@casens To see this in my inbox:

"Metaculus Notifications <notifications\@mg2.metaculus.com>

Resolved: Will at least one nuclear weapon be detonated in Ukraine before 2023?

Was probably the scariest moment of the whole year to me.

Russia told their diplomatic staff at their Ukraine embassy to leave the country. Source from Russian State Media

— edited by ersatz

I think that the community is *vastly* underestimating the enormous size of the Russian army that is currently heading towards the city. A 40-mile-long (64-kilometer) military convoy made up of armored vehicles, tanks, towed artillery and other logistical vehicles. If historical precedents can teach us anything even with a heroic resistance, the Ukrainians will be overwhelmed and defeated in a matter of days. I’m at 95% and I suspect that I’m even a little conservative, my estimate of a 1 in 20 for the city to hold for a whole month is entirely based on ...
[*Two-year update on my personal AI timelines*](https://www.lesswrong.com/posts/AfH2oPHCApdKicM4m/two-year-update-on-my-personal-ai-timelines) by Ajeya Cotra (posted 3rd Aug 2022) > * ~15% probability by 2030 (a decrease of ~6 years from 2036). > * ~35% probability by 2036 (a ~3x likelihood ratio\* vs 15%). > This implies that each year in the 6 year period from 2030 to 2036 has an average of over 3% probability of TAI occurring in that particular year (smaller earlier and larger later). > * A median of ~2040 (a decrease of ~10 years from 2050). > This ...

The Professional Superforecasters of Good judgment are now at 59%.

@ethanbas Why not? To answer seriously, because assigning 99% implies that you put this prediction in a class of 100 predictions of which you are wrong in only one of them. It's incredibly unlikely that you could create such a class for predictions on world events, and especially not for political world events.

@akarlin I think that in a world where they don't intend to invade they have every incentive to give the appearance that they are about to do so, including removing diplomats. This doesn't update me.

New paper by DeepMind: [Making sense of sensory input](https://www.sciencedirect.com/science/article/pii/S0004370220301855) > In a new paper, our team uses unsupervised program synthesis to make sense of sensory sequences. This system is able to solve intelligence test problems zero-shot, without prior training on similar tasks Full abstract: > This paper attempts to answer a central question in unsupervised learning: what does it mean to “make sense” of a sensory sequence? In our formalization, making sense involves constructing a symbolic causal the...

The Professional Superforecasters of Good judgment are at 56%.

— edited by ersatz

@eibbett I don't think that his disagreement with Twitter's far left on Israel and the BDS movement has any significant effect on his chances of being elected, indeed, he merely endorses a very mainstream view, see for example: German parliament condemns 'anti-Semitic' BDS movement. On the other hand, I would have significantly lowered his chances of being elected if he had aligned himself with Twitter's far left on this issue.

I would be interested in the reasoning of people who think that the odds are under 50% based on the information available

@(mumpskin) > Russia's shelling of Stanytsia Luhanska in Ukrainian government-controlled territory in Donbas hit a kindergarten, injured two teachers, and knocked out power in the village. The aggressor in Donbas is clear - Russia. (1/2) https://twitter.com/USEmbassyKyiv/status/1494346636587409410 > This attack, as with so many others, is a heinous Russian violation of the Minsk Agreements and again demonstrates Russia’s disregard for Ukrainian civilians on both sides of the line of contact.(2/2) https://twitter.com/USEmbassyKyiv/status/1494346638202...

I think over 85% is probably overconfident and under 65% probably underconfident.

Knowing how over-regulated the EU is I very much doubt it. I’ll start with 15% based on the odds of countries like Estonia doing it.

@Jgalt That seems very significant, I update to 80% from 75% on that alone.

Good Judgment’s Professional Superforecasters are now at 63% for NO invasion before 31 May 2022. (For the record, I'm at 75% for a positive resolution of this Metaculus question.)

> 16 February 2022 - Good Judgment's professional Superforecasters currently see a 63% probability that a Russian invasion of Ukraine will not take place before June 2022. By "invasion" we mean Russia "sending ground national military forces into Ukraine without consent from the Ukrainian government in Kyiv." Factors driving this consensus forecast include: (1) the expected cost for Russia given the likelihood of Ukraine's resistance and international sanctions that would follow; (2) a strong stand that the US, UK, and parts of the EU took; (3) ongoing d...

@BrunoYammine A minor point but “he said, she said” would imply that both parties have similar credibility, but I don’t think that’s accurate. If the West says that Russia isn’t pulling back its troops and Russia says that actually it is, the thing we learn is that Russia thinks that falsely claiming to do so is in its interest.

— edited by ersatz

@RedBox Nothing has changed strategically because of Russian military modernization, the US would not go to war with Russia in the 90s because of nuclear weapons and will not go to war today for the same reason. Their conventional military forces are totally irrelevant.