Skip to main content

On AI

Just a collection of great opinion pieces that shaped my own view of the currently overhyped generative AI. Here I will shortly outline or quote the central aspects mentioned. The order has no meaning, it is just the order in which these articles came to my mind. If any of these summaries look interesting to you, please read the original article, condensing them into a few bullet points necessarily leaves out most of the nuance and richness of the original authors thoughts.

Speech by Meredith Whittaker

When she recieved the Helmut Schmidt Zukunftspreis (english)

  • AI is a marketing term
  • AI grew out of the toxic surveillance buisiness model and is now used to justify and cement the dominance of big tech companies over all of our lives
  • AI is highly unreliable
  • Therefore, deploying it in war to decide who will be killed is horrible

Please stop externalizing your costs directly into my face

By Drew DeVault (english)

  • AI web scrapers put enormous load onto every webserver and their sysadmins and is a giant wast of time and resources
  • As a response to this, the Web is now more closed than ever with cloudflare CAPTHA's or instances of Anubis in more and more places.

„Viele Menschen haben für Metas Profit gelitten“

By Ingo Dachwitz on netzpollitik.org (german)

  • Big tech is neocolonialistically outsourcing labelling of AI training data to the global south
  • this includes being exposed to e.g. sexual abuse content the whole day without getting any psychological support or even an appropriate salary

How much electricity does AI consume?

By James Vincent on TheVerge (english)

  • Generating four images uses as much energy as fully charging a smartphone
  • Data center energy usage is estimated to roughly double from 2022 to 2026 with a total energy consumption comparable to that of germany. Before the AI hype, data center energy usage was relatively stable.

These are not the same

By Tante (english)

  • AI does not scale and running it does not financially make sense
  • The current AI hype is a bubble which will burst eventually
  • The interesting question is "What does the bubble leave behind after its burst?"
  • With the AI bubble, this will be the massive infrastructure in terms of massive data centers filled with graphics cards. The most likely outcome seems to be that it will be used by companies like Palantir for doing mass surveillance for governments which is not something that will benefit society.

Are "AI" systems really tools?

By Tante (english)

  • "Tools are not just “things you can use in a way”, they are objects that have been designed with great intent for a set of specific problems"
  • "When I say “AI” here I am not talking about specialized machine learning models that are intended for a very specific case. [...] I am thinking about “AI” as it is pushed into the market by OpenAI, Anthropic etc.: “AI” is this one solution to everything (eventually)."
  • "And here the tool idea falls apart: ChatGPT isn’t designed for anything."

How AI is killing joy

By Leah Oswald (english)

  • AI is like frozen pizza
    • "The results of AI are kinda acceptable… sometimes, and it’s fast and cheap, because we ignore the invisible costs or the poor quality. AI is the fast food of creative processes. It works, but if we consider it fairly, it also sucks."
  • "AI is stopping our continuous development towards something new. We start to focus entirely on the past."
  • "AI is creating a small devil on your shoulder that nags you every time you see something you like if maybe it was created by a human or instead by a dead thing that has no actual idea about all this."
Back to top