Tobias Dänzer

252 karmaJoined Jun 2021


Topic contributions

What exactly does the "Request for Feedback" button do when writing a post? I began a post, clicked the aforementioned button, and my post got saved as a draft, with no other visible feedback as to what was happening, or whether the request was successful, or what would happen next.

Also, I kind of expected that I'd be able to mention what kind of feedback I'm actually interested in; a generic request for feedback is unlikely to give me the kind of feedback I'm interested in, after all. Is the idea here to add a draft section with details re: the request for feedback, or what?

Let me justify my complete disagreement.

I read your comment as applying insanely high quality requirements to what's already an absolutely thankless task. The result of applying your standards would be that the OP would not get written. In a world where criticism is too expensive, it won't get produced. This is good if the criticism is substance-less, but bad if it's of substance.

Also, professional journalists are paid for their work. In case of posts like these, who is supposed to pay the wages and provide the manpower to fulfill requirements like "running it by legal"? Are we going to ask all EA organisations to pay into a whistleblower fund, or what?

Also, for many standards and codes of ethics, their main purpose is not to provide a public good, or to improve epistemics, but to protect the professionals themselves. (For example, I sure wish doctors would tell patients if any of their colleagues should be avoided, but this is just not done.) So unequivocally adhering to such professional standards is not the right goal to strive for.

I also read your comment as containing a bunch of leading questions that presupposed a negative conclusion. Over eight paragraphs of questions, you're questioning the author and his sources, but the only time you question the source of the investigation is when it puts them in a positive light. Thus I found the following phrasing disingenious: "I don't know the answers to some of these, although I have personal hunches about others. But that's not what's important here."

Overall, I would be more sympathetic towards your perspective if the EA Forum was drowning in this kind of, as you call it, amateur investigative journalism. But I don't think we suffer from an oversupply. To the contrary, we could've used a lot more of that before FTX blew up.

Finally, instead of the decision-making algorithm of judging by the standards of professional investigative journalism, I suggest an alternative algorithm more like "does this standard make outcomes like FTX more or less likely". I think your suggestion makes it more likely.

This seems like a great project, so thanks a lot for doing this! <3

Suggestions & feedback

  • Begin future summary posts with a short but huge standard disclaimer of the main shortcomings of (these / all) AI summaries: can't guarantee accuracy; the emphasis might be wrong; middle sections of long posts are skipped for now; etc.
  • Besides the author byline, add additional info pulled from the original posts: like length or reading time; or amount of discussion (none / little / lots)
    • (Since you already have the author bylines, you could link the author names to their user profiles. Though that might just result in pointless clutter.)
    • (You could even flag posts which feature disproportionate amounts of scholarship, by e.g. identifying posts which link to a lot of academic papers, or which have lots of footnotes. But that has unclear benefits and feasibility, and is likely very out of scope.)
  • Currently summaries are grouped by EA Forum vs. LessWrong, but that seems like insufficient structure given the ~16 posts in each section. And I don't know how posts are sorted right now. So a third level of hierarchy seems warranted. Maybe these posts could be automatically grouped or sorted into topics like "AI" or "Longtermism" or something?
    • I don't know where the topics and the categorization would come from, though. Maybe from looking at the post tags? Unfortunately posts are tagged by hand and thus tags are inconsistently applied.
  • Spelling: "Lesswrong" section header -> "LessWrong"
  • There were some issues with the bylines, which I reported in this comment thread.
  • Summary limitations: Detect when a post contains mostly content which can't be properly summarized: probably images; probably some types of embeds (Manifold? Twitter?); maybe also other stuff like tables or LaTeX math formulas.

Regarding summary tone & format

  • In addition to the current output as a post on the EA Forum / LW, you could also make a spreadsheet of summaries (with columns for e.g. week, title, link, authors, summary paragraph), add new summaries to this spreadsheet, and then link to the spreadsheet at the beginning of each post. I'm not sure how useful that would be, but it would allow for e.g. follow-up automations; quarterly / yearly summaries of these weekly summary posts; etc.
  • Regarding the formatting of the summaries, one question to ask is "Who are these summaries for?" or "What value do summaries add?" or "What questions do summaries answer?".
    • E.g. a crucial one is "Should I read this?" / "Who would benefit from reading this?". Currently these summaries answer that question implicitly, but you could instead / in addition ask ChatGPT to answer this question explicitly.
  • Regarding the summary tone, I currently find the summaries a bit dry, so I don't find them particularly enjoyable to read. But I'm not sure if there's a tone which is less formal without being less accurate.

Opportunities for collaboration

  • Maybe you could collaborate with the people from Type III Audio in some way? They generate automatic audio transcripts for new posts while you generate automatic text summaries; so there might be some potential for cross-pollination.
  • I vaguely recall there also being some efforts to auto-tag posts via LLMs; that's another avenue for collaboration.

Another username issue: If there are multiple authors, they're ignored here, e.g. for "Impact obsession", the attribution for Ewelina_Tur is missing. Same with "CE alert", "Dimensions of pain", etc.

I've noticed one more username issue: The listed name is the one in the URL (e.g. jessica-liu-taylor) instead of the username (e.g. jessicata). Or e.g. Bill Benzon becomes bill-benzon. It's not clear to me why e.g. Raemon doesn't become raemon, though.

For context, here's Matt Yglesias on the state of UK housing policy:

The nature of the UK housing issue should be familiar to Americans, especially those who read Slow Boring. There is a lot of demand for living in Greater London (and also Oxford, which isn’t far away), and that demand is not met, which leads to high prices.

But the extent of the problem can be hard for an American to grasp.

For example, as best I can tell, the average size of a home in the United Kingdom is 1033 square feet versus 968 square feet in New York City, so it’s not that Greater London is experiencing a housing squeeze comparable to Greater New York City — the UK as a whole is experiencing a housing squeeze similar to that of NYC. London is worse, and not coincidentally, the UK doesn’t have an equivalent of Dallas or Phoenix or Atlanta — a big, cheap city that is growing fast.

Worst of all, as Anthony Breach from the Centre for Cities points out, English homes are getting smaller over time because the market is so squeezed.

I posted for the ~first time in the EA forum after the SBF stuff, and was pretty disappointed by the voting patterns: almost all critical posts get highly upvoted (well, taking into account the selection effect where I wouldn't see negative-karma posts), seemingly regardless of how valid or truthseeking or actionable they are. And then the high-karma comments very often just consist of praise for writing up that criticism, or ones that take the criticism for granted and expand on it, while criticism of criticism gets few upvotes.

(Anyway, after observing the voting patterns on this comment thread of mine, I see little reason with engaging on this forum anymore. I find the voting patterns on LW healthier.)

Where do the multiple "inzuence" typos come from? A weird effect of copy&pasting? (I've seen this stuff in pdfs when "fl" was displayed as a separate character.) A typo in the original that was fixed there but not here?

IIRC commenters disputed whether / to which degree MIRI's secrecy & infohazards policy was in any way worse than typical NDAs for big companies.

IIRC re: Michael Vassar, the problem was not so much the connection to him, but that several people around Vassar (the author included) had experienced drug-induced psychosis, which made their criticisms and reported experiences suspect. My sense of the post was that it described innocuous facts and then considered them to be bad by analogy to e.g. Leverage.

Re: mental health, I agree that the MIRI world view is likely not good for one's mental health, but I wouldn't consider that a "ground fact" about MIRI, but rather (assuming one buys into that worldview) a problem with the world being the way it is. For instance, it sure would be better for everyone's mental health if AI alignment were universally agreed upon to be trivially easy, but unfortunately that's not the case.

based on what I see to be a fairly superficial reading of karma/comment count.

I read a significant fraction of the comments in that thread when it first appeared (though not all of them). I'm stressing those data points so much because that thread is still getting cited to this day as if it's undisputed, legitimate and broadly community-endorsed criticism, merely because it has positive karma. Hence I think stressing how to interpret the karma score and number of comments is a very important point, not a superficial one.

To their credit, the EA and LW communities love to question and criticize themselves, and to upvote all criticism. Unfortunately, that lends credence to weak or epistemically dubious criticisms far beyond what would be merited.

Load more