Steve Jankowski
Assistant Professor in New Media Histories at UvA. Curator of Keywords for Studying Media, Culture & Information https://textaural.com/keywords/. I study Wikipedia, utopian computation, design and consensus politics, often all at once.
- I'm doing some manual data entry for a citation analysis and Google just blocked me. I'm as fast as an automated system 😄
- Huh. Rockwell's "Freedom of Speech" painting-turned 2020s meme was used in 1997 for Wired Magazine's editorial on "The Digital Citizen". m.media-amazon.com/images/I/A1a...
- Interesting that this figure is part of the archetypal form of what the digital citizen is expected to look like and how their dissenting actions are framed.
- Reposted by Steve JankowskiNew @ursulakleguin.com print from @skeuomorphpress.org arrived
- Read @laurenfklein.bsky.social's arxiv.org/abs/2502.19190 which presents 8 claims for humanities AI research. They suggest "2) GenAI requires an expanded definition of culture." This is 👍 & connects to work (mine 😉) on culture and technique: journals.sagepub.com/doi/full/10....
- They also have a point "; 3) Generative AI can never be representative" – I've been advocating for using the term "indexical" over "representative", as I feel representative in some cases means proxy/approximate, whereas indexical (following Pierce) means "to point to something" real.
- 🚨 CFP: In Defense of the Commons 🚨 Communication+1 Deadline: January 15, 2025 openpublishing.library.umass.edu/cpo/news/24/ Looking for papers to diagnose problems & explore the strategies of resistance, repair, and rehabilitation necessary to sustain vital public digital (and analog) resources.
- Some fun with Internet periodization. 1980s: Communalization of the Internet (Rheingold, 1993) 1990s: Webification of the Internet (Lévy, 2001) 2000s: Googlization of the Internet (Siva 2012) 2010s: Platformization of the Internet (Helmond 2015) 2020s: Synthesization of the Internet (Berry, 2025)
- Looking for a list of recent work that theorizes the current moment of making the Internet (and the web) "AI-Ready", just like how the Internet was made "platform-ready" (developer APIs) and "Google-ready" (SEO). I think Berry's work is in the right direction link.springer.com/article/10.1...
- We need to be looking at things like RAGs (arxiv.org/abs/2005.11401) , agents.md, and the other ways that developers are thinking about this (like on hackernews: www.google.com/search?clien...
- And also W3C www.w3.org/reports/ai-w...
- Reposted by Steve Jankowski🚨Calling all new media scholars🚨 If (and I emphasize if) a 2nd edition of the New Media Reader happened to be in the works, what additions would you like to see? What do you think could be excised? @docmofo.bsky.social @n-w-f.bsky.social @psssssssss.bsky.social mitpress.mit.edu/978026223227...
- I don't know how this wasn't on my radar, but Robert Darnton published his _Business of Enlightenment_ with a CreativeCommons license in 2016. He really set the stage for thinking of the materiality of encyclopedias. What a gift! archive.org/details/Busi...
- I love seeing old lectures published as webpages. 2004: gyre.umeoce.maine.edu/physicalocea...
- Every three minute startup pitch is a comedy bit of capitalist theatre.
- On Nov 3, I sampled 1K Grokipedia v0.1 articles and found that about 55% of the 1000 articles were based on Wikipedia because they used its CC license. I repeated the method on those same articles today and those CC licenses are gone. Grokipedia v0.2 looks to be all generated text now.
- Reposted by Steve Jankowski🚨 Bookmatch is back! 🚨 From now until December 2, make a donation of any amount to n+1 and we’ll send you the Bookmatch quiz—a personality test that will generate a reading list tailored to your tastes and whims. Try the quiz here: secure.givelively.org/donate/n1-fo...
- Reposted by Steve Jankowski“I see the TikTok, I see more, I get interested, I look it up online.” www.theverge.com/cs/features/...
- 1980s: Personal computer > Noun object 1990s: Computer network > Connected noun objects 2000s: Cloud computing > Global verb 2010s: Computation > Noun from process 2020s: Compute > Nouned verb
- Reposted by Steve JankowskiI have a piece in @techpolicypress.bsky.social today: With Grokipedia, Top-Down Control of Knowledge Is New Again Trying to consider the nature of the project and its vision of neutrality vs. @wikipedia.org's (without getting into e.g. article comparisons). www.techpolicy.press/with-grokipe...
- It looks like Grokipedia is at least a partial fork of English Wikipedia (which is allowable under Wikipedia's license). Verbatim paragraphs on the "Encyclopedia" page and includes Wikipedia artifacts (like the use of Wikitext: {{cite book}}). [Left image is Grokipedia; Right image is Wikipedia]
- Telltale signs about whether it is generated or a Wikipedia-forked page might be found in the References section. The "Nobel Prize for Literature" (which was suggested on the front page of Grokipedia) has just URLs for the references. Compare that with the Wikipedia-fork of "Encyclopedias."
- Well I'm a dumb-dumb. It literally says that it is a version of the Wikipedia page. So no mystery there.
-
View full threadEdit: "edit history" is the wrong term. There are no dates attached to the edits.
- Fine. I will have to wade more into this eventually. But for now, it took me literally 1 minute of reading my first article on Grokipedia before I couldn't help but make a head tilt: "Pre-Columbian Indigenous Societies" makes absolutely no sense for an article on "Canada."
- The "Canada" article also includes what looks like an artifact of the generation process where the Grok is talking to itself ... maybe? "– wait, no wiki, alternative"
- Also: here's the auto-suggest for articles related to abortion between Grokipedia and Wikipedia.
- @elissawelle.bsky.social Just saw the article about Grokipedia, and if anyone is looking into academic/research background about AI, Wikipedia and encyclopedias, either myself or some of my co-authors work might be useful. www.theverge.com/ai-artificia...
