Measuring impact — EA bias towards numbers?

July 26th, 2023

TLDR: Effective Altruism likes numbers, therefore there is bias towards projects that are easily quantifiable. What about big impactful ideas that make intuitive sense but are not that easy to put into a spreadsheet?

I guess we need build a better models, this is part of the game, part of the process.

Cross-posting, cross-linking: The link to this post is now shared on Effective Altruism Forum. Blockchain based platform such as mirror.xyz reduces the probability of “being cancelled” in case someone doesn’t like the opinions / proposals / ideas.

(it’s part of the metacrisis that people are too afraid to say what they really think, silent majority, social desirability bias and the entire multiverse of cancel culture)

Face to face conversations at Funding the Commons in Paris

Featuring twitter.com/TheDevanshMehta and Johan:

Already received some early feedback from a person hosting me at EA house in Paris:

I disagree that EA has that much of a bias towards "easily" quantifiable projects. GiveWell was founded in fact to find ways to quantify and compare the impact of different charities in different domains. This was an incredibly large undertaking over a decade, but now that they have the numerical models they are the foundation of the global health and development side of EA.

On balance, I think the truth is in the middle… The bias is still there that’s why need to get better at modelling, predicting, measuring, evaluating, adding new dimensions of value, accounting for negative externalities and cobenefits.

So here is the big 💡💡💡

Great African Railway - 10000km - 6000 miles - Cape Town to Cairo.

It was in the making 100 years ago:

https://en.wikipedia.org/wiki/Cape_to_Cairo_Railway

It was possible 100 years ago but then Great Depression, WW2, collapse of the British Empire.

Loads of the segments are already completed.

There is plenty of historical resources, research, papers about it.

The bridge at Victoria Falls looks really impressive:

Image used without permission: https://www.britannica.com/place/Victoria-Falls-waterfall-Zambia-Zimbabwe (but we liked it the most)

Random facts:

not just the railway but telegraph, information travels faster than physical atoms
ships and train similar speed = using ferries on Great Rift Lakes of Africa
the bridge is so close to the waterfall to feel the mist

How did you figured out the railway?

Food system. Global hunger. World Food Programme. Food distribution.

Even if you provide the food, what about the distribution network?

It is rather sparse: https://en.wikipedia.org/wiki/African_Union_of_Railways

Africa is bigger than it seems on the map: https://www.visualcapitalist.com/mercator-map-true-size-of-countries/

How do you finance it?

I live in London. I recently visited Paris. Colonial exploitation. Colonial reparations.

It would be morally and politically acceptable, something similar to en.wikipedia.org/wiki/Loss_and_damage

Geopolitical stability?

Obviously. The main concern. As I was analysing the project, a new conflict appeared:

But I wasn’t discouraged, it motivated me even more.

Economic prosperity can end the conflicts.

I do not know much about PMC (private military contractors) but maybe after the Ukraine-Russia peace treaty, the mercenaries from the Wagner Group will be looking for a new job? International contingent of PEACEKEEPERS, checks and balances, making sure the bad guy do not go rogue.

What are the other concerns and 2nd and 3rd order effects?

AI can do a great job in providing insights into the interdependencies.

For me the main thing (“how did you figured out the railway”) is the food distribution.

The risk-reward ratio is there. The cost-benefit analysis there. Assuming money and geopolitical stability is resolved. But geopolitical stability needs to be resolved anyway, and the investment / impact / loss & damage money will come anyway. So may as well do something that brings long term benefits for the entire continent.

How to you measure, evaluate, quantify it?

Definitely seeking help, advice and collaboration in this area.

Relatively simple model of a flower shop: https://www.getguesstimate.com/models/3206

Other examples of evaluating impact

Onboarding 3 billion people not connected to the internet

Reading. Writing. Numbers. English. Or local language, not repeating colonialism.

Personalised AI tutor - education and economic opportunity.

Starlink terminal in each village.

Maybe even a mesh network created by the devices but that would affect the battery life. Maybe some alternative battery preserving tooling: scuttlebutt.nz

https://www.statista.com/statistics/273018/number-of-internet-users-worldwide/

https://www.businessofapps.com/data/android-statistics/

There is enough old iOS and Android devices capable of accessing the internet.

Lessons from One Laptop Per Child:

That was 2006… Completely different reality.

Universal basic water

I was thinking about the cultural aspects of African women going for water. Social aspect of water trips can be replaced by social aspect of education.

Starlink terminal will require some power source, excess power can be transformed into fresh water:

Technology is there. And if you don’t know what’s going on - it’s about the money.

If you could prevent majority of the diseases, what would be the impact?

Slow internet in the UK

I was near the UK parliament during the protests. I thought the slow internet is caused by the large amount of people and government intentionally putting some devices to slow down communication between protesters.

Went for a visit. Internet in the building was non-existent. Another security feature? Just imagine some very influential person trying to pass the contact details.

Same logic applies for trains. Productivity lost of not being able to access internet - surely it has value but non-obvious how to quantify.

Chineese fishing boats

I’m concerned about local people who need take more risk (go further into the ocean) to catch their fish and ensure protein for local population.

If a father fisherman dies, how do you you evaluate impact on a child?

Chineese water management

Certain laws - such as access to water - en.wikipedia.org/wiki/Riparian_water_rights - made sense in California in the 1800s… If you are by the river - your horse can drink as much water as he / she / they / them wants.

These days it is much more complex, see the China Vietnam Mekong situation:

you might consider consulting with experts in these specific fields

What would be even better - interdisciplinary experts who see the bigger picture, who understand the interdependencies, especially these related to human behaviour.

Assembling a working group: impact measurement

These projects do not have instantly obvious business model.

These projects require initial planning and expenses, long before projects are completed.

The impact is not easily quantifiable, spread over time, will need to add “time” as additional dimension of measurement.

Please connect:

BaseX.com - solving the problem of misaligned incentives of the economic system in the context of climate change. We are fixing bugs and getting ready for the release, it’s coming…