Comment author: Khorton 17 December 2017 08:36:19PM 0 points [-]

I motion to move all unpaid internships from "jobs" to "projects."

Comment author: Khorton 17 December 2017 12:21:29PM 1 point [-]

What happens if a non-EA wins?

Comment author: Khorton 14 December 2017 10:14:25AM 1 point [-]
Comment author: Khorton 10 December 2017 03:42:43PM 0 points [-]

What's the deal with the stars? What makes a project 1* or 3*?

Comment author: BenHoffman 02 December 2017 09:58:43PM 2 points [-]

We should also expect this to mean that countries such as Australia and China that heavily weight a national exam system when advancing students at crucial stages will have less corrupt educational systems than countries like the US which weight locally assessed factors like grades heavily.

(Of course, there can be massive downsides to standardization as well.)

Comment author: Khorton 03 December 2017 02:07:32AM 0 points [-]

I'd find this pretty surprising based on my knowledge of the Canadian (Albertan) & British education systems. Does anyone have evidence for standardized exams decreasing "corruption"? (Ben, I'm not sure exactly what you meant by corruption here - do you mean grades that don't match ability, or lazy teaching, or something else?)

Comment author: ateabug 31 October 2017 11:48:22PM *  3 points [-]

I'm a Canadian currently studying public policy in London. I'm planning to write my dissertation on AI policy and gender, so naturally I'm fascinated by your organization.

Out of curiosity, what is the connection between AI policy and gender you're looking at?

Comment author: Khorton 01 November 2017 12:35:00PM 1 point [-]

I'm tentatively planning to look at the government's role in regards to AI that is discriminating or perceived to be discriminating based on sex. For example, if an AI system was only short-listing men for top jobs, should the government respond with regulation, make it easier for offended parties to challenge in court, try to provide incentives for the company to improve their technology, or something else entirely?

I just started my MA a month ago, though, and won't be seriously focusing on my dissertation until May, so I will have a much better idea in six months. :)

Comment author: WyattTessari 31 October 2017 06:20:37PM *  5 points [-]

Hi Dony,

Great questions! My name is Wyatt Tessari and I am the founder.

1) We are doing that right now. Consultations is a top priority for us before we start our advocacy efforts. It's also part of the reason we're reaching out here.

2) Our main comparative advantage is that (to the best of our research) there is no one else in the political/advocacy sphere openly talking about the issue in Canada. If there are better organisations than us, where are they? We'd gladly join or collaborate with them.

3) There are plenty of risks - causing fear or misunderstanding, getting hijacked by personalities or adjacent causes, causing backlash or counterproductive behaviour - but the reality is they exist anyway. The general public will eventually clue in to the stakes around ASI and AI safety and the best we can do is get in early in the debate, frame it as constructively as possible, and provide people with tools (petitions, campaigns) that will be an effective outlet for their concerns.

4) This is a tough question. There would likely be a number of metrics - feedback from AI & governance experts, popular support (or lack thereof), and a healthy dose of ongoing critical thought. But if you (or anyone else reading this) has better ideas we'd love to hear them.

In any case, thanks again for your questions and we'd love to hear more (that's how we're hoping to grow...).

Comment author: Khorton 31 October 2017 11:02:42PM 5 points [-]

Hi Wyatt,

I'm a Canadian currently studying public policy in London. I'm planning to write my dissertation on AI policy and gender, so naturally I'm fascinated by your organization.

The topics you're planning to discuss, especially the risk of a general artificial intelligence, seem quite sensitive. You didn't say a lot about your background. What relevant experience does your team have at handling sensitive issues or framing political debates? (I mean in your day jobs; I know the nonprofit is new.)

Kirsten

Comment author: zdgroff 30 October 2017 06:57:12PM 3 points [-]

Thanks for writing this up! I learned some things here even having managed press previously for Direct Action Everywhere.

One thing I'd note that might help with some bruised emails (and let me know if you disagree) is that this business is highly random. Getting an op-ed published depends entirely on the judgments of a small number of people.

Also, one comment:

Pitch the same piece to multiple outlets at once

This is definitely standard press advice, but I'm actually curious if this is wise. Given that (I think) it's more often than not the case that the chance of getting published is quite low, if a piece is time-sensitive, doesn't this dramatically lower your chances of getting published (many places don't notify of rejections), while if you submit to multiple, the most that happens is you burn someone who might not have been a good contact anyway? I've been wondering if the common wisdom is wrong here.

Comment author: Khorton 30 October 2017 09:34:34PM *  2 points [-]

If you really want to publicize something time-sensitive, perhaps you could pitch to multiple publications with query letters personalized to each? You could end up writing 2+ articles or op-eds on the same topic (open letter about factory farming) but with different angles and tones (focus on famous people who are concerned vs focus on animal rights vs focus on risks to humans).

I've seen this option suggested online eg here: https://www.theadventurouswriter.com/blogwriting/multiple-query-letters-magazines/

[Edited for clarity.]

Comment author: zdgroff 27 October 2017 06:36:31PM *  7 points [-]

Your portrait of what the EA community could be is a beautiful one and made me tear up. You hit the nail on the head many times in this post on the subtle connections between things that I think can be hard to identify: the connection between heart and head, the E and the A, the overuse of jargon, and the hero worship, and so on. I have to say that as a fairly straight-passing gay man with immense amounts of privilege, even I feel many of these pressures and am often put off by the alpha-male machismo you often see in EA spaces.

I’ve witnessed discrimination and harassment, and heard of assault, in EA-ish spaces, and it seems pretty clear that this is contributing to the gender gap. I’ve definitely exhibited some of the combative and argumentative behaviors you mention. When I got into the EA community a few years ago, I began in global poverty and animal advocacy circles, and I found they were much better on these issues than the community is now, sadly. (That’s with both of those areas’ having plenty of problems.)

I think Kelly moved us toward a type of dialogue on this issue that is lacking in the world, and I hope we can have more of it. Right now, discussions around diversity and inclusion seem polarized between the sort of “rationalist” discussion that’s snarky and dismissive on the one hand and an ostracizing mob mentality on the other hand. I don’t want to say EA should chart a middle path, because I think we should lean toward being overly zealous on diversity and inclusion rather than away, but I think EA and its aligned movements (animal advocacy in my mind) would benefit from a conversation that is at the same time inclusive and data-based. I don’t think the world has that type of conversation very often.

The lack of conversations that are both inclusive and data-based seems to lead to pretty bad results, where diversity and inclusion are may not be promoted in the most effective ways, and people opposed to diversity and inclusion harbor suspicions about the world (e.g. that discrimination does not exist) that continue to fester unaddressed.

From my exploration of these matters, I’ve come to see that generally, when one reads about data on discrimination, differences between groups, etc. one finds that (a) discrimination exists and can be quite powerful; (b) there are differences between genders, but the differences are subtle and go in varied directions (e.g. men are more combative, and women are more collaborative, as Kelly notes); and (c ) these differences are not the reason for the vast majority of gaps that we see.

I think that because discussion about differences between genders is often consigned to the more diversity-hostile corners of the internet, though, ideas that would be proven wrong by the data go unchallenged. Again, I think if we were to have the right sort of conversation on these issues, we would find that discrimination is indeed the primary cause of the gender gap in EA, but without that conversation, people will not be convinced. (And if an honest conversation engaged with data and personal experiences came to the conclusion that this was not the case, that would probably be good information to have.)

For instance, I read the Damore memo, but then saw this graph which seems to be pretty good evidence that the vast majority of the gap in tech is not from biological differences (and so likely some iteration of discrimination, implicit or explicit). I don’t remember where I came across this graph, but it was very helpful to me. Without looking at the whole picture, though, one can look solely at the individual components of the picture (e.g. Damore’s arguments on specific gender differences) and come to conclusions that would be put in doubt with fuller information.

As an additional reason why I think EA is a movement that could have the right conversation on this, I think that EAs recognize a moral principle similar to equality of interests, where differences in personal traits do not lead to moral differences. It seems that in many diversity and inclusion conversations, both the right and the left consider personal trait differences to imply moral differences, and I think EAs can challenge and move beyond that assumption–though with care and only after we start improving on our demographics.

This is a very challenging issue because, as noted in a comment below, racism and sexism have long been motivated by biological essentialism, and it’s extremely disturbing to have people talk about a group you are a part of in this way. (As a Jew, I can say that I feel discomfort with the conversation about Jewish values below, for instance, though I don’t have a strong opinion on its propriety.) I think that the way to deal with this problem is to exercise caution when speaking about these sorts of things, to avoid casual discussion of them, and to have a higher evidence standard for when we talk about these things. I think that our community can learn the appropriate maturity to do that, though.

Anyway, all this is to say that I hope that as this conversation goes on, we can bring data to bear and recognize the implications of the way we speak for others in this community. Words and ideas do cause harm, and we should be utilitarians about the way we speak. With appropriate caution, though, I think that EAs can have a conversation that gets to the heart of the matter and offers a model for how these conversations can be had.

————————————

For those looking for examples of places where these discussions could be valuable, I have a few:

  • Gender and cosmopolitan values–The Better Angels of Our Nature cites feminism as one of the reasons for declines in all sorts of violence (war, sexual violence, torture), and I’ve seen enough data to match my intuition that feminism is also very good for animals. I think there are lots of things to explore empirically in this domain (that likely would have implications for the A vs. E debate), but they probably involve engaging with uncomfortable questions about where these gender differences arise.

  • On another note, animal advocates will often assert that if we focus on multiple causes, we will solve our diversity and inclusion problem. I think this is a very important claim to test, because focusing on multiple causes may be quite costly. I’m fully supportive of focusing on creating justice within our movements and groups, e.g. by aggressively fighting sexual assault and getting rid of income barriers, but I think the claim about movements’ outward focus is a debatable one that really needs to be empirically explored.

  • Similarly to the above note, animal advocates often work on issues to promote diversity and inclusion including things like fighting urban food deserts without looking into the evidence around them. This could not only hinder direct impacts but also create the impression that advocates’ diversity and inclusion efforts are an afterthought without the same rigor applied to it that advocates apply to work for animals.

Comment author: Khorton 27 October 2017 07:50:04PM 1 point [-]

"fighting urban food deserts without looking into the." I think there's a word or phrase missing.

Comment author: Khorton 26 October 2017 10:50:29PM 9 points [-]

I'm seeing a lot of comments questioning the literature around diversity improving performance. EA prizes accuracy, so that's a good thing.

However, I'm concerned we're falling into two very common traps of requiring women to prove themselves more competent than men and status quo bias.

In general, I'd expect teams to be diverse unless a non-diverse team can be proven more effective. Because so many EA leaders are currently white men, I can imagine some reasons why we might have less-diverse teams in the short-term, but my baseline expectation would always be to prefer a more diverse team all other things being equal.

View more: Next