Director of Epoch.
Currently working on:
Macroeconomic models of AI takeoff
Trends in Artificial Intelligence
Forecasting cumulative records
Improving forecast aggregation
I also run Connectome Art.
Jaime Sevilla
Karma: 2,207
FWIW I reviewed and redid the analysis in Heinrich’s paper.
The statistical relationship seems quite robust, but the analysis of causation is purely qualitative.
I am out of my depth assessing the psychological and historical evidence they cite, and it might well be true that kinship is an important component of it. However the usual standard for evidence in econometrics is a natural experiment, which they did not do.
Thank you Michael!
I write these mostly for self-reflection and fun!
Then I post them for praise, being a role model and promoting my projects (which has led to some cool opportunities).
Unclear which of these is the main benefit :3
Thank you for your kind words!
Some quick notes on productivity:
I am in a very fortunate position to be able to pursue projects that I find very motivating. I often find myself working on what I find most interesting and pursuing obsessions.
At Epoch we have a dedicated slack channel for daily commitments. I found this useful to prioritize what to work on and get annoying tasks done.
I am finding coaching with Katie Glass useful to reflect on what I coud improve. Examples include being more careful with accepting random requests for help, blocking off time for focused work when people cannot book a meeting with me, and interweaving walks with problem solving.
I collaborate early and often with people. For effective collaboration, is important to 1) have a lead person leading the project and seeing it done and 2) making it clear what everyone is supposed to contribute to the project.
My usual day looks like: start working at 10:00, write a list of daily goals, work on these goals until lunchtime, take short break, then work until 17:00.
The two most useful ways to motivate me to work productively in a project are 1) a weekly meetings to discuss advances and priorities, 2) having an idea of the concrete goal I am aiming for (a paper, a website, a conference, etc) and 3) having a clear idea of what the next step is.
Bahamian Adventures: An Epic Tale of Entrepreneurship, AI Strategy Research and Potatoes
There is also an Epoch summary of it!
I added a section with an overview!
I didn’t keep good track of them, but like ~5 of them were spam.
They included some bizarre texts like someone who wanted to sell us a new “discovery” of a medicine and two people self promoting.
The funniest one is one titled “mega-earthquakes” that talked about… you guessed it, the Christian apocalypse.
No real obstacle- we just neglected to add trackable links. We will do so next time!
Results of a Spanish-speaking essay contest about Global Catastrophic Risk
Impressively well called, and congratulations on the prize!
That´s a good point—I expect most of these discussions to lead to edits rather than publications.
I downvoted because 1) I want to discourage more conversation on the topic and 2) I think its bad policy to ask organizations if they have any projects they decided to keep secret (because if its true they might have to lie about it)
In hindsight I think I am overthinking this, and I retracted my downvotes on this thread of comments.
I recall three in depth conversations about particular Epoch products. None of them led to a substantive change in publication and content.
OTOH I can think of at least three instances where we decided to not pursue projects or we edited some information out of an article guided by considerations like “we may not want to call attention about this topic”.
In general I think we are good at preempting when something might be controversial or could be presented in a less conspicuous framing, and acting on it.
Laplace’s rule of succesion is often used by forecasters to set base rates. What does he think of that? Is this a good rule of thumb?
Moreover, Laplace’s rule gives different results depending on how finely you subdivide time (eg saying that there has been one year of global pandemic in the last 20 years will give different results that if you say there has been 12 months of pandemic in the last 20*12 months). How should we account for that inconsistency when applying Laplace’s law?
How should one aggregate different forecasts? Is External Bayesianity a compelling criteria for an aggregation procedure? How about marginalization?
META LEVEL REPLY
Thinking about the ways publications can be harmful is something that I wish was practiced more widely in the world, specially in the field of AI.
That being said, I believe that in EA, and in particular in AI Safety, the pendulum has swung too far—we would benefit from discussing these issues more openly.
In particular, I think that talking about AI scaling is unlikely to goad major companies to invest much more in AI (there are already huge incentives). And I think EAs and people otherwise invested in AI Safety would benefit from having access to the current best guesses of the people who spend more time thinking about the topic.
This does not exempt the responsibility for Epoch and other people working on AI Strategy to be mindful of how their work could result in harm, but I felt it was important to argue for more openness in the margin.
OBJECT LEVEL REPLY:
Our current publication policy is:
Any Epoch staff member can object when we announce intention to publish a paper or blogpost.
We then have a discussion about it. If we conclude that there is a harm and that the harm outweights the benefits we refrain from publishing.
If no consensus is reached we discuss the issue with some of our trusted partners and seek advice.
Some of our work that is not published is instead disseminated privately on a case-by-case basis
We think this policy has a good mix of being flexible and giving space for Epoch staff to raise concerns.
Announcing Epoch: A research organization investigating the road to Transformative AI
I have downvoted this comment.
I broadly resonate with the message that EAs should focus on the things that make them unique and that we should uphold the mentality of figuring out the most impact.
But I think some parts of the EA mindset would be very useful to tackle some other important issues like reproductive rights, and I think we should encourage playful and scientific exploration of topics.
These explorations are good exercises of cost effectiveness analysis, will help us find new problems to tackle and curiosity is a great value to promote.
Let co-authors access post analytics
I can get around this by asking the main coauthor to share the analytics, but I´d rather I could access them myself.
Thank you Miguel!