For reference, I surveyed battles/sieges listed not including ones involving Kyiv. Mauripol was first surrounded sometime in the period March 2-4, so it's been holding out for 8-10 days. Mauripol has about 500k population. Kherson, which is around half the size of Mauripol (250k population) fell about 4 days after it was surrounded Kharkiv, with 1.5 million people, has been the site of battles s...

@cjlemontea if "people" are anything like me, they forget which questions they forecasted on until they get emails about them

I think the community is maybe underrating the possibility of Ukr surrender + implementation of Minsk II (among other things). I don't think it's especially likely, but 1% seems too low

I'm at 57%. Rough reasoning is: 70% Russia eventually takes Kyiv, 20% conditional that it takes longer than the deadline. They seem to have switched to a more patient strategy lately.

Kind of surprised this is over 50%, given the well documented incumbent advantage and the fact that Biden has already beaten Trump when Trump had the incumbent advantage

It looks like in the last 30 years two countries out of 10-20 phased out their lowest coin denominations….

I would expect the rate to be higher in the next 30 years; multiplying a bunch of fudge factors together gets me to 12% for the USA in the next 4 years

@Tamay FWIW, I think this report is not very good. They depend a lot on the argument that the epidemiology of SARS-CoV2 is abnormal, but my notes from when I looked into it say “H1N1 (2009) may have only spilled over once”. I think it’s a bit ridiculous not to include that one in the reference class of “zoonotic respiratory pandemics”.

— edited by orion.tjungarryi

Base rate of hostile takeover success is apparently about 35%……

The poison pill might lower this somewhat; the fact that the bid was public might raise it a little. Over all I'm at 30%

@(AvrahamEisenberg) I agree - the question is quite explicit that this is about the release of a pathogen from a laboratory, not simply attributable to the activities of laboratory staff. I am now against resolution. (I hadn't noticed this before) For more explanation: ">50% likelihood of laboratory released pathogen or sampling accident" is consistent with ">50% likelihood of laboratory released pathogen", but so is ">50% likelihood of laboratory pathogen release or natural origin". The latter is clearly not enough for resolution, and the former has a...

"This week, the three-month process of aligning the telescope began – and over the last day, Webb team members saw the first photons of starlight that traveled through the entire telescope and were detected by the Near Infrared Camera (NIRCam) instrument. "

Resolves yes @admins…

I think the amount of noise being made about a potential invasion supports the hypothesis that Russian authorities believe there's some chance of peacefully getting what they want if they apply enough pressure, and that they would prefer this outcome against the hypothesis that they are already determined to invade.

— edited by orion.tjungarryi

I wonder if people are underestimating the scope for outcomes of the type "AI that is bad for a lot of people, but neutral/good for a lot of people too"

— edited by orion.tjungarryi

The biggest issue for me judging this is whether the genome of SARS-CoV-2 supports an artificial origin hypothesis, and how strongly it does so. For [Rootclaim](, evidence from the SARS-CoV-2 genome is strongly in favour of a lab origin (Bayes factor of around 100). [Anderson et. al.]( seems to claim that the SARS-CoV-2 genome is evidence *against* an artificial origin and [Jeffrey Ladish](

How does this resolve if ChatGPT says it’s trained by RLHF but doesn’t correctly explain the additional supervised learning step?

More generally: how many correct details are needed for positive resolution?

@QI92756340QI I've gotta wonder if all that stuff will take more than 6 months though

I've considered:

  • 25% chance Russian leadership feels desperate/panicky on account of both war going badly and economic trouble; 5% chance of Nuke conditional

  • 33% not panicked over war but upset about sanctions cause desperate/panicky feelings; 1% chance of nuke conditional ("high end of baseline risk")

  • 5% Non-Russian nuclear armed actors escalate due to accident or doubling down on investment in Ukraine; 25% chance of Nuke conditional

Overall 2.8%, plus baseline risk from other events gets to 3-4%.

— edited by orion.tjungarryi

@orion.tjungarryi TL;DR: if you agree with rootclaim about the evidence regarding where the virus broke out, then if:

  • P(wrong family|pandemic,zoo,cleavage) < A * P(wrong family|pandemic,lab, cleavage) and
  • P(chimera|pandemic, zoo) < B * P(chimera|pandemic, lab)

and A*B<1/30, then lab escape is >50% likely, otherwise zoonotic is >50% likely.

— edited by orion.tjungarryi