Backlinks

• 11_04_24_ClassNotes (TO)

  index

  How can we do something with AI that is note evil?

  • Plenty of people working in GLAM are doing thoughtful things with machine learning.
  • People working in GLAM used targeted questions and problems with their AI/machine learning
• Bender, Haigh, Other Observations (A)

  index

  **Emily Bender, "Thought experiment in the National Library of Thailand. [link]([https://medium.com/@emilymenonbender/thought-experiment-in-the-national-library-of-thailand-f2bf761a8a83](https://medium.com/@emilymenonbender/tho

• CLIP Experiment Lab Notes (TO)

  index

  Seeing the potential applications of CLIP as a result of using it today has been eye opening. The first place my mind went was to the potential for this to be an incredibly useful tool from GLAM. If every images of piece in t

• Free Writing 09_30_24 (TO)

  index

  Free Writing:

  • With what I understand now regarding AI I realize that I really had no clue what AI actually was before entering the class.
  • My understanding of AI prior to the course was that it scraped the internet 'learned' f
• Free Writing 10_07_24 (TO)

  index

  AI in the News.

  • articles tend to fall into one of two camps consistent with the 'Two War Visions of AI' reading, those being the 'accelerationists' and the 'doomers'.
  • When looking at some of the news outlets in Canada we see
• Getting PixPlot to Work with the GeoCities Dataset

  index

  After trying (and failing) to retrace my steps and gather a new dataset of GeoCities screenshots, I returned to my saved collection of images and tried to input it to PixPlot.

  With 250 images from the Enchanted Forest neighbo

• Language Models Can only Write Poetry - Allison Parish (TO)

  index

  • There is a widespread apocalyptic fear among those who write literature about how these language models have become so strong and able to produce poetry (for example)
  • Parish goes as far as to claim that computers can only w
• On LLM-Clip

  index

  What can we use LLM-Clip for to do historical research?

  The most obvious after exploring LLM-Clip looks like taking archives/datasets/etc. and using vibe searches to find parallels and similarities among images for sorting purp

• Perchance Doesn't Know Mountaineering History

  index

  I chose to use perchance.org to generate images of my historical prompt: 1996 Mount Everest Disaster.

  That is exactly how I inputted it to perchance, and I found my results to be a bit lackluster. Over and over again I was gener

• PixPlot_Close_Reading(TO)

  index

  The first issue I encountered was in the Google Collab notebook where I had issues with the Craiyon AI model. I moved to the Perchance model where I ran into the issue of my images being saved as a .jpg instead of .png. I then re

• Project Bestiary

  index

  Exploring latent space + GAN Image creation + Medieval bestiaries

  (thisisnotabesitary)

  I thought it would be cool to build a type of medieval bestiary using image Generative Adversarial Networks to create images, maybe even ba

• Retraining GPT-2 on a Medieval Bestiary

  index

  What animal is pure? thou hast studied nature.' Again he says: 'Not many.' Again: 'Not many.' Again: 'Not many indeed.' Again: 'Not many.

  Source material: Aberdeen Bestiary, a 12th century illuminated manuscript detailing

• Sept 30 - The Map Is Not the Territory

  index

  Computer systems - a confusion of map and territory

  • A question of semantics and understanding. The issues brought forward in Karawynn Long_Reading "Language Is a Poor Heuristic for Intelligence" about seeing language = i
• The Map and the Territory - Artificial Inelligence Being Taken as Actual Intelligence

  index

  The problem of AI is still very indistinct to me. Throughout the readings people have been twiddling their fingers and biting their fingernails worrying about the mass reproduction of biases into the architecture of society (as I

• Theocratic AI Rule

  index

  • How we frame AI and talk about it and the religious undertones that emerge
  • TESCREAL folks applying religious ideologies to AI and surrounding discourse Good/evil, heaven/hell, sin/penance, end of the world/final damnation, e
• There Was No ‘First AI Winter’ - Thomas Haigh (TO)

  index

  • Early AI as a branded effort to receive funding for computer science departments, but before too long was an important actual area of study.
  • Haigh explains that the ‘First AI winter’ (a period of downturn in AI research) is a
• Thought Experiment in the National Library of Thailand - Emily Bender (TO)

  index

  • Language models like ChatGPT only produce likely sequences of words, ‘If the text makes sense it is because we, the reader, make sense of it”
  • Systems cannot learn meaning but only form of language.
  • The linguistic forms creat
• Week 1 Readings - AL

  index

  There is a human element that is irreplaceable by AI, and an intelligence formed by human experience will never be completely replaceable by AI.

  There is no meaning behind the creations of AI, its simply fulfilling a request bas

• Week 2 Readings

  index

  I found this week to reinforce the idea that language models do not really understand what it is they are replying to or even saying, they are essentially giving an answer based on a probability.

  I enjoyed the thought experiment

• Week 6 Readings Notes

  index

  What struck me the most from this weeks readings was the study by Birhane et al., it highlighted that these models are using datasets that contain a high proportion of sensitive or dangerous content, including violent content. C

• Week 7 Reading Notes

  index

  Ai should be used as a supplemental tool, not a full replacement for workers and original human thinking and interaction.

  Many believe that it should completely replace human experiences

  AI can be used as a tool that provides n

• What Is ChatGPT Doing… And Why Does It Work - Stephen Wolfram (TO)

  index

  • Essentially what ChatGPT and other LLMs are doing is asking the question “Given the text so far, what should the next word be?” repeating this again and again after each new word based on the huge corpus of information that
• Why Can't AI Do It Itself

  index

  • The need for reinforcement learning human feedback The human cost of RLHF is clear, but what can we do about it? How can we go about this differently? There is a huge human cost for safeguards, but without the safeguards built
• Note on Karawynn Long

  index

  About the author: Karawynn Long is a Mexico-based sci-fi and fantasy author and novelist. Very aware of her social standings and intersectionality.

  1. Heuristic as an adjective vs in psychology are very different. IE. learn
• Sg-Sept9

  index

  HIST4805a

  ---

  Overall Plan for this Course

  • this course is an EXPERIMENT. Who knows where it'll lead?
  • who I am and why I'm interested in doing this ex
• Week 5 Garden and Discussion Notes

  index

  Accelerationism and Doomerism The coup at OpenAI basically a struggle between a nervous committee of professional worrywarts who believe the way AI is being developed will be harmful, especially to the already marginalized, v