All of AM's Comments + Replies

Book Review: Free Will

He doesn't go into this in the book, but I am fairly sure that Harris would agree with your consequentialist take of "acting as if they had free will". I have heard him speak on this matter in a few of his podcasts around "the hard problem of consciousness" with Dennett, Chalmers and a neurosurgeon that I can't find the name of (I remember him being british). 

As I understand him, his view is to not view criminals (or anyone) as "morally bad" for whatever they have done, but to move directly on to figuring out the best possible way to avoid bad things ... (read more)

2JBlack2moSo long as it doesn't increase the number of people committing crimes once to gain access to cushy luxurious rehabilitation programs, sure! In my personal life, I do notice a tendency that when I do categorize people as being "morally wrong", this is more normally associated with wanting them to become morally right than wanting them punished for being morally wrong. The latter seems just one (probably ineffective) way to achieve the former. I don't seem to see this tendency in many other people around me, so I suspect I'm in a minority. I don't think this relies on any particular position in the free will discussion though. I've seen some people "punish" objects for adversely affecting them by yelling or striking at them, and certainly not for reasons of the object having free will. It seems more of an innate urge than any philosophical belief.
£2000 bounty - contraceptives (and UTI) literature review

Helpful resource for whoever ends up doing this: Contraceptive Technology. It's a huge book that summarises almost all effectiveness studies that have been done on contraceptives, including the definitions of perfect and typical use (very important when comparing contraceptives). It also has detailed summaries of side effects, medical interactions, description of method of action and well researched "advantages" and "disadvantages" sections — it's basically what doctors use to decide how to prescribe birth control. 

Source: I have used this book myself in research, I work for a birth control app company.

Review of "Why AI is Harder Than We Think"

Good points! 
Yes this snippet is particularly nonsensical to me

an AI system could be“superintelligent” without any basic humanlike common sense, yet while seamlessly preserving the speed, precision and programmability of a computer

It sounds like their experience with computers has involved them having a lot of "basic humanlike common sense" which is a pretty crazy experience in this case. When I explain what programming is like to kids, I usually say something like "The computer will do exactly exactly exactly what you tell it to, extremely fast. You ... (read more)

Review of "Why AI is Harder Than We Think"

Great and fair critique of this paper! I also enjoyed reading it and would recommend it also just for the history write up.

What do you think is the underlying reason for the bad reasoning in fallacy 4? Is the orthogonal it thesis particularly hard to understand intuitively or has it been covered so badly by media so often that the broad consensus of what it means is now wrong?

3electroswing7moHmmm...the orthogonality thesis is pretty simple to state, so I don't think necessarily that it has been grossly misunderstood. The bad reasoning in Fallacy 4 seems to come from a more general phenomenon with classic AI Safety arguments, where they do hold up, but only with some caveats and/or more precise phrasing. So I guess "bad coverage" could apply to the extent that popular sources don't go in depth enough. I do think the author presented good summaries of Bostrom's and Russell's viewpoints. But then they immediately jump to a "special sauce" type argument. (Quoting the full thing just in case) I really don't understand where the author is coming from with this. I will admit that the classic paperclip maximizer example is pretty far-fetched, and maybe not the best way to explain the orthogonality thesis to a skeptic. I prefer more down-to-earth examples like, say, a chess bot with plenty of compute to look ahead, but its goal is to protect its pawns at all costs instead of its king. It will pursue its goal intelligently but the goal is silly to us, if what we want is for it to be a good chess player. I feel like the author's counterargument would make more sense if they framed it as an outer alignment objection like "it's exceedingly difficult to make an AI whose goal is to maximize paperclips unboundedly, with no other human values baked in, because the training data is made by humans". And maybe this is also what their intuition was, and they just picked on the orthogonality thesis since it's connected to the paperclip maximize example and easy to state. Hard to tell. It would be nice if AI Safety were less disorganized, and had a textbook or something. Then, a researcher would have a hard time learning about the orthogonality thesis without also hearing a refutation of this common objection. But a textbook seems a long way away...
Menstrual cycle effects—Clue study summary and commentary

Great write-up! I work at another femtech company and will share this with some of my colleagues. 

Three thoughts/comments:
1. Clue is very focused on period dates (being a period tracker) and doesn't have very accurate ovulation data. Therefore, cyclical changes in metrics that are affected by ovulation rather than by period will look dimmer due to differences between the length of peoples' Luteal phase.  E.g. the BBT signal would probably be much sharper if the x axis was "days relative to ovulation" rather than "days relative to period", since t... (read more)

How do you optimize productivity with respect to your menstrual cycle?
Answer by AMFeb 09, 20213

I don't menstruate, but I work at Natural Cycles (a data-driven birth control app) with data science, and look for these types of patterns a lot — our users are not using Hormonal Birth Control though, so the sample is biased in that way. 

Clue (a popular period tracker app) recently released one of the best studies I've seen on the mood changes over the menstrual cycle, but unfortunately it is not open access. The authors shared a pdf with me after I emailed them though, DM me if you would like a copy (sci-hub doesn't seem to have the paper yet), or email the authors directly, they are very helpful. 

Anti-Aging: State of the Art

True, circulatory diseases would be a big win, but do you think the marginal buck there is likely to do as much as a marginal buck focused on aging giving the amount of funding allocated to each? If we add the R&D budgets focused on circulatory diseases to the treatment cost of circulatory diseases (potential profit pool for pharma companies), my intuition says that the number would be ~20-100x the total amount of funding to aging-stopping or -reversing technology. What do you think the ratio would be?

Anti-Aging: State of the Art

Definitely a bug! It was my first and only foray into D3.js so there are a lot of bad states you can get into fairly easily. Might rebuild it in something else one day.

3JackH1yI think it would be worth rebuilding if you have time. If you do, make sure to share it on Longevity Subreddit [https://www.reddit.com/r/longevity/]. You will get a lot of interest in it there.
Anti-Aging: State of the Art

Love this article.

After reading the The Fable of the Dragon-Tyrant a few years ago after my father died, I went into a deep dive on this and ended up making a calculator, comparing the impact of eliminating various causes of death on average / median lifespan. It's very simplistic, but I found it interesting to use to illustrate how ageing contributes to death:

http://life.analogmantra.com/ 

2jmh1yVery interesting. Assuming we eliminated everything but accidental causes looks like we should live to about 120+ years. I think Sinclair had said that was the expected lifespan as well. Taking the tool at face value, it seem that both personally and socially effort focused on circulatory diseases should give the biggest bang for the buck. Then again I didn't run through different cases of combination so...
2JackH1yReally love the app, great work! Just a bug I found (I think it's a bug?) - if I untick all the boxes, the median age of death goes to 0.