People%27s_Platform>. >> >> The commons itself is also increasingly contentious as an institution. >> Anna Tsing’s latent commons or Fred Moten and Stefano Harney’s undercommons >> question the very being of the commons as a stable form, or as a definite >> space clearly at odds with the Creative Commons movement responsible for >> Common Crawl. These critiques of the commons point at an advantage in >> ambiguity, though I worry that the uneven application of copyright and the >> law continues to advantage forms over others. Corporate piracy over peer >> sharing as one immediate example. >> >> These reactions, I think, help us to imagine what a public culture might >> be after free culture’s demise. What remains, however, is to recapture that >> productive engine of Kelty’s recursive publics that drive cryptocurrency >> and web3 today in building other hyper-capitalist futures. If commons-based >> peer production is a joke. If free culture is a ruse. What ways of doing >> with technology remain? Just as Jackie Wang has productively re-read >> theories of control through racial capitalism, I see a continued challenge >> of reimagining democracy, addressing its historic exclusions, and colonial >> underpinnings. >> >> Perhaps as a final word I stay with Cristian Dunbar-Hester’s closing >> sentence in Hacking Diversity. “A focus on technology itself... may be >> confining to a social justice agenda. It might be possible to build more >> democratic technology -- undoubtably, it is possible -- but at the same >> time, democratic praxis should never be limited to a technological >> imaginary” (p. 242). Therein, we might consider what was free culture as a >> proxy for new prototypes of being, of citizenship that never succeed but >> are still needed. >> >> >> # distributed via <nettime>: no commercial use without permission >> # <nettime> is a moderated mailing list for net criticism, >> # collaborative text filtering and cultural politics of the nets >> # more info: http://mx.kein.org/mailman/listinfo/nettime-l >> # archive: http://www.nettime.org contact: nett...@kein.org >> # @nettime_bot tweets mail w/ sender unless #ANON is in Subject: > > -- Be good, Fen
# distributed via <nettime>: no commercial use without permission # <nettime> is a moderated mailing list for net criticism, # collaborative text filtering and cultural politics of the nets # more info: http://mx.kein.org/mailman/listinfo/nettime-l # archive: http://www.nettime.org contact: nett...@kein.org # @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
