Sean is one of the under-sung heroes who helped build FHI and kept it alive. He did this by--among other things--careful and difficult relationship management with the faculty. I had to engage in this work too and it was less like being between a rock and a hard place and more like being between a belt grinder and another bigger belt grinder.
One can disagree about apportioning the blame for this relationship--and in my mind, I divide it differently than Sean--but after his four years of first-hand experience, my response to Sean is to take his view seriously, listen, and consider it. (And to give it weight even against my 3.5 years of first-hand experience!)
As a tangent, respectfully listening to people's views and expressing gratitude--and avoiding unnecessary blame--was a core part of what allowed ops and admin staff to keep FHI alive for so long against hostile social dynamics. As per Anders' comment posted by Pablo here, it might be useful for extending EA's productive legacy as well.
I want to take this opportunity to thank the people who kept FHI alive for so many years against such hurricane-force headwinds. But I also want to express some concerns, warnings, and--honestly--mixed feelings about what that entailed.
Today, a huge amount of FHI's work is being carried forward by dozens of excellent organizations and literally thousands of brilliant individuals. FHI's mission has replicated and spread and diversified. It is safe now. However, there was a time when FHI was mostly alone and the ember might have died from the shockingly harsh winds of Oxford before it could light these thousands of other fires.
I have mixed feelings about encouraging the veneration of FHI ops people because they made sacrifices that later had terrible consequences for their physical and mental health, family lives, and sometimes careers--and I want to discourage others from making these trade-offs in the future. At the same time, their willingness to sacrifice so much, quietly and in the background, because of their sincere belief in FHI's mission--and this sacrifice paying off with keeping FHI alive long enough for its work to spread--is something for which I am incredibly grateful.
A small selection from the report:
Bostrom has stated: “I wish it were possible to convey the heroic efforts of our core administrative team that were required to keep the FHI organizational apparatus semi-performant and dynamic for all those years until its final demise! It is an important part of the story. And the discrepancy between the caliber of our people and the typical university administrators - like Andrew carpet bombing his intray with pomodoros over the weekends... or Tanya putting in literal 21 or 22 hour workdays (!) for weeks at an end. Probably not even our own researchers fully appreciate what went on behind the scenes.”
21 and 22 hour workdays sounds like hyperbole, but I was there and it isn't. No one should work this hard. And it was not free. Yet, if you ever meet Tanya Singh, please know you are meeting a (foolishly self-sacrificing?) hero.
And while Andrew Snyder-Beattie is widely and accurately known as a productivity robot, transforming into a robot--leaving aside the fairytales of the cult of productivity--requires inflicting an enormous amount of deprivation on your human needs.
But why did this even happen? An example from the report:
One of our administrators developed a joke measurement unit, “the Oxford”. 1 Oxford is the amount of work it takes to read and write 308 emails. This is the actual administrative effort it took for FHI to have a small grant disbursed into its account within the Philosophy Faculty so that we could start using it - after both the funder and the University had already approved the grant.)
This again sounds like hyperbole. It again is not. This was me. After a small grant was awarded and accepted by the university, it took me 308 emails to get this "completed" grant into our account.
FHI died because Oxford killed it. But it was not a quick death. It was a years-long struggle with incredible heroism and terrible casualties. But what a legacy. Thank you sincerely to all of the ops people who made it possible.
Christiano and Amodei are both at OpenAI. Jan Leike, Shane Legg, and Miljan Martic are all at DeepMind. (Jan Leike is also former FHI and is still a research associate with us :) ).
I wrote this in a google doc and copy-pasted, without intending the numbers to be links to anything. I'm not really sure why it made them highlight like a hyperlink.
Sean is one of the under-sung heroes who helped build FHI and kept it alive. He did this by--among other things--careful and difficult relationship management with the faculty. I had to engage in this work too and it was less like being between a rock and a hard place and more like being between a belt grinder and another bigger belt grinder.
One can disagree about apportioning the blame for this relationship--and in my mind, I divide it differently than Sean--but after his four years of first-hand experience, my response to Sean is to take his view seriously, listen, and consider it. (And to give it weight even against my 3.5 years of first-hand experience!)
As a tangent, respectfully listening to people's views and expressing gratitude--and avoiding unnecessary blame--was a core part of what allowed ops and admin staff to keep FHI alive for so long against hostile social dynamics. As per Anders' comment posted by Pablo here, it might be useful for extending EA's productive legacy as well.
Sean thank you so much for all you did for FHI.
The Shrimp of Humanity Institute shut down two days ago :(
Fortunately, its legacy lives on in the dozens of other longshrimpism organizations it helped to inspire.
I want to take this opportunity to thank the people who kept FHI alive for so many years against such hurricane-force headwinds. But I also want to express some concerns, warnings, and--honestly--mixed feelings about what that entailed.
Today, a huge amount of FHI's work is being carried forward by dozens of excellent organizations and literally thousands of brilliant individuals. FHI's mission has replicated and spread and diversified. It is safe now. However, there was a time when FHI was mostly alone and the ember might have died from the shockingly harsh winds of Oxford before it could light these thousands of other fires.
I have mixed feelings about encouraging the veneration of FHI ops people because they made sacrifices that later had terrible consequences for their physical and mental health, family lives, and sometimes careers--and I want to discourage others from making these trade-offs in the future. At the same time, their willingness to sacrifice so much, quietly and in the background, because of their sincere belief in FHI's mission--and this sacrifice paying off with keeping FHI alive long enough for its work to spread--is something for which I am incredibly grateful.
A small selection from the report:
21 and 22 hour workdays sounds like hyperbole, but I was there and it isn't. No one should work this hard. And it was not free. Yet, if you ever meet Tanya Singh, please know you are meeting a (foolishly self-sacrificing?) hero.
And while Andrew Snyder-Beattie is widely and accurately known as a productivity robot, transforming into a robot--leaving aside the fairytales of the cult of productivity--requires inflicting an enormous amount of deprivation on your human needs.
But why did this even happen? An example from the report:
This again sounds like hyperbole. It again is not. This was me. After a small grant was awarded and accepted by the university, it took me 308 emails to get this "completed" grant into our account.
FHI died because Oxford killed it. But it was not a quick death. It was a years-long struggle with incredible heroism and terrible casualties. But what a legacy. Thank you sincerely to all of the ops people who made it possible.
Shrimp Mentoring.
Shrimp of Humanity Institute.
Giving Shrimp We Can.
Apparently my wife has been making huge bets on the internet with money from a hidden 'fiat@' account.
Christiano and Amodei are both at OpenAI. Jan Leike, Shane Legg, and Miljan Martic are all at DeepMind. (Jan Leike is also former FHI and is still a research associate with us :) ).
I wrote this in a google doc and copy-pasted, without intending the numbers to be links to anything. I'm not really sure why it made them highlight like a hyperlink.
The closest seems to be really well done analytic philosophy. I recommend Nick Bostrom's work as a great example of this.
I also think that is seems similar to what mathematicians, theoretical physicists, and model builders frequently do.
Other good examples would probably be Thomas Schelling in IR and nuclear security. Coase in economics. Maybe Feynman?