markov_user

Karma: 411

markov_user Aug 29, 2021, 9:29 AM
2 points
0 ∶ 0
on: Lessons from Running Stanford EA and SERI
Great post, thanks for sharing! Pretty much exactly the type of post I had been hoping for for a while. Just hearing that one success story of a local group that was in a more or less similar state as mine (albeit arguably in a higher potential environment), but made it into something so impressive, is very inspiring.
Given I only have ~10h per week available to spend on EA things (and not all of them go into community building), I was particularly happy to hear your ⁸⁰⁄₂₀ remark. I do wonder if it’s possible to move a local group onto a kind of growth trajectory at only, say, 6h per week, or if that’s just a lost cause. Maybe I should just spend the majority of these 6h looking for a person with more time and motivation to take over the role. :)
Currently we’re definitely leaving a lot of low hanging fruit on the table (or tree) though. And a lot of that may be due to relatively trivial issues and inconveniences. Some examples of such limiting factors (and I do wonder if similar things are true for other small local groups):
- I’ve heard fellowships mentioned & recommended a lot in the last 1-2 years, but have a fairly limited understanding of the concrete details. Should we run our own one? Should we redirect people to other online fellowships? What would I even tell people in order to motivate them to do so? Timing also needs to be taken into account.
- Fear of organizing things and (almost) nobody (new) showing up. We had quite a few talks and such that ended up only being heard by our core team, although we were hoping to attract some new faces. That being said, our marketing was often pretty shy rather than aggressive.
- Lack of detailed knowledge about the European data protection regulation and its implications prevents us/me from systemizing our “funnel” (which hardly exists). I have no idea if it’s even legal to have a database of names / email addresses / other personal information of people, whether we’d need to inform them beforehand, etc.
- Most of our small number of members are busy with their own things / studies / careers and have hardly any capacity to engage with the group beyond one weekly social/discussion, so there’s little room for organizing bigger things or spending more time on community building, and I find that situation somewhat demoralizing
- We have a whatsapp group and a Slack workspace. Whatsapp is great to get new people on board quickly, but it’s surprisingly difficult to get them to sign up on Slack, and if they do we can never rely on them seeing new messages, or looking in there at all. Right now our Slack workspace is almost exclusively used by our few core members, and others hardly ever engage.
- I feel very aversive to “push” people to do things, and wonder if that’s a necessary skill to have for a community builder, or ideally people should be motivated enough that they only need to be “enabled”/supported instead.

markov_user Aug 28, 2021, 12:29 PM
1 point
0 ∶ 0
on: Philosophy Web—Project Proposal
It sounds interesting, albeit to be fair a bit gimmicky as well. To me at least, which may not mean much: I can imagine taking a few minutes to play around with such a tool if it existed, maybe find some contradiction in my beliefs (probably after realizing that many of my beliefs are pretty vague and that it’s hard to put these hard labels on them), and get to the conclusion that really my beliefs weren’t that strong anyway and so the contradiction probably doesn’t matter all that much. I can imagine others would have a very different experience though (and maybe my expectation about myself is wrong as well of course).
I’d be interested in your thoughts on a few questions:
1. Can you describe an example “user journey” for Philosophy Web? What beliefs would that imaginary user hold, how would they interact with the software, what would come out, just as one prototypical example?
2. Would there be other, maybe simpler ways for that imaginary user to get to the same conclusion, not involving Philosophy Web? What bottleneck prevents people from making these conclusions?
3. Who would be the primary target audience for this? What would make the tool “effective”? Are you primarily thinking about EAs getting to a more self-consistent belief set? Philosophy students? Everyone?
4. What are the most likely ways in which such a project would fail, given you found the necessary support to build it?
5. Does the project’s success depend on some large number of users? What’s the “threshold”? How likely is it to pass that threshold?
6. What would be the smallest possible version (so MVP basically) of the project that achieves its primary purpose? Could something be prototyped within a day that allows people to test it?
7. Assuming the project is built and completed and people can use it as intended—what are the most likely reasons for members of your target audience to not find it useful?
As an additional note, I’m quite a fan of putting complex information into more easily digestible forms, such as mind maps, and could imagine that “data structure” in itself being quite valuable to people merely to explore different areas of philosophy, even to a limited degree. I’m not quite sure though if the project entails such a web being presented visually, or if users would only see the implications of their personal beliefs.

markov_user Aug 28, 2021, 11:47 AM
2 points
0 ∶ 0
on: Open call for EAs with passion for meta-learning <3
Just wanted to say I very much like the idea, although I’ll probably not get involved myself. I was very happy about the anki deck of EA key numbers that was published two months ago, and would find it great if there were more ways to easily add important EA ideas to one’s anki deck (e.g. you mention the 80,000 Hours key ideas in the google doc, great idea!).

markov_user Aug 25, 2021, 6:49 PM
7 points
0 ∶ 0
on: How much money donated and to where (in the space of animal wellbeing) would “cancel out” the harm from leasing a commercial building to a restaurant that presumably uses factory farmed animals?
It would be quite surprising to me if your idea did not work out, simply because doing good for animals via donations tends to be really low cost (but might depend on what “a lot more money” really means in your case). Imagining for instance that for each and every restaurant in the world some non-negligible cut of the rent (say 5%) would go into effective animal charity, my super rough 3 minute Fermi estimate says that would amount to something in the order of $10 billion per year. Given that about 80 billion land animals are slaughtered each year, that would mean that at a cost effectiveness of sparing 8 animal lives per dollar donated (which doesn’t sound entirely unrealistic), your suggested approach to leasing to restaurants would, on a global scale, not only be net positive, but very theoretically end factory farming of land animals (obviously not in practice given diminishing marginal returns). It’s a very hypothetical argument, but maybe it adds something.

Apart from that, maybe there’s a way to attract more vegetarian/vegan restaurants in particular? No idea about the concrete processes and legislature around that, but maybe you have some power in that regard.

markov_user Aug 21, 2021, 9:45 AM
3 points
0 ∶ 0
on: Teaching You How To Learn post 1 is live!
Some random thoughts from me as well:
- I wonder if different people may have quite different bottlenecks with regards to how to learn most effectively, and it may be not so much about “do these things” but rather “from these typical bottlenecks, which one affects you the most?”
- the framing of “The best way to learn” seems a bit dangerous to me; even if “scientifically proven”, it still basically just means that it works well on average, but not necessarily for everybody. While active recall and spaced repitition probably are indeed very general, it might be good to add a few notes regarding how people might differ.
- on a similar note, 80,000 Hours tends to incorporate “reasons why you might disagree” or “where we’ve been wrong in the past” kind of sections and articles, which too I feel would help a little. E.g. “things Anki isn’t ideal for”, which definitely exist.
- maybe a relevant part of effective learning is to be more aware of one’s true motives in doing things, be it getting a degree, reading non-fiction books, having an anki routine etc., and whether one’s truly doing this to learn things, and if so for what exact purpose
- related to this, there are different dimensions to learning, similar to productivity: what are you learning (and why), how are you learning, and how much time are you spending. So basically direction, quality, quantity of the learning process. It seems that many resources, maybe including your site, mostly focus on the quality part, whereas the direction part may be even more important and comparably neglected.
- during one of the EAG Virtual conferences I talked to somebody who was involved in creating a free ebook on the most effective learning strategies for students during pandemic times; wasn’t able to find it again so far, but if I do I’ll add a link
- I personally would find it very useful to get some better/clearer mental models of learning and knowledge. Maybe the kind of thing Spencer Greenberg tends to do, e.g. in his podcasts, where he frequently goes into “Well I think X can be broken down into 4 categories: …” mode and suddenly X makes way more sense than it did before that breakdown.
- for a long time I’ve been of the conviction that the way we tend to structure information is highly suboptimal. I’m mostly referring to linear texts about things. 1. Texts are good for some things, but by far not for everything, 2. we’re not at all using our brain’s immense capability for spatial and visual processing, 3. texts are static and non-interactive, 4. while you have things like table of contents, chapters/headlines and some formatting, it’s not an ideal implementation of “different zoom levels”, and there are certainly better ways of letting people learn things on a very high level first and then “zoom in” further. As a learner, you have to take what you’ve got of course. But the other side of the coin—how can you make learning for others easier as a content provider of any sort? - seems very important as well, and I think such a page would be in a great position to experiment with such ways, and not rely on classical linear text form.
About the concrete project:
- I think providing anki cards at the bottom of your posts is a great idea
- 80,000 Hours tends to have small summaries of their articles at the top, which I would find useful here as well
- The Key Ideas Guide post is currently very text-heavy, which makes sense since it’s in progress and you probably want to focus on the ideas themselves rather than the presentation. For the future though I think it would make it much more digestible if there was a bit more variety to it, be it pictures, graphs, or even just some formatting tweaks. E.g. one or two screenshots from actual anki cars would be a start, or a graph of the forgetting curve.
- Style-wise, you’re using parantheses a lot in your post, which I can totally relate with—I do it all the time e.g. when exchanging messages with people or writing forum posts and comments. But it does still seem sumoptimal to me, as it hurts the reading flow, and may be a sign one’s not focusing on what’s actually essential.
- The post to me feels quite a bit like it’s trying to sell me something. I was almost expecting a “subscribe to my newsletter to get a FREE ebook!” while reading. :) This is something 80,000 Hours avoid pretty well by being very open and grayscale about things.
- I find it great that you’ve just started doing it and putting it out there looking for feedback; I’m working on one or two vague similar-ish projects (not related to learning though) and didn’t yet manage to get over my semi-perfectionist “I’ll just make sure I have something good before showing it to anybody” attitude, although I know that’s a bad approach
- minor note, at one point you write “(god this bold is intense)” although there’s nothing actually bold; maybe the formatting got lost somewhere on the way?
Some counter points on drawbacks/challenges of Anki:
- you need to be rather conscientous to use it effectively; missing a week can easily break the habit of daily ankiing, because you’re suddenly looking at potentially 100s of flashcards to review
- it might push people to go for memorizing (often useless) facts rather than really learning and understanding deeper concepts
- also, adding anki cards to your deck now feels like progress; e.g. after reading a book (or chapter), you might have a feeling that not creating new cards is bad. This might nudge you to add useless cards rather than nothing, degrading the quality of your deck over time. I find it really hard to prevent this personally. After reading a book and going through my notes, if I add nothing to my Anki deck, I feel like having read the book was a waste of time. So I’m motivated to add things simply to feel better about the sunk cost. But looking at my deck honestly, I’m almost sure 50% of the stuff in there doesn’t really add anything to my life.
- setting up such a system and getting into it takes a lot of work and willpower, and many people may just not be willing to go that far (even if it does indeed pay off in the long term)
That all being said, if I went back to university, I’d definitely use Anki and I’m sure it would improve my performance a lot compared to my time there in the past where I didn’t know what spaced repetition even is. I’d just say that it’s maybe something like 40% of my personal ideal learning system, and there would be a lot beyond that (e.g. how to watch lectures, how to take notes, how to work on actual exercises, the fact that explaining things to others is very helpful, how to motivate yourself, how to plan and build a reliable system, …).

markov_user Aug 6, 2021, 4:06 PM
3 points
0 ∶ 0
on: How to Train Better EAs?
I recently read Can’t Hurt Me by David Goggins, as well as Living with a SEAL about him, and found both pretty appealing. Also wondered whether EA could learn anything from this approach, and am pretty sure that this is indeed the case, at least for a subset of people. There is surely also some risk of his “no bullshit / total honesty / never quit” attitude to be very detrimental to some, but I assume it can be quite helpful for others.

In a way, CFAR workshops seem to go in a similar-ish direction, don’t they? Just much more compressed. So one hypothetical option to think about would be to consider scaling it up to a multi-month program, for highly ambitious & driven people who prioritize maximizing their impact to an unusually high degree. Thinking about it, this does indeed sound at least somewhat like what Charity Entrepeneurship is doing. Although it’s a pretty particular and rather “object-level” approach, so I can imagine having some alternatives that require similarly high levels of commitment but have a different focus could indeed be very valuable.

markov_user Jul 17, 2021, 12:25 PM
4 points
0 ∶ 0
on: Building my Scout Mindset: #1
Thanks for making this public, found it really interesting to follow your train of thought. Also, despite hearing about it in the past, I had completely forgotten about Julia’s book. Added it to my reading list now. :)

markov_user Jul 13, 2021, 11:08 AM
3 points
0 ∶ 0
on: Call for participants to test a pilot forecasting training program
How much time should a participant roughly allocate for this? How much time are we supposed to spend on each of the questions? For how many days/weeks/months will this be running?

Is “start by finding someone to practice with” something one should do before signing up, i.e. should people sign up in groups of 2? Or does that matching of participants happen once you’ve got enough together? If the latter, do you have control over which of the two roles you get? I couldn’t yet make that much sense of the descriptions of what backcaster and retriever are doing exactly, specifically the “pick a date” part and how the date influences things.

What degree of forecasting experience are you looking for? Or all types of people? Would it make sense for people to sign up when they’ve gone through a lot of calibration training in the past?

And a side note, the first paragraph on the linked page seems to have been pasted twice.

markov_user Jul 4, 2021, 1:16 PM
3 points
0 ∶ 0
in reply to: PeterSlattery’s comment on: Anki deck for “Some key numbers that (almost) every EA should know”
+1 to that! Really cool, thanks for doing this. :)

markov_user Jun 25, 2021, 8:49 PM
4 points
0 ∶ 0
on: On Sleep Procrastination: Going To Bed At A Reasonable Hour
Thanks a lot for the thorough post Emily! I like the framing of staying up late as a high-interest loan a lot. And I agree that reading Why We Sleep may indeed be quite useful for certain people, despite its shortcomings. You make a lot of good points and provide several interesting ideas, plus the post is written in a very readable way, and the drawings are great.

Not that much else to add, except two tiny nitpicks regarding your estimation:
- you equated “being 30% less productive” with “taking 30% more time to complete things”, but actually being 30% less productive would mean you take ¹⁰⁰⁄₇₀ − 1 = ~42% longer. (a more obvious example of this would be that being 50% less productive means you require twice the time = 100% more, not 50% more)
- concluding your estimation, your multiplication characters were interpreted as formatting, making the “0.254.9 + 0.502.1 + 0.10*0.4” part quite confusing to read. You could use × or • instead.

markov_user Jun 4, 2021, 9:19 AM
1 point
0 ∶ 0
in reply to: DC’s comment on: How can I best use product management skills for digital services for good?
Same here. :)

markov_user Apr 16, 2021, 6:01 PM
1 point
0 ∶ 0
on: Statistics for Lazy People, Part 2
Neat! Small mistake: “What is the probability that it will still be working after eight twenty years” should probably be “after twenty years”. And multiple data points are exciting indeed!

markov_user Apr 2, 2021, 4:32 PM
16 points
0 ∶ 0
in reply to: EdoArad’s comment on: Announcing “Naming What We Can”!
Perfect! In the end the impact will of course be orders of magnitude higher, as a slightly better name of any particular organization will affect tens if not hundreds of thousands of people in the long run. And there may even be a tail chance of better names increasing the community’s stability and thus preventing collapse scenarios. I think overall you really undersold your project with that guesstimate model only focusing on this post only, as if that was all there is to it.

markov_user Apr 2, 2021, 12:42 PM
30 points
0 ∶ 0
on: Announcing “Naming What We Can”!
I believe there are a few serious flaws in your guesstimate model:
- a year has 365.2421905 days, not 365.25. That’s not even rounded correctly!
- Smiles per QALY should multiply days in a year with smiles in a good day, instead they are added. They don’t even have the same unit, how can you add them! Insanity!
- the post’s karma is far outside of even your 99% interval
Everything else seems quite correct and I agree with your CIs and conclusions.

Also, please find a new name for guesstimate.

markov_user Apr 2, 2021, 12:19 PM
1 point
0 ∶ 0
on: Statistics for Lazy People, Part 1
Nice post! Found it through the forum digest newsletter. Interestingly I knew Lindy’s Law as the “Copernican principle” from Algorithms to Live By, IIRC. Searching for the term yields quite different results however, so I wonder what the connection is.

Also, I believe your webcomic example is missing a “1 -”. You seem to have calculcated p(no further webcomic will be released this year) rather than p(there will be another webcomic this year). Increasing the time frame should increase the probability, but given the formula in the example, the probability would in fact decrease over time.

markov_user Jan 30, 2021, 9:52 AM
2 points
0 ∶ 0
on: EA Münster Predictions 2021
“Bei 80% der Treffen der EA Münster Lokalgruppe in 2021 waren mehr als 5 Personen anwesend”—how will cancelled meetups (due to lack of attendees, if that ever happens) count into this? Not at all, or as <=5 attendees? (kind of reminds me of how the Deutsche Bahn decided to not count cancelled trains as delayed)

Also, coming from EA Bonn where our average attendance is ~4 people, I find the implications of this question impressive. :D

markov_user Jan 2, 2021, 11:39 AM
2 points
0 ∶ 0
in reply to: MaxRa’s comment on: One’s Future Behavior as a Domain of Calibration
I see, so at the end of the day you’re assigning a number representing how productive the day was, and you consider predicting that number the day before? I guess in case that rating is based on your feeling about the day as opposed to more objectively predefined criteria, the “predictions affect outcomes” issue might indeed be a bit larger here than described in the post, as in this case the prediction would potentially not only affect your behavior, but also the rating itself, so it could have an effect of decoupling the metric from reality to a degree.

If you end up doing this, I’d be very interested in how things go. May I message you in a month or so?

markov_user Jan 2, 2021, 11:30 AM
2 points
0 ∶ 0
in reply to: MaxRa’s comment on: One’s Future Behavior as a Domain of Calibration
Good point, I also make predictions about quarterly goals (which I update twice a month) as well as my plans for the year. I find the latter especially difficult, as quite a lot can change within a year including my perspective on and priority of the goals. For short term goals you basically only need to predict to what degree you will act in accordance with your preferences, whereas for longer term goals you also need to take potential changes of your preferences into account.

It does appear to me that calibration can differ between the different time frames. I seem to be well calibrated regarding weekly plans, decently calibrated on the quarter level, and probably less so on the year level (I don’t yet have any data for the latter). Admittedly that weakens the “calibration can be achieved quickly in this domain” to a degree, as calibrating on “behavior over the next year” might still take a year or two to significantly improve.

markov_user Dec 31, 2020, 3:50 PM
3 points
0 ∶ 0
on: One’s Future Behavior as a Domain of Calibration
I personally tend to stick to the following system:
- Every Monday morning I plan my week, usually collecting anything between 20 and 50 tasks I’d like to get done that week (this planning step usually takes me ~20 minutes)
  - Most such tasks are clear enough that I don’t need to specify any further definition of done; examples would be “publish a post in the EA forum”, “work 3 hours on project X”, “water the plants” or “attend my local group’s EA social” – very little “wiggle room” or risk of not knowing whether any of these evaluates to true or false in the end
  - In a few cases, I do need to specify in greater detail what it means for the task to be done; e.g. “tidy up bedroom” isn’t very concrete, and I thus either timebox it or add a less ambiguous evaluation criterion
- Then I go through my predictions from the week before and evaluate them based on which items are crossed off my weekly to do list (~3 minutes)
  - “Evaluate” at first only means writing a 1 or a 0 in my spreadsheet next to the predicted probability
  - There are rare exceptions where I drop individual predictions entirely due to inability to evaluate them properly, e.g. because the criterion seemed clear during planning, but it later turned out I had failed to take some aspect or event into consideration^[1], or because I deliberately decided to not do the task for unforeseeable reasons^[2]. Of course I could invest more time into bulletproofing my predictions to prevent such cases altogether, but my impression is that it wouldn’t be worth the effort.
- After that I check my performance of that week as well as of the most recent 250 predictions (~2 minutes)
  - For the week itself, I usually only compare the expected value (sum of probabilities) with actually resolved tasks, to check for general over- or underconfidence, as there aren’t enough predictions to evaluate individual percentage ranges
  - For the most recent 250 predictions I check my calibration by having the predictions sorted into probability ranges of 0..9%, 10..19%, … 90..99%.^[3] and checking how much the average outcome ratio of each category deviates from the average of predictions in that range. This is just a quick visual check, which lets me know in which percentage range I tend to be far off.
  - I try to use both these results in order to adjust my predictions for the upcoming week in the next step
- Finally I assign probabilities to all the tasks. I keep this list of predictions hidden from myself throughout the following week in order to minimize the undesired effect of my predictions affecting my behavior (~5 minutes)
  - These predictions are very much System 1 based and any single prediction usually takes no more than a few seconds.
  - I can’t remember how difficult this was when I started this system ~1.5 years ago, but by now coming up with probabilities feels highly natural and I differentiate between things being e.g. 81% likely or 83% likely without the distinction feeling arbitrary.
  - Depending on how striking the results from the evaluation steps were, I slightly adjust the intuitively generated numbers. This also happens intuitively as opposed to following some formal mathematical process.
While this may sound complex when explaining it, I added the time estimates to the list above in order to demonstrate that all of these steps are pretty quick and easy. Spending these 10 minutes^[4] each week seems like a fair price for the benefits it brings.
1. ↩︎
  An example would be “make check up appointment with my dentist”, but when calling during the week realizing the dentist is on vacation and no appointment can be made; given there’s no time pressure and I prefer making an appointment there later to calling a different dentist, the task itself was not achieved, yet my behavior was as desired; as there are arguments to be made to evaluate this both as true or false, I often just drop such cases entirely from my evaluation
2. ↩︎
  I once had the task “sign up for library membership” on my list, but then during the week realized that membership was more expensive than I had thought, and thus decided to drop that goal; here too, you could either argue “the goal is concluded” (no todo remains open at the end of the week) or “I failed the task” (as I didn’t do the formulated action), so I usually ignore those cases instead of evaluating them arbitrarily
3. ↩︎
  One could argue that a 5% and a 95% prediction should really end up in the same bucket, as they entail the same level of certainty; my experience with this particular forecasting domain however is that the symmetry implied by this argument is not necessarily given here. The category of things you’re very likely to do seems highly different in nature from the category of things you’re very unlikely to do. This lack of symmetry can also be observed in the fact that 90% predictions are ~10x more frequent for me in this domain than 10% predictions.
4. ↩︎
  It’s 30 minutes total, but the first 20 are just the planning process itself, whereas the 3+2+5 afterwards are the actual forecasting & calibration training.

markov_user Dec 30, 2020, 12:19 PM
1 point
0 ∶ 0
on: Announcing the Forecasting Innovation Prize
“Before January 1st” in any particular time zone? I’ll probably (85%) publish something within the next ~32h at the time of writing this comment. In case you’re based in e.g. Australia or Asia that might then be January 1st already. Hope that still qualifies. :)