Unfortunately, I was really wrong and am now going to make a PCR test and find a plane ticket to somewhere outside Russia. Fortunately, I also have Israeli citizenship, so I have somewhat more options

— edited by mishasamin

@ersatz I don't want to live in an Iran. Also, I own a US company that won't be able to use the US financial institutions if I'm still a resident of Russia

“Emergent bartering behaviour in multi-agent reinforcement learning”

https://www.deepmind.com/publications/emergen…

“Agents learn to produce resources in a spatially complex world, trade them with one another, and consume those that they prefer. We show that the emergent production, consumption, and pricing behaviours respond to environmental conditions in the directions predicted by supply and demand shifts in Microeconomics. We also demonstrate settings where the agents' emergent prices for goods vary over space, reflecting the local abundance of goods. After the price disparities emerge, some agents then discover a niche of transporting goods between regions with d...

https://ai.googleblog.com/2022/06/minerva-solving-quantitative-reasoning.html “In “[Solving Quantitative Reasoning Problems With Language Models](https://storage.googleapis.com/minerva-paper/minerva_paper.pdf)”, we present Minerva, a language model capable of solving mathematical and scientific questions using step-by-step reasoning. We show that by focusing on collecting training data that is relevant for quantitative reasoning problems, training models at scale, and employing best-in-class inference techniques, we achieve significant performance gains...

https://arxiv.org/abs/2206.15378 (DeepMind) “We introduce DeepNash, an autonomous agent capable of learning to play the imperfect information game Stratego from scratch, up to a human expert level. Stratego is one of the few iconic board games that Artificial Intelligence (AI) has not yet mastered. This popular game has an enormous game tree on the order of 10^535 nodes, i.e., 10^175 times larger than that of Go. It has the additional complexity of requiring decision-making under imperfect information, similar to Texas hold'em poker, which has a signifi...

https://www.deepmind.com/publications/perceiver-ar-general-purpose-long-context-autoregressive-generation "Real-world data is high-dimensional: a book, image, or musical performance can easily contain hundreds of thousands of elements even after compression. However, the most commonly used autoregressive models, Transformers, are prohibitively expensive to scale to the number of inputs and layers needed to capture this long-range structure. We develop Perceiver AR, an autoregressive, modality-agnostic architecture which uses cross-attention to map long-...

Metaculus is currently at 54% median prediction, and, well, I’d be happy to find anyone I know to bet 1:1 that Russia won’t invade Ukraine. Yes, the Russian troops are being moved, a lot of troops. CIT regularly reports more info on the troops movement towards the border region (https://citeam.org/russian-exercise-or-not-a-exercise.html?2). But no, I don’t believe Russia is going to invade Ukraine. They have no actual interest in it, they’ve got a lot of internal problems, the surveys show the invasion won’t improve their ratings that continue to decl...

I submitted a question on whether it will be acknowledged by Metaculus admins before 2023 that Russian troll factories tried to influence Metaculus

[Metaculus News](https://metaculus.medium.com/900-forecasters-on-whether-russia-will-invade-ukraine-f8b36f7dee0f): *"Today, the world’s attention is riveted by the intensifying military brinkmanship playing out on the Russian-Ukrainian border. To find a signal in all the noise, Metaculus statistically aggregates the predictions of the crowd and gives more weight to those with the best track records and the most forecasting skill."* "​​The current Metaculus Prediction gives Russia a 65% chance of invading new territory in Ukraine in 2022. This is 14 per...

https://www.deepmind.com/blog/discovering-nov…

“we adapted AlphaTensor to specifically find algorithms that are fast on a given hardware, such as Nvidia V100 GPU, and Google TPU v2. These algorithms multiply large matrices 10-20% faster than the commonly used algorithms on the same hardware”

“YouTube can be blocked in Russia before the end of the next week or maybe even today” - RIA (a Russian state-controlled news agency) citing an anonymous source

“The decision was expected the last week, but Instagram was blocked which might have postponed the decision on YouTube”

https://t.me/rian_ru/154363

https://openai.com/blog/vpt/

“We trained a neural network to play Minecraft by Video PreTraining (VPT) on a massive unlabeled video dataset of human Minecraft play, while using only a small amount of labeled contractor data. With fine-tuning, our model can learn to craft diamond tools, a task that usually takes proficient humans over 20 minutes (24,000 actions). Our model uses the native human interface of keypresses and mouse movements, making it quite general, and represents a step towards general computer-using agents.”

There’s a website with translations of Belarusian news: https://www.voiceofbelarus.com/

@KnowName @akarlin the question is about whether or not Navalny will become the president or prime minister of Russia. IF he dies before becoming either, the question will resolve negatively. It's wrong to resolve the question if he's still alive and has chances. Resolving ambiguously if Navalny is still alive by 2500 makes sense.

When predicting 95% yes, I was actually thinking about whether or not any Metaculus user would be tested positive for the coronavirus in 2020. I was pretty confident someone would catch COVID-19 even though the users here were probably more cautious about the coronavirus than the population. But I totally wasn't thinking about *P(user reports here|tested positive)*. On the retrospective, it's obvious that the reporting is the part that might've gone wrong even if my predictions about the pandemic were correct. Well, - 129 virtual points and + a lesson a...

“To test Minerva’s quantitative reasoning abilities we evaluated the model on STEM benchmarks ranging in difficulty from grade school level problems to graduate level coursework.” “MATH: High school math competition level problems” - 50.3% (previous published SOTA 6.9%) On the MATH benchmark, PaLM achieves 8.8%; davinci-002 achieves 19.1%. “MMLU-STEM: A subset of the Massive Multitask Language Understanding benchmark focused on STEM, covering topics such as engineering, chemistry, math, and physics at high school and college level.” - 75% (SOTA 55%) ...

Aric Toler from Bellingcat noticed that the videos about the evacuation by the DRN and LNR leaders were recorded on February 16 https://twitter.com/AricToler/status/14947392…

@(akarlin) (1) when Russian soldiers vote, the process is observed by their bosses, and they assume there’s no secrecy on how they vote. It doesn’t at all represent the views of the soldiers. Also, there weren’t any actual liberal candidates on the ballot of any federal elections for a decade. Navalny ran for the Moscow mayor position in 2013, but those weren’t federal elections; Sobchak (a goddaughter of Putin) was a “liberal candidate” in the 2018 presidential elections in order to make the elections appear more legitimate (and Navalny later revealed ...

Good Judgment predicts 45% no invasion and 13% invading only the Donetsk and/or Luhansk Oblasts

Update: 27% no invasion and 18% invading only the Donetsk and/or Luhansk Oblasts

— edited by mishasamin