This is a meetup report - let me know if you want me to stop posting these. It's intended to give a very rough flavour of what was discussed, not as accurate or complete minutes.

Attendance: 4 (snowy weather may have reduced attendance somewhat)

This was the first Toronto EA meetup for over a year. I'm no longer branding it as a "THINK" meeting, but I'm still reusing the THINK meetup group. It was held in my apartment.

We started the discussion on the effectiveness of Oxfam.

• GiveWell are decidedly lukewarm about them
• The Life You Can Save like them
• They have goals that are hard to measure, e.g. education, lobbying

Oxfam certainly evaluate the impact of some of their programs.

But there's a contrast between this and a more data-driven approach, e.g. GiveDirectly. I described a data driven approach as:

• Scientific studies measure the relationship between some intervention and some outcome we actually care about (e.g. lifespan, health outcomes)
• Orgs measure something that's easier to measure, e.g. counting the number of interventions
• From this and the original study we can estimate the impact the org is having

Data driven approach:

• Advantages: Reduces amount of work needed for donors to decide?
• Disadvantages: Paucity of data, Most people don't have skillset to apply data driven approach
• (we came up with some others as well but I didn't record them)

The issue came up: How do I know which authorities to listen to? (e.g. GiveWell versus other charity evaluators)

Hypothesis: bimodal distribution, with small donors doing hardly any research and large donors paying someone to do research for them

Reasons for corporate donation matching:

• tax reasons (if people are givers then matching their donations is a way of essentially paying them more without paying extra tax)
• makes company look good
• recruitment?

A problem that the sort of people working for orgs are unlikely to be good at a data-driven approach. It's a small percentage of the population to start with, and orgs are likely made up of:

• people who care about the project
• people who know about business and managing donors
• specialists with knowledge about projects (these are the most likely to be good at data).

Why would priorities of large donors conflict with chasing effectiveness?

• relationship development
• whatever's in vogue (e.g. viewing organisations as businesses)

We talked about Oxfam and MIRI (I never thought I'd be lumping those two together!) as examples of orgs working on something speculative with hard-to-measure impact. Brian Tomasik was mentioned as someone who's given reasons for giving to MIRI.

Joseph talked about his Neural net to predict earthquakes

This of course got us on to discussing the causes of publication bias:

• journals not accepting it
• corporate bias
• desk drawer effect

EA is all about opportunity cost - if you're giving to SCI then that's money you're not giving to Oxfam etc.

We talked about utility:

• This article came up (on comparing different ways of helping)
• QALYs only capture only a small component of what we actually care about
• What would a virtue ethics or deontological EA look like?
• Utilitarianism is deontology with one rule.

We talked about the money that we're spending on things other than charity.

• Is economic significance of the 90% of income that I'm spending on myself greater than the 10% that I'm spending on charity?
• In particular, fair trade and other ethically sourced suppliers such as Mountain Equipment Co-op
• We weren't sure if fair trade was effective, but didn't get around to digging up any research pointing one way or the other
• Opportunity cost - money spent on buying ethical stuff could have come out of charity budget
• Etheopian coffee - highly priced on international market, but farmers are paid very little. Market inefficiencies - high barrier to entry. If market was more efficient, price would go down at consumer end so demand would go up, benefitting the farmers. Import tarriffs also get in the way.
• Because of supply/demand, a more efficient market should benefit both consumers and farmers
• If things cost the same, utilitarianism says give to the fairer one

Don't try to choose the best one - optimisation is intractable.

• Choose "better" rather than "best".
• Don't condemn people who aren't doing the best they can - applaud changes in the right direction

Low hanging fruit in purchasing:

• don't buy a load of useless crap (living frugally)
• cheapest long term, including reasonable quality stuff second hand

Animal suffering

• animal has experience?
• can't measure degree of experience but we can measure complexity of behaviour
• is consciousness on a gradient?