I deeply appreciate how you feel about EA becoming a self-perpetuating but misaligned engine. It’s much stronger writing than what I’ve told people (usually on discord) when they bring up EA movement building as a cause area.
I think more can be said about TMM. One angle is patience, the idea that we can think of EA institutions as being an order or several of magnitude more wealthy in the future instead of thinking of them as we currently think of them. Combine this insight with some moderate credence in we are not in the hinge of history, you could turn the hold option from the S-process into a full-throated choice to invest. This way, the quality bar (the difficult search for fund-worthy people/projects) wouldn’t be lowered by any of the TMM pressures.
However. There are some of the considerations about institutions becoming self-perpetuating and misaligned, as a special case of social scenes or movements becoming self-perpetuating and misaligned. It would be really difficult to reconcile patience with your pointing out of the fact that we don’t know how to Do Actual Thing, as a society we might be getting worse at it, and as a movement EAs aren’t exceptional. As an intuition pump, imagine a researcher trying to figure out when, why, and how harvard arrived at Not Doing Actual Thing: clearly the size of the endowment isn’t the key variable of the expected impact of harvard relative to a counterfactual universe in which harvard was Doing Actual Thing.
My vision is for somebody to found an org for one 6-18 month project and promise to shutter the org immediately after. Combine this with stellar impact metrics and I think you have a recipe for EA busting out of it’s Not Doing Actual Thing shackles (the impact metrics part is the hard part).
My vision is for somebody to found an org for one 6-18 month project and promise to shutter the org immediately after. Combine this with stellar impact metrics and I think you have a recipe for EA busting out of it’s Not Doing Actual Thing shackles (the impact metrics part is the hard part).
If you shut down the org after 18 months, how will you evaluating them based on impact metrics in any meaningful way?
I was probably assuming third party evaluator. I think the individuals should be free to do another project while they wait for the metrics to kick in / the numbers to come back. I think if the metrics come back and it turns out they had done a great job, then they should gain social capital to spend on their future projects, and maybe return to a project similar to the one they shuttered in the future.
You’re right that this is a problem if the metrics are expected to be done in house!
Metrics are everywhere and always a problem. If the project doesn’t continue and the metrics are used to judge the person’s performance, it’s even more of a Goodhart issue, so I’d be very cautious about judging via known metrics, unless a given situation provides a very good fit.
I deeply appreciate how you feel about EA becoming a self-perpetuating but misaligned engine. It’s much stronger writing than what I’ve told people (usually on discord) when they bring up EA movement building as a cause area.
I think more can be said about TMM. One angle is patience, the idea that we can think of EA institutions as being an order or several of magnitude more wealthy in the future instead of thinking of them as we currently think of them. Combine this insight with some moderate credence in we are not in the hinge of history, you could turn the
holdoption from the S-process into a full-throated choice to invest. This way, the quality bar (the difficult search for fund-worthy people/projects) wouldn’t be lowered by any of the TMM pressures.However. There are some of the considerations about institutions becoming self-perpetuating and misaligned, as a special case of social scenes or movements becoming self-perpetuating and misaligned. It would be really difficult to reconcile patience with your pointing out of the fact that we don’t know how to Do Actual Thing, as a society we might be getting worse at it, and as a movement EAs aren’t exceptional. As an intuition pump, imagine a researcher trying to figure out when, why, and how harvard arrived at Not Doing Actual Thing: clearly the size of the endowment isn’t the key variable of the expected impact of harvard relative to a counterfactual universe in which harvard was Doing Actual Thing.
My vision is for somebody to found an org for one 6-18 month project and promise to shutter the org immediately after. Combine this with stellar impact metrics and I think you have a recipe for EA busting out of it’s Not Doing Actual Thing shackles (the impact metrics part is the hard part).
If you shut down the org after 18 months, how will you evaluating them based on impact metrics in any meaningful way?
I was probably assuming third party evaluator. I think the individuals should be free to do another project while they wait for the metrics to kick in / the numbers to come back. I think if the metrics come back and it turns out they had done a great job, then they should gain social capital to spend on their future projects, and maybe return to a project similar to the one they shuttered in the future.
You’re right that this is a problem if the metrics are expected to be done in house!
Metrics are everywhere and always a problem. If the project doesn’t continue and the metrics are used to judge the person’s performance, it’s even more of a Goodhart issue, so I’d be very cautious about judging via known metrics, unless a given situation provides a very good fit.