April 3rd notes
ending term with a look at google - fears over computer access and fraud + the real world consequences - enshittification of everything
Slide - earliest screenshot of google (1998) - copyrighted by standford U - organizing information online
-
1990s vs 2020s web experience
-
GOOGLE
- in search of ways to exist on the web
- larry page begins google as a dissertation on the mathematics of the web
- inspired by the way academic literature has backlinks (citations)
- "if i can operationalize that, the backlinks that point to what's important"
- sets up webcrawler
- the idea of surfacing important information from a structural perspective rather than keyword approach
- originally AGAINST advertising
- dot com boom journalism*
- auctions for keywords + advertising
- early friendship between yahoo + google (stock in exchange for dropping a lawsuit)
- Famously has to bring in parental supervision for 2001 CEO, Page.
- in 4 years (1999 -> 2003) the amount to buy google goes from 1 million --> 3 billion
- dot com boom has crested by 2004
- google has its ipo by 2004, valued at 52 billion by 2005.
- yahoo, the early darling of the internet, fixes to build their own search engine (2004/2005)
- web 2.0, instead of building a site you are logging in to someone else's
- 2008 click fraud is a big problem
- google is using cookies to track individuals by 2008 (trying to mitigate the ad fraud)
- by 2010s google buys up other services or launches similar services to kill off competititors and keeps bad news under wraps.
-
the web as history vs history of the web
- communication studies / theories of mass communications
- John Thompson: “mass communication” -> “the institutionalized production and generalized diffusion of symbolic goods via the transmission and storage of information/communication." (1990)
- search engines can fit into this definition, but they dont produce narratives or stories.
- 2006, "google is simply an aggregator of information" (eric schmidt)
- they created the lens in which we view the web "i got it off of google"
- search engines do not access the whole web - things that are not accessible to the web crawler. favouring certain types of content.
- kremlinology
- tech is intertwined with the social, political, and economic context in which is arises. (E van couvering, 2008)
-
network science (hot in the early 2000s!)
- how to get value out of the web
-
the problem for someone putting stuff online - attention scarcity
- sitting in the spot of intermediation between our eye balls and where we wanna go
- Stop shitting in the well
- trusting results is a problem
-
early google webcrawlers
- travel the links on a webpage
- popular hub websites vs millions of low ranking websites
- this process was framed as democracy in its early days
- google = democracy (links are a type of vote)
- well linked pages are more prominent
- mainstream voices are reproduced and reinforced
- how we know google now vs. then
- original goal of organizing the worlds data has mixed motivations
- stay on their site, but click on the advertising
- paid placements (links within search results) as a move away from banner ads
- advertisements blended in with actual content
- birth of search engine optimization
-
Safiya noble -> hypervisibility as a means of rendering black women and girls invisible. TECHNOLOGY IS NOT NEUTRAL
- structurally and algorithmically we are pushed towards the status quo.
- studies google from a perspective based in black intersectional feminism
- the way that different identities intersect, particularly online
- searching white girls vs. black girls showing racially charged pornography for the latter.
- google responded instantly so that they could not replicate these searches.
- our own identity as we intersect with google (and culturally with the search engine) shows how all these horrible currents are pushed to the surface
-
recently firing of ethics researchers
- 2007 a company called metaweb launched a project called freebase - an open databasa that expressed data not in relational data, but as a graph.
- power of network analysis to traverse the graph to surface answers (triples)
- google buys this in 2010, contextual clues aiding in search engine results.
- announces knowledge graph in 2012 -> web on top of a web.
- drawing on wikipedias materials without attribution (reducing use of wikipedia)
- annotated vision of the world --> core data for large language models.
- take in everything they can without attribution or recompense.
- ethical researchers are fired for pointing this out (breaking copyright, don't be evil, etc.)
- too much money to be ethical
-
aaron swartz (rip) and his influence, work, and impact
- "Aaron's act was undoubtedly political activism, and taking such an act in the physical world would, at most, have a meant he faced light penalties akin to trespassing as part of a political protest. Because he used a computer, he instead faced long-term incarceration." - EFF
- this brings us into the last decade-ish
-
Last basic concept, the enshittification
- term coined by cory doctorow
- we are seeing how platforms die.
- Make a platform attractive, hold users hostage, abuse business customers to claw back the value for themselves.
-
Amazon for example!
- Sells goods and ships below cost, website is clean.
- get big fast
- locks you in with ebooks and audiobooks
- making it hard for original users to leave (cost switching)
- amazon prime is "free shipping" so why leave?
- amazon marketplace = trapping businesses by subsidizing their costs
- can't buy them anywhere but amazon, though
- amazon traps businesses and consumers
- preferred placement = extra fees
- surpluses directed to users, then once locked in they go to suppliers, once locked in they go to shareholders.
-
Facebook's enshittification ("'They trust me. Dumb fucks' - Zuckerberg")
- originally created to creep on girls
- early log in data was used to try and breach harvard crimson student reporters
- facemash inspires facebook but without the hot or not feature
- you can't take your social graph out of facebook
- shoving posts from companies/accounts you don't follow onto your feed
- readers don't go to the publications, completely dependent on facebook
- facebook & google have an illegal agreement where FB gets an illegal advantage in advertising auctions in exchange for FB canning its own service.
-
end-to-end principle (the idea that networks should be designed so that willing speakers' messages would be delivered to willing listeners' (doctorow)
-
Give things away, make it hard to leave, give you away.
- "Advertising funded search engines will be inherently biased towards the advertisers and away from the needs of consumers" (Page and Brinn 1998)
difference between content creators and writers/artists - perhaps the difference between the relationship with the tool - sharing ideas and monetization - platform receiving monetization (pushing content) / making money FOR a platform
- Drawing class to a close around the 2012 mark
- break throughs in neural networks and the emergence of convolutional neural networks for image classification leading to improvements in performance.
- ai is scary.
We made it to the end!
- Wednesday will be building elements for final memex
- themes and intersections
- head start on final journey
- easter monday = contractually obligated to be in class (we only come if help needed!)