TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Critiqued coronavirus simulation gets thumbs up from code-checking efforts

15 pointsby tanhalmost 5 years ago

4 comments

seesawtronalmost 5 years ago
I am not sure if it&#x27;s a thumbs up when they said the code was &#x27;horrible&#x27; and &#x27;buggy mess&#x27;. Key points to note:<p>The authors reason that they had to adapt an older version of the simulations code and hence did not have time to write everything from scratch. I think we have all been there.<p>Reproducibility doesn&#x27;t imply that the underlying model assumptions were free from flaws. As Taleb pointed out in another talk, &quot;It was wrong to use point-estimates for distributions with heavy tails&quot;<p>The lead author Neil Ferguson was unfortunately caught up in another controversy where he left his house during lockdown measures to meet his &quot;married lover&quot;. He had to resign from some of his positions but this shouldn&#x27;t affect the research work he published.
评论 #23491648 未加载
jgalt212almost 5 years ago
This article is all you have to know about the man who is more responsible than any other for the &quot;shut down the world&quot; strategy (which he thought did not apply to him).<p>Coronavirus: Prof Neil Ferguson quits government role after &#x27;undermining&#x27; lockdown.<p><a href="https:&#x2F;&#x2F;www.bbc.com&#x2F;news&#x2F;uk-politics-52553229" rel="nofollow">https:&#x2F;&#x2F;www.bbc.com&#x2F;news&#x2F;uk-politics-52553229</a>
dr_dshivalmost 5 years ago
It&#x27;s a strange but good example of the effects of AI on society.<p>It tells me that we are far away from being able to produce models that incorporate and aim to optimize a broader notion of human wellbeing.
rurbanalmost 5 years ago
The criticism was mainly on the wrong model and wrong data, not the code quality. A good model must be robust and verifiable, this is neither robust nor verifiable. Just compare it to a good model, like the Göttingen model with Bayes inference, which successfully predicted results two weeks into the future. Multiple times. <a href="https:&#x2F;&#x2F;science.sciencemag.org&#x2F;content&#x2F;early&#x2F;2020&#x2F;05&#x2F;14&#x2F;science.abb9789.full" rel="nofollow">https:&#x2F;&#x2F;science.sciencemag.org&#x2F;content&#x2F;early&#x2F;2020&#x2F;05&#x2F;14&#x2F;scie...</a><p>Ferguson&#x27;s model was always exponentially wrong, you can even say catastrophically wrong given the economic costs.<p>Nature still defending this catatrophy, interesting. So they also switched to non- science, supporting politics. The political angle is that they want to suppress extreme right-wing shifts as after the last such political&#x2F;economical crisis 1930, caused by fatefully wrong political decisions. As they happened after and because of the Ferguson study.