I'm still moving stuff over from a Google Docs document. I recommend looking at that until I finish moving it over here.
TODO: Add tables
TODO: Copy over links
TODO: Add notes suggested by others.
Me and Zander interviewed Andy Matuschak in the SuperMemo.Wiki Discord in October 2020. You can check out the fll inter
view is available on YouTube.
- If you are interested in such events, join the Discord channel and sign up for my [newsletter] to get notified when there are interesting events like this.
- Thanks to Zehnn in the SuperMemo.Wiki Discord for doing the timestamps!!!
TODO: Add table
- Knowledge work is a "Wicked Domain" (Ericsson)
- Unclear what virtuosity looks like in knowledge work compared to ballet, music
- Nobel prize winners rarely become coaches, don’t record what fundamental skills people ought to develop to attain their levels of knowledge and success, so there’s no corpus of knowledge.
- Examples of fundamental skills for knowledge workers: attention management, developing inklings over time, communicating successfully in a team.
- What exercises could be used to develop knowledge work skills?
- Daily writing activities?
- Might be interesting to look at religious study practices for inspiration?
- Athletes and musicians pursue virtuosity in fundamental skills much more rigorously than knowledge workers do (Note by Andy Matuschak)
- Knowledge work (Note by Andy Matuschak)
- Core practices in knowledge work are often ad-hoc (Note by Andy Matuschak)
- We don’t know how to measure knowledge worker productivity (Note by Andy Matuschak)
- Knowledge work rarely involves deliberate practice (Note by Andy Matuschak)
- Peak: Secrets from the New Science of Expertise (Book by Anders Ericsson)
- Summary of Deliberate Practice (Paragraph from Wikipedia)
- The Making of an Expert (Article by Anders Ericsson)
¶ Question 2: How important are your notes to your learning and creative process? Can you give an example of a creative breakthrough or idea that you were able to generate through your note taking practices?
- Has been doing this for 1.5 years, no not sure how significant yet
- Most of his dramatic insights predate his note-taking system.
- More consistently makes progress on developing inklings.
- Quantum country emerged from 2 years of convos between himself and Michael Nielsen.
- Traditional mechanisms don’t enable the development of inklings well.
- More progress on inklings in parallel, plus more with original components, not just summaries
- Timeful Texts - written in just 3 hours based on 100s of notes - had already done MOST of the thinking.
- Why Books Don't Work - took 80 hours because he did the thinking at the same time as the writing.
- Timeful Texts (Essay by Andy Matuschak and Michael Nielsen)
- Why Books Don't Work (Essay by Andy Matuschak)
- Quantum Country (Essays by Andy Matuschak and Michael Nielsen)
- Still difficult to say; but he can develop ideas better for sure.
- Doesn't think he has published anything really significant over the past 1.5 years (check back in 3 years!)
- Andy Matuschak’s Notes
- Livestream of Andy’s Writing Practice
- Andy spent 5 years at Khan
- After he left, he wanted to stop thinking about formal educational contexts
- Agrees with some substantial fraction of the sentiments advanced by Illich, Woz with regards to Anti-educationalism, but more pragmatic.
- At Khan: INCREDIBLY difficult to make progress because kids already checked out, results from projects were biased / distorted.
- He has decided not to pay attention to traditional education systems.
- Problem of Schooling (Book by Piotr Wozniak)
- Deschooling Society (Book by Ivan Illich)
- Khan Academy R&D Reports
- Not all note takers don't have a serious context, just some.
- His favorite "serious context of use" example (from M. Nielsen) is NASA working on the Apollo program
- Goal was to put people on the moon, they actually developed a bunch of extra valuable knowledge, tools etc during the process
- Cathedrals - Goal was religion, developed loads of architectural knowledge
- Lots of the literature on Spaced Repetition sucks – not being undertaken in a serious context.
- Quantum Country: quantum country is earnestly trying to be a primer for quantum computing. Written for graduate students – for people who want to study it seriously.
- Spaced Repetition for Effective Learning (Literature review by Gwern)
- ‘Systems’ thinking — ideas from the Apollo space programme on effective management and a possible ‘systems politics’ (Article by Dominic Cummings)
- Still a mystery to him!
- Application prompts are questions which intentionally discourage memorization.
- Help you to apply what you learned.
- The application prompts are based on a concept, the phrasing of the question changes each time (cycles over 2 years)
- Intended to help transfer learning
- Working with Authors
- Expensive to write, but cheaper than exercises
- People sometimes aren’t confident in knowledge gained through SRS
- Learner’s perceptions of their knowledge vs reality after spaced vs massed learning
- Taxonomy of Note Types (Note by Andy Matuschak)
- Procedural SuperMemo (Blog by George Zonnios)
- Poor transfer in cases where the question can be "pattern matched”
- Clues in the syntactic formulation of the question.
- GPT3 is not great for formulating questions.
- What if you created items and then have GPT3 generate syntactic variations.
- What are the fundamental nouns and verbs in an SRS?
- What are the “primitives” of the system?
- No successful representation of progress in the systems - they show you the interval, statistics, but these don’t show progress towards the user’s real world goals.
- SRS seem like they don’t work because you spend most time on cards you are forgetting.
- Tools developed in academic environments are not developed and tested in serious contexts of use.
- Insight through making loop
- Everyone in SRS understates the role of emotion
- Most people don’t stick with SRS
- Insights through making (Note by Andy Matuschak)
- Analyse an ethical decision you faced using some philosophical framework.
- Analyse past experiences - when someone made an action that surprised you.
- How much value comes from writing the prompts vs engaging with them over time.
- Unusual applications of spaced repetition memory systems (Note by Andy Matuschak)
- Templates / scaffolds may help users to bootstrap their prompt writing skills.
- Andy feels like he still isn’t great at writing prompts.
- Struggled especially during 1st year, but benefitted from being able to ask
- Plans to run workshops on prompt writing for authors
- Then use that to write a book - like a book length version of the 20 rules.
- He has noticed that there are certain common templates for certain types of knowledge, but also lots of things that resist templatization.
- The 20 rules of formulating knowledge in learning (Article by Piotr Wozniak)
¶ Question 12: SRS allows you to gain and retain more knowledge. How, if at all, does it change the time allocation strategy in learning. Is it better to be a specialist or a generalist.
- SRS is downstream of learning activities.
- Doesn’t really change the fundamental mix.
- Learning is always in the service of curiosity or serious projects.
- Doesn’t study for the sake of studying.
- Learns in response to a need - programmer’s perspective.
- Common failure mode for people using SRS to try to learn everything, learn a bunch of trivia.
- Good strategy - eminent performer in one domain and pretty good in a couple of others.
- He had the book in front of him, not from memory.
- Makes markings to index the book.
- When he does deliberate study, eg. on deliberate practice, he makes multiple passes on the source material, making markings, reads again more carefully, and makes a map of the key ideas.
- Livestream of Andy’s Writing Practice
- Very interesting.
- Building his own prototype system, but not high priority right now.
- Rewarded by the sense of freedom
- There’s a feeling when reading a physical book that when you put it down, you will never go and pick it up again.
- Nouns and verbs wrong in SM, workflow wrong, doesn’t understand it enough to criticize it fully.
¶ Question 15: From Woz: “I just spoke with friends about converting the whole operation of humanity into a huge concept network. People would operate as neurons, define their needs, define their interests and define their capacities. They will organize into an efficient network of co-operation and knowledge exchange. We need software infrastructure to implement it. What do you think Andy?”
- Collective + networked intelligence literature.
- Currently lack the framing for a successful system of this kind.
- Military operations research - coordinating actions effectively between many people.
- Michael Nielsen - Reinventing Discovery - how to direct expert attention effectively.
- What are the subproblems eg. Attention.
- Worries that the ideas about the underlying system are too unclear.
- As a systems designer you are reifying properties of the underlying system, but the underlying theories are not powerful enough yet.
- Reinventing Discovery (Book by Michael Nielsen)
- Polymath Project (Article on Wikipedia)