Ten common statistical mistakes to watch out for when writing or reviewing a manuscript (Or when writing Nintil posts)
The history of the structure of science funding in the US
Innovation Growth Lab, a collection of RCTs on what works in science policy
The benefits of allowing grant program management (extend, stop, modify grants as they go) vs the academic system where grants are usually more hands-off.
China keeps making progress in quantum communications
SENS's spinoff Oisin extends life in mice by 20% (Which is an okay figure in aging interventions) via senolytics, human trial to start next year.
Estimating the True Complexity of Comprehensive Rejuvenation - Aubrey de Grey. de Grey is still convinced that the SENS strategy is right, and that the 7 problems of aging he identified over a decade ago are still probably the only things that need addressing.
The unfortunate paradox is that the very aspects of “junk science” that we so properly criticize—the reliance on indirect, highly variable measurements from nonrepresentative samples, open-ended data analysis, followed up by grandiose conclusions and emphatic policy recommendations drawn from questionable data— all seem to occur when we suggest our own improvements to the system. All our carefully-held principles seem to evaporate when our emotions get engaged. This is similar to a pattern noted by Gelman and Loken (2012) that academic statisticians only rarely seem to use statistical principles in designing and evaluating their teaching.
Aaronson op-ed on quantum supremacy
Efficiency is ruining hedge funds, who now have to retreat to harder markets (e.g. private equity)
London's Ultra Low Emissions Zone has been effective in cutting air pollution by a third
Apparently brain drain (researchers emigrating) does not necessarily lead to knowledge drain, as researchers going abroad facilitate knowledge transfer back to the original country (Of course, to a point, if everyone emigrates there would be no researchers left!)
SSC on Indian economic reform
Tiny proteins - with <100 aminoacids - turn out that do more than initially thought.
David Perell on Boeing
Tesla buys 30 year old company to speed up battery production
An article critical with nuclear fusion startups
Earlier in my Links post (29) I mentioned that we are winning the war on cancer. I should have also linked this take from SSC on the relative importance of the sources of that progress: earlier screening detection (Which partly helps treatment, but also confounds statistics), improvement in treatment, and social changes like a reduction in smoking. (ht/ Mohammad Ansarin)
How to increase performance 100x given the same algorithm, giving the code to the CPU the way it likes it.
The most comprehensive review of the IAT, the test that measures implicit bias. Not good news for it.
Open Grants. What do grant applications look like?
Building a beautiful UI with Tailwind CSS
Trying to squeeze science into a formal schema
Inequality in former communist countries, a thread (featuring a blogpost I wrote)
A sizable fraction (80%) of US' raw materials to manufacture drug generics come from China
Attitudes towards LGBTQ individuals have shifted substantially in recent times - a glaring debunking of strict evolutionary explanations of morality -. Why have those changes happened?
Samsara, by Scott Alexander
Sarah Constantin on the rationality of VC investing . Many things to like, some to dispute, eg. she claims that VCs prefer male to female founders, however that's not true in every study that has looked at the question, for example this
Thread on the PG&E California blackout shenanigans
SpaceX's Starlink, an intro
Quantifying the waste in formatting papers: The study found that most of the 203 authors spent 1 to 3 days or more on reformatting alone [...] Based on their survey results, the authors estimated that the total time spent reformatting the 2.3 million scientific articles published annually translates into a global cost of over $1 billion.
Niskanen Center on recent calls for super-high marginal tax rates for the rich.
Alessandro Strumia, a CERN physicist that got kicked out of it for claiming things like "Physics is not sexist" or "Physics was built by men" (Which seems true to me as a first approximation) is about to publish a paper detailing the evidence behind his claims.
There is a mismatch between two ways of calculating Hubble's constant. It was hoped that with better measurement, the discrepancy would go away, but it has not.
Mutation in gene NPSR1 discovered that enables people to sleep less
Building less flawed metrics
You may have noticed there is no Nintil newsletter - this won't change - and that there is RSS instead. Recently I've become aware of the fact that contrary to what I thought, RSS is not in widespread use and people have been abandoning it. That post explains why.
On the merits of OpenAI's recent breakthrough in robotic manipulation
Adan Marblestone's three part series on climate technology
Gwern's link to Readings on the Principles and applications of Decision Analysis
Study tries to value how much people would be willing to pay to forego things like search engines or emails.
When colleges are established, those regions seem to get a boost in patents per year. The same thing is true if instead a region gets a prison or an insane asylum (!?) The read of that thread is also worth reading (e.g. this about R&D tax credits)
The science of lobbying
I ask twitter for examples of evolution making a species simpler