Builds web apps (eg viewpoints.xyz) and makes forecasts. Currently I have spare capacity.
Nathan Young
Karma: 17,346
I like Ben personally.
I don’t intend to quote tweet him, but I’d like someone to make a kind of defence.
Ben Landau-Taylor tweeted this a couple of days back:
It has been annoying me, since I don’t think it’s accurate. Here is my proposed response (These aren’t tweets, it’s a scheduling app where I draft):
I would appreciate criticism.
I think it would allow many very online slightly anxious people to note how the situation is changing, rather than plug their minds into twitter each day.
I think the bird flu site helped a little in my part of twitter to tell people to chill out a bit and not work themselves into a frenzy earlier than was necessary. At least one powerful person said it was cool.
I think that civil servants might use it but I’m not sure they know what they want here and it’s good to have something concrete to show.
Seems notable that my modes is that OpenPhil has stopped funding some right wing or even centrist projects, so has less power in this world than it could have done.
Ben Todd writes:
Most philanthropic (vs. government or industry) AI safety funding (>50%) comes from one source: Good Ventures, via Open Philanthropy.2 But they’ve recently stopped funding several categories of work (my own categories, not theirs):
Many Republican-leaning think tanks, such as the Foundation for American Innovation
...
Likewise almost all the people I know 2nd hand who have power in the new admin are via either twitter or rationality, both spaces where EA has, to my eye, sought to distance itself in recent years.
I have been tempted to make a US democracy dashboard, tracking the likelihood according to prediction markets of different democracy outcomes.
Similar to https://birdflurisk.com
Should anyone wish to fund this to cause it to happen faster, they are welcome to dm me.
Oppenheimer
Alexander Hamilton
I would want to see a big database of the exact wording of the statement that I could look up or at least some large random sample.
Paul Christiano
Nixon
Napoleon
Kissinger
Einstein
Von Neumann
Eliezer Yudkowsky
Yeah I have long thought this.
Though sometimes it is true, eg if someone makes an insigntful comment and then I write it up, it really is reasonable to say thanks to them for making but that errors are my own.
Bird flu is probably fine right now. Let’s not cry wolf.
I’ve been looking into H5N1 bird flu and built this dashboard (https://birdflurisk.com)
To me the indicators suggest it’s likely gonna be fine. You can see the forecasts are pretty low and even if these resolve positive it probably won’t be a big deal to humans (see note in a sec)
I think it’s worth becoming well calibrated on risk ie only crying wolf when there is a wolf and right now I see no wolf, so as a community we improve our calibration by saying “bird flu will almost certainly be fine”
That said, it probably will involve farms full of chickens being tortured to death if they catch the disease. This is tragic. I suggest it requires a different comms strategy though.
Also there may be inflation with the political ramifications of that.
Let me know what would make the dashboard more useful to you.
I like this, and have been trying a similar visual approach using squiggle. I agree that LLM estimation using squiggle seems tractable and that it could help turn many text outputs into quantifiable/comparable numerical outputs.
I am interested in creating a space to compare/rank these outputs. @Ozzie Gooen do you see squiggle hub as the space for this?
I find it very funny that such a huge donation basically happened by accident. Surreal stuff.
I dunno, by how much? Seems contingent on lots of factors.