I think it may make sense to repeat this question but without the self-resolving feature. Because it seems like currently people should predict the public attitude towards the lab release (prediction is heavily adjusted towards "maybe it would reach 3% soon" or "My analysis shows that lab release is very likely but the commitee won't say it's >50% likely"). I'd probably be more interested in what is people's actual subjective confidence of lab release. The ["public agencies claim by 2024"](https://pandemic.metaculus.com/questions/3682/before-the-end-of-...

Seems like some metaphorical use of words "civil war" when describing a social unrest could trigger a positive resolution.

In June, there was around 137K COVID deaths worldwide according do worldometers counts.

I don't have sources on hand but the current public consensus seems to be that there's not much evidence for synthetic origin (at least in my perception). However, as China may begin to push [the story](https://www.the-scientist.com/news-opinion/chinese-officials-blame-us-army-for-coronavirus-67267) of the connection between US troops and the virus origination, there is a very small chance of "The Chinese Center for Disease Control and Prevention" making the claim about synthetic origin. The question references "Chinese virology laboratory" so foreign l...
I liked [this blogpost](https://moalquraishi.wordpress.com/2020/12/08/alphafold2-casp14-it-feels-like-ones-child-has-left-home/) on AlphaFold2. It is by professor Mohammed AlQuraishi from Columbia University where he researches protein structure prediction among other main research directions of his lab. It examines what we know about the model and its implications for different parts of biology in detail. "Although I never wrote it in the CASP13 post, my expectation (back then) for when we would nail all the details was another 10 years, i.e., not unti...

gwern estimates that OpenAI may currently receive around $120M in yearly revenue from its API.

Do mixture of experts networks count? I think the question needs to be precise on this. Here's a [2017 paper](https://arxiv.org/abs/1701.06538) with a net in hundreds of billions of parameters, which contradicts the question assumption that "largest models have gone from having 94 million parameters in 2018". The recent 600B [Gshard model](https://arxiv.org/abs/2006.16668) is also a sparse MoE. MoE and dense networks seem to have a very large discrepancy in the amount of compute that they require: e.g. 2B dense transformer was trained with an order of ma...
@(TheSingleMalt) It is my impression that likely the majority of people here might place more weight on the virus escaping from some facilities connected with research in accident, rather than on it being the result of "purposeful manipulation". The last Nature paper sentence that you cite - "However, since we observed all notable SARS-CoV-2 features, including the optimized RBD and polybasic cleavage site, in related coronaviruses in nature, we do not believe that any type of laboratory-based scenario is plausible." - confuses me a bit, because all the...
By a rough count 5/12 states with the most COVID-19 deaths per capita in 2020 over the whole year voted for Trump https://covid.cdc.gov/covid-data-tracker/#cases_deathsper100k (but those states that didn't vote for Trump at the top have really high numbers so you can't outrank them easily) 13/18 states with the most COVID-19 deaths per capita in 2020 over the whole year in the remaining list (without those 12 top states) voted for Trump 11/12 states with the most COVID-19 deaths per capita in 2020 over the last week voted for Trump https://covid.cdc....
My prior view - they may eventually ban it because they have completely failed to control the flow of conversation on the internet via propaganda. It just seems unlikely to happen in 2021-2022. It would require a massive acceleration of their invasion of people's rights, or otherwise extreme protests which are not on the horizon. And they would have neither technical capabilities nor the desire to block it completely like it is done in China - it would work with a VPN. It seems difficult to comprehend and decipher the influence of the new law. I wouldn'...

Today in Hrodna (population 370K) the local power structures satisfied all the demands of protesters on the city level. Among other things, the police freed all detained peaceful protesters and apologized, the law was changed to allow them to protest everywhere in the city, they were promised TV coverage, and directors of local enterprises were ordered not to punish workers on the strike. (source: @nexta_live in Telegram)

— edited by alexlyzhov

The strikes have petered out really quickly. Workers on the largest or most profitable enterprises no longer organize large strikes. Example. The opposition news orgs don't write much about strikes anymore. Without such action, I don't see any likely way this is resolving positive even if you double or triple the number of people on the streets because the window of opportunity has closed.

A random "N501Y" Twitter search about the new UK set of mutations pulled this: https://twitter.com/firefoxx66/status/1338535621636333569 "As with other mutations: it is unlikely that a few mutations render a vaccine ineffective. The vaccine causes your body to recognise many parts of spike, so a few changes should not mean it stops working. However impact of more changes is less sure - scientists investigating this now." (Co-developer of @nextstrain) https://twitter.com/firefoxx66/status/1340360016054906882 "Will it impact the vaccine? I don't think ...

Hmmm, the community is at 99% but the current mean score probably falls short and I would say there is >1% on the next CASPs being cancelled because John Moult who co-founded CASP said that "in some sense, the problem is solved". For example, look what happened with ImageNet challenge: looks like it lasted for ~5 years after the advent of CNNs.

— edited by alexlyzhov

Picture of flight available in NASA live stream around 40 minutes in https://www.youtube.com/watch?v=p1KolyCqICI

— edited by alexlyzhov