Antioxidants are widely to be held either good or non-effective. But in some cases they can be bad, in particular promote cancer (in mice)

What do we know about R&D spillovers and productivity?

From a public policy perspective, our findings indicate that the case for public support for R&D investment (input subsidies) may not be as strong as has been assumed so far. We have established empirically that the direct productivity effect of own R&D is comparable to and usually larger than that of spillovers. We have also demonstrated that the spillover effect is larger when firms/industries/countries possess a higher level of absorptive capacity, which is a positive function of ownR&D investment. Given these findings, the R&D gap - i.e., the gap between actual and socially optimal levels of R&D investment at the firm level - maybe smaller and more heterogeneous than what has been assumed so far.

The R&D gap may be narrower because firms are aware of the need to invest in own R&D and build absorptive capacity as a basis for gains from knowledge externalities. Also, the R&D gap is likely to be heterogeneous, due to different firm characteristics (age, size, market share, etc.) and different industry characteristics (e.g., technology type, technology frontier and nature of competition). Therefore, we argue that direct or indirect support for R&D investment may be tooblunt an instrument for securing additional R&D effort by supported firms.

Inequality trends heavily depend on how it is measured, one more case

A combination of three drugs (rapamycin, lithium, and trametinib) extend lifespan IN FLIES!

Automated discovery in the Sciences, Derek Lowe discusses a review of its applications.

Book argues that there are more conflicts now that earlier in History, criticising Steven Pinker's arguments to the contrary.

Culture seems to be what matter for entrepreneurialism, rather than national investment in R&D, VC funding, or availabiliy of STEM education? Have not read it yet, but worth looking into.

Harking, sharking, and tharking, or how to approach the evaluation of results when doing science. The tiny brain version here is that you get some data and you figure out what the data means without any prior hypothesis. The medium brain version is that that's bad because you can just keep slicing the data until you get significant results that are actually fake, which is why preregistration -announcing in advance how you will slice the data, and sticking to it- is so important. And the galaxy brain is to pre-commit, but then also be free to speculate post-hoc, clearly saying that this is what you are doing.

My ongoing literature review on peer-review and grantmaking practices.

The effect of corporate scientists on firm's innovation activity: a literature review

Computer scientists David Gelernter publicly rejected Darwinism, Jerry Coyne on why he is mistaken.

What happens to companies when their founding CEOs die? They become less innovative apparently.

It is believed that the cost of pet healthcare has been rising as fast as human healthcare in the US. But maybe not.

A great list of underinvestigated fields, by Matthew McAteer

Thread featuring Michael Nielsen, Alexey Guzey, and myself, among others, on why we sleep.

How can we develop transformative tools for thought?

A review of the usefulness of structural equation modeling

Me on the weirdness of naming conventions of genes and proteins , which sometimes lead to funny situations

An exercise: meta-rational phenomena

Data on how well VCs do as a whole (also this) Adjusting for risk, the returns are what one would expect from regular financial theory.

How do we move the needle on progress?

How often do studies done in mice replicate? Twitter replies. The IN MICE! meme is justified: Only 37% replicate in humans, and many of the animal studies are low quality anyway.

Stian Westlake and Sam Bowman's pamphlet on what the best policies for the UK would be.

Is it even possible to record every single neuron in the brain?

There are more people employed in nonprofits than manufacturing, in the US

Quantum supremacy: Scott Aaronson and Gil Kalai

New Doug Irwin paper on the history of US trade policy

Airfares in the US vs Canada. An example of how progress (in price reduction) is not just about newer technology

A list of all (?) aging biotech companies

"From the National Museum of Ireland’s archaeology section: in the Bronze Age, it seems that a way of swearing fealty to a king was to ritually suck their nipples", and more from Anton Howes' visit to Dublin

Does offering economic incentives raises IQ? A meta-analysis seemed to claim so, but looking at particular studies within the literature seems to show that it does not. Thread.

Do cancer drugs act in the way it was thought they do? Apparently older studies did show that inhibiting gene expression via RNA interference worked for some of those genes. But what if it was also knowing down other genes? In this paper, they used CRISPR, which is more precise to one by one knock out the genes directly and see exactly how they interact with the drugs that target them

One of five genetics papers contains errors because of Excel?

Massive breakthrough from Tesla in battery research

Bret Victor's overview of climate change and what a technologist can do about it.

A paper claimed to have reverted human aging (As measured by the Horvath clock) in a sample of 9. Some critiques

Some experimental evidence in favor of self-deception really helping better deceive others

I asked twitter: Are there languages more efficient than others? Apparently most language have a similar efficiency, except for Japanese, which seems worse. But sounds cool.

Some pictures of African architecture