Comment author: Peter_Hurford  (EA Profile) 02 April 2017 01:10:42AM 0 points [-]

In looking over your quantitative model on cage-free campaigns, I notice that you conclude cage-free campaigns have a mean direct effect of ~204 QALYs per $.

ACE also has a model of cage-free campaigns (actual inputs are in the bottom-right quadrant) and they conclude cage-free campaigns have a mean direct effect of 65 animals spared per dollar. This isn't in QALYs, but since they're all chickens and egg-laying chickens live an average of 1 year, I assume they mean ~65 QALYs per $.

These two estimates differ by an order of magnitude and I'm unsure who is right. You could both be right, since this is an estimate for MFA in particular, rather than corporate campaigns as a whole, and ACE seems to give less weight toward MFA on these campaigns.

But the more important part is that ACE's estimate also includes a substantial risk of negative impact, which is more concerning. This comes from the fact that ACE thinks there is a substantial probability that the impact to animal welfare could be net negative, perhaps stemming from the discussion between Direct Action Everywhere and OpenPhil.

I was curious what you thought about the differences in your two models, especially with regard to the risk of cage-free being negative?

I do realize this isn't particularly important because you ultimately do not make a cage-free donation, but it seems like food for thought. It's also interesting to evaluate model uncertainty by comparing two independent models.

Comment author: Peter_Hurford  (EA Profile) 01 April 2017 11:55:51PM 1 point [-]

Is there a deadline for filling out the application?

Comment author: nikvetr 01 April 2017 10:10:30PM *  1 point [-]

Ah, gotcha. But re: code review, even the most beautifully constructed chains can fail, and how you specify your model can easily cause things to go kabloom even if the machine's doing everything exactly how it's supposed to. And it only takes a few minutes to drag your log files into something like Tracer and do some basic peace-of-mind checks (and others, e.g. examine bivariate posterior distributions to assess nonidentifiably wrt your demographic params). More sophisticated diagnostics are scattered across a few programs but don't take too long to run either (unless you have e.g. hundreds or thousands of chains, like in marginal likelihood estimation w/ stepping stones... a friend's actually coming out with a program soon -- BONSAI -- that automates a lot of that grunt work, which might be worth looking out for!). :]

(on phone at gym with shit wifi so can't provide links/refs atm, sorry!)

Comment author: Peter_Hurford  (EA Profile) 01 April 2017 10:49:46PM 0 points [-]

Do you have any good textbooks or educational resources to learn these kinds of techniques?

Comment author: nikvetr 01 April 2017 07:56:38PM *  3 points [-]

Yay for Bayesian regression (binomial, I'm guessing? You re-binned your attitude and donations responses? I think an ordered logit would be more appropriate here and result in less of a loss in resolution, or even a dirichlet, but then you'd lose yer ordering)! Those posteriors look decently tight, though I do have some questions!

I'm a little confused on what your control was, exactly. You have both points and distributions in your posterior plots, but you don't have any control paragraph blurb in you google doc questionnaire. How did you evaluate your control? Did you give them a paragraph entirely unrelated to EA? These plots are the posterior estimates for p_binomial when each dummy variable for treatment is 0? Is "average treatment effect" some posterior predictive difference from the control p (i.e. why it's exactly 0)?

On a related (and elucidatory) note, could you more explicitly clarify which models you fitted, exactly? Did you do any model comparison or averaging, or evaluate model adequacy? You mention "controlling for other variables in the survey" but I don't see any e.g. demographic questions in your questionnaire. You said you "examined these relationships overall and among the critical subgroup of those with at least a bachelor’s degree" -- did you do this by excluding everyone without a bachelor's, or by modeling the effects of educational attainment and then doing model comparison to test the legitimacy of those effects (I'd think looking at the posterior for the interaction between your paragraph and education dummies would be the clearest test)? Did you use diffuse, "uninformative" priors (and hyperpriors)? Which ones, exactly?

I assume that since this is a hierarchical analysis you used MCMC (HMC?) to do the fitting. Are your posterior distributions smoothed substantially, e.g. with a kernel density estimator? Or did you just get fantastic performance? What diagnostics did you run to ensure MCMC health? How many chains did you run? Did you use stopping rules? In my experience, hierarchical regression models can be pretty finicky to fit as they get more complex.

Kudos on not just using some wackily inappropriate out-of-the-box frequentist test!

edit: also, what are the boxplot-looking things? 95% HPDIs? CIs? Some other %? Ah wait they're the sd of your marginal samples?

Comment author: Peter_Hurford  (EA Profile) 01 April 2017 08:23:28PM 2 points [-]

On a related (and elucidatory) note, could you more explicitly clarify which models you fitted, exactly?

It would be cool to provide the code, for both learning and verification purposes.

Comment author: Peter_Hurford  (EA Profile) 31 March 2017 07:38:57PM *  3 points [-]

If Clinton had won there would still be a Republican House and Senate

These two outcomes are correlated, so it's unclear to me how true this statement is.

-

If Clinton had won, it’s likely Democrats would have to wait until 2028 until they control the Presidency and Congress

I agree that had Clinton won and had the Republicans still won the House and the Senate, it would likely be awhile until the Democrats would control all three, given the strong trend of midterm loss.

Comment author: kierangreig 12 October 2016 04:29:43PM *  8 points [-]

1) What are the main points of disagreement MIRI has with Open Phil's technical advisors about the importance of Agent Foundations research for reducing risks from AI?

2) Is Sam Harris co-authoring a book with Eliezer on AI Safety? If yes, please provide further details.

3) How many hours do full time MIRI staff work in a usual working week?

4) What’s the biggest mistake MIRI made in the past year?

Comment author: Peter_Hurford  (EA Profile) 30 March 2017 04:31:41PM 0 points [-]
Comment author: Zeke_Sherman 28 March 2017 02:36:30AM 2 points [-]

Pareto Fellowship was shut down? When? What happened?

Comment author: Peter_Hurford  (EA Profile) 28 March 2017 02:52:21AM 6 points [-]

We do not plan to continue the Pareto Fellowship in its current form this year. While we thought that it was a valuable experiment, the cost per participant was too high relative to the magnitude of plan changes made by the fellows. We might consider running a much shorter version of the program, without the project period, in the future. The Pareto Fellowship did, however, make us more excited about doing other high-touch mentoring and training with promising members of the effective altruism community.

From CEA's 2017 Fundraising Report.

Comment author: Tor 27 March 2017 01:00:55AM 3 points [-]

I wanted to dedicate myself to making youtube-videos at some point, but I have another project that I'm prioritizing instead so I'm not spending much time on this at the moment. However, with enough outside help I think that together we might achieve a good impact.

Making videos for existing channels (like e.g. Vox) without expecting payment from them, is a possibility to look into, although convincing them probably would be hard, and the requirements for quality would be challenging. The most likely scenario is publishing to a channel of our own though.

Here is a google-document with some relevant thoughts (ideas and drafts for episodes, etc): https://docs.google.com/document/d/1iVRb85Dkb6e04M0mAkiCJwPg7Y29cjUQboXy-OdCcus/edit#

Video-production is a kind of task where a large fraction of the work can be done while listening to podcasts and such once one has done the necessary learning, but many hours are required. If anyone could be interested in working on this, and could see themselves potentially making this project a significant priority in their lives for several years, then feel very free to search up "Tor Barstad" on Facebook and get in touch for a video-conversation or something :)

In response to comment by Tor on Concrete project lists
Comment author: Peter_Hurford  (EA Profile) 27 March 2017 01:35:58AM 2 points [-]

You and others considering prioritizing this may be interested in Charity Science Outreach's shallow review of content marketing.

Comment author: RyanCarey 26 March 2017 06:33:01AM 4 points [-]

You could do unconditional basic income but why would you start with that when we haven't even created a facility for people to fund credible proposals yet? Seems better to reboot EA Ventures or Impact Certificates first (given that the EA community is a bit bigger, and that some of the reasons for previous failure were related to circumstance).

Comment author: Peter_Hurford  (EA Profile) 26 March 2017 02:48:03PM 16 points [-]

I guess another important next step would be learning from why similar things like EA Ventures, Impact Certificates, and the Pareto Fellowship didn't get more traction and were shut down.

Comment author: Daniel_Eth 26 March 2017 01:56:12AM 7 points [-]

This makes me wish we had basic income - I feel like the need for some income to fulfill basic needs stops people from "taking risks" and pursuing these sorts of projects.

Comment author: Peter_Hurford  (EA Profile) 26 March 2017 04:07:24AM 8 points [-]

I think it could be possible to set up a general sort of EA Fund for this sort of thing, sort of like how there is one for political activism. That could be a missing step on our quest to turn money into talent.

How long do you think someone would have basic income for before they could either "prove" their project and get actual donations / fundraising based on merit or they could go back to a day job? How much funding do you think this would take?

View more: Prev | Next