[homework: asking, writing, thinking, doing, 🔴 red-teaming 😈] Critique what you have built; Real-World Data; Historiography of Data, Incentives, & AI

[irisrobyn]

Reading

Pro tip: try using an app on your phone or computer to read aloud to you at 1.5x speed! This can save time and make it easier to absorb information while not being tied down to a computer or device visually.

• [x] Review the bug report we submitted: https://github.com/duckdb/duckdb/issues/7015 and the fix to avoid the bug we ran into for analyzing Zulip data: https://github.com/onefact/datathinking.org-codespace/blob/main/notebooks/in-class-notebooks/230420-zulip-chat-duckdb-loading.ipynb
• [x] Review several editions of Style, Clarity and Grace (https://bookshop.org/p/books/style-lessons-in-clarity-and-grace-joseph-williams/8980182 - Libgen)
• [x] [climate harms] Read https://arxiv.org/ftp/arxiv/papers/2104/2104.10350.pdf
• [x] [global context thanks to @indrekromet] Read https://www.palladiummag.com/2023/02/23/the-west-lives-on-in-the-talibans-afghanistan/
• Preparation for guest lecture
  • [x] Read https://arxiv.org/abs/2302.10329
  • [x] https://www.economist.com/science-and-technology/2023/04/05/it-doesnt-take-much-to-make-machine-learning-algorithms-go-awry (https://archive.ph/5l1k3#selection-1038.0-1038.1)
• [x] [David Foster Wallace on Consumption] https://harpers.org/wp-content/uploads/2008/09/HarpersMagazine-1996-01-0007859.pdf
• [x] [Cross-Cultural Research and Ethnography] Read excerpt from https://www.thefieldstudyhandbook.com/ - https://drive.google.com/file/d/1F9zr1pHesGPSOTtXLQu4hTxOGu7zoIRx/view?usp=sharing
• [x] [Status games, social engineering] Read the chapter on status games in Impro (https://bookshop.org/p/books/impro-improvisation-and-the-theatre-keith-johnstone/11721204 or Libgen)
• [x] [Closure on the data journey begun in the first class on COVID] https://www.wsj.com/articles/covid-origin-china-lab-leak-807b7b0a
• [x] [History of Computing and Data Thinking] Read the Epilogue of Computer Age Statistical Inference by Hastie et al., https://hastie.su.domains/CASI_files/PDF/casi.pdf
• [x] [Basics of GitHub] https://github.com/onefact/github-starter-course
• Systems Thinking and Incentives:
  • [x] [Context on how big 4 consulting firms may/may not be replaced by AI] https://www.lrb.co.uk/the-paper/v44/n24/laleh-khalili/in-clover (https://archive.ph/yC1YQ)
  • [x] [High finance & geopolitics] https://www.theguardian.com/world/2016/jul/28/1mdb-inside-story-worlds-biggest-financial-scandal-malaysia
  • [x] [High finance & geopolitics] https://www.bloomberg.com/news/articles/2014-08-21/mugabes-bailout-och-ziff-investment-linked-to-zimbabwe-despot#xj4y7vzkg?leadSource=uverify%20wall
  • [x] [Predatory Investing] https://www.newyorker.com/magazine/2018/08/27/paul-singer-doomsday-investor
  • [x] [Global Geopolitics of AI] Chapter 1 - China's Sputnik Moment - and Chapter 4 - A Tale of Two Countries - of Kai-Fu Lee's AI Superpowers (https://bookshop.org/p/books/ai-superpowers-china-silicon-valley-and-the-new-world-order-kai-fu-lee/6960836 or Libgen)

Doing

• [x] Edit https://en.wikipedia.org/wiki/Data_thinking using the learner definitions of data thinking
• [x] Try playing one round of https://lost-at-sql.therobinlord.com/
• [x] Install https://marketplace.visualstudio.com/items?itemName=charliermarsh.ruff and read the documentation, and activate the extension for python files opened in Visual Studio Code: https://beta.ruff.rs/docs/tutorial/
• [x] Red-teaming Previous Homework:
  • [x] [Set timer: 10 minutes maximum] Before asking GPT (to avoid biasing yourself!), write your own critique of your homework. Questions to consider could be: what could be improved? What doesn’t make sense in the visualization? What doesn’t make sense in the writing?
  • [x] [Set timer: 10 minutes maximum] Ask GPT-4 to critique the homework or the visual using your favorite data thinking definition we have so far.
  • [x] [Set timer: 10 minutes maximum] Add this critique as a comment on the homework’s github issue, and link to the critique in Zulip.
  • [x] [Set timer: 10 minutes maximum] Repeat this exercise for the previous homework of one other person in the class

Creating

• [x] Generate art or diagrams or logos using https://beta.dreamstudio.ai/generate and post on Zulip chat
• [x] Use https://poly.cam/ to scan an object or location of your choosing; share the link on Zulip (bonus: feel free to try BlenderGPT: https://github.com/gd3kr/BlenderGPT to modify the 3D scan)
• [x] Use duckdb to load the Zulip data into a SQL database, and use altair to visualize the data, following https://github.com/onefact/datathinking.org-codespace/blob/main/notebooks/in-class-notebooks/230420-debugging-duckdb-altair-falcon-3-1-1-service-requests.ipynb (run this notebook with this data: https://data.cityofnewyork.us/Social-Services/311-Service-Requests-from-2010-to-Present/erm2-nwe9 - and try changing the data source to be the Zulip data and post a visualization of the chat data on Zulip)

Thinking

• [x] In a comment on this issue, propose a research question that you will aim to answer the upcoming final homework (and cross-post it on Zulip for feedback); include a rationale for why this topic matters to you, and decisions you need to make, and a clear articulation of the stakes. This can be a paragraph or several. Perhaps start with your experience report from the first homework.

Listening

• [x] Listen to this podcast with Emily Bender: https://open.spotify.com/episode/2ILGlkAXAt4xfuKHwIV2on

Large Language Model Access Checklist

• [x] Try GPT-4 at https://chat.openai.com/chat (ask on Zulip if you need access)
• [x] Try https://serge.chat/
• [x] Try https://bard.google.com
• [x] Try https://console.anthropic.com/claude (ask on Zulip if you have not received an invite)
• [x] Try https://www.phind.com/
• [x] Try Copilot (https://github.com/features/copilot; free with GitHub Education Pack)
• [x] Try Bing Chat (https://bing.com/chat)
• [x] Try Huggingface Chat (hf.co/chat)

[irisrobyn]

An emotional prelude: I think I still want to carry on with the narrative that maybe innovation has had some affects on the mental health of society. The idea is that as innovation accelerates we don't have time to collectively adapt to it. This could possibly lead to innovation uprooting our lives instead of making it easier (eg losing jobs or entire industries, losing basic routines humans have evolved on). I have this idea that if you looked at a brain from 30 years ago and now you could visibly see the differences or brainwave imaging definitely shows intense differeneces due to the way we experienced world then and now. I think the main thing is how much information we go through which very closely relates to our emotions and hormones. 100 years ago a person knew maybe 40 people - completely made up statement just to illustrate the idea. Today we can follow people from around the world, keep up with what everyone from our childhood is doing, what horrible things are happening on every continent, how local changes are shaping worldwide crisis and so on. Not to even mention being constantly online for your own network - no more "He called to say he's taking a bus - should be here in an hour" and wondering where he is in an hour, but more imporantly no more solitude waiting for the bus, self-reflection on the bus, interacting with the world around us to find the way in a new city once off the bus. Now it's local, global, pop, alt, political and meme news waiting for the bus, someone's weird way of making a casserole and a public feud between a broken up couple on the bus and quick gps with the same music you never really chose while navigating to your destination. The bigger idea is that we are seeing the massive and deep-rooted affects of modernization and globalization over a long period of time - they simply could not be predicted since it is so all-embracing. Now we are looking at another leap and before blindly taking it, we should maybe rule out the possibility that we were better off before coming this far.

Hypothesis: Accelerating innovation plays a role in declining mental health.

Considerations:

1. Mental health data over the years is most likely unreliable. It is a widely known fact that mental health conditions have been poorly documented til recent years. Many conditions have only lately been described.

2. Mental health is a complex issue rarely with one concrete culprit. Even if the data shows mental health declining in well documented periods and a correlation with innovation acceleration, this is a very primitive way to describe a very complex and big system with infinite players. This is why I am well aware that the hypothesis is unprovable, scientific method simply does not allow for this kind of approximation. I see the value in just the discussion. I don't think there is much to do about stopping innovation and I don't belive in it, but I do think innovation should be deliberate, thought through and cautious as well as truly user-friendly in the sense that it considers how the user can healthily adapt to it.

3. Everything has a learning curve and we tend to focus on the negative. Maybe AI is the next technology that fixes all of our problems. Most people today live a much more comfortable life than 20 or a 100 years ago. We have a bias for memories - nostalgia, and a neverending motor trying to make the present and future better - pointing to it's flaws.

4. I am setting myself up for failure. The data is probably hard to find and analyse. There really can't be any conclusions made even if a correlation is found. The topic is too broad and emotional.

5. I don't have any better ideas and this has been bothering me for months now. I am going to try to see this through but a warning and declaration of bias: the whole concept of this work is rooted in generational depression which is quite bleak. I do not do well with writing.