In an earlier links, I linked to a post critical with the "polyunsaturated fatty acids (PUFAs) are bad for you" thesis. This other post is a reply to that one. (For what's worth I still think that saturated fat is bad and that olive oil is all the oil you need so why bother with seed oils :)

"Charge more", niche keyboards edition

Dominic Cummings on how to really take over a government (it's the deep state what you need)

A rare paper showing a reversal in atherosclerotic plaque (in mice) using cysteamine (which is FDA approved btw). Out of curiosity, I looked at the funders of this, and it's the British Heart Foundation (No major funders like NIH involved)

Where are the large Republican cities, asks Bryan Caplan. His own tentative answer is unconvincing to me though.

Progress in treating cancer in mice with a combined mRNA therapy

A book on longevity research written in 1985. Much has changed since then!

Bottlenecks in Science and Technology 2021 now has some videos online!

Retrotransposons as tumor suppressors in blind mole rats

How likely are clinical trials to succeed? Here's the data on that. There seems to be an uptick in approval rates since ~2018, now it's more likely that a successful Phase 1 eventually ends in approval.

Adam Marblestone just published a number of essays on bottlenecks in various fields (I contributed to the aging one), but also biological resilience, brain mapping, and nanofabrication

Attempt at deriving "breakthrough innovations" from a sample of patents.

Lab-grown meat is further in the future than commonly thought

Reminder that the standard model of particle physics does not fully agree with all experimental data: There are longstanding deviations from known physics that are known to be so with very high confidence. An example.

Scott Alexander (and comments) on modern architecture. Pushback from Rohit. I tend to agree with Scott. Modern architecture, I think, lacks detail. When it does have detail then it perhaps looks better to us? Take a minimalistic construction consisting of interlocking cubes. The design can be glanced in one go, and there's no more visual depth to it so it seems boring. In contrast, something like the Sagrada Familia (the most beautiful building in the world, imho) has a lot of detail. Look at this beautiful ceiling! Such feeling of awe is impossible to approach with a flat slab of concrete for a roof and openings on the sides to get some light. Modern architecture may be novel, but is lazy; trad architecture is obviously not novel, but it takes hard work to actually design one building in a traditional style, when there are more elements added there's more complexity about what those will be and where they go. Buildings like the Sagrada Familia (and others in Rohit's post) show how we can have both modern beautiful buildings without having to resort to the nth copy of roman temples that government buildings and banks are so keen to emulate.

Delivering gene therapies to places other than the liver, spleen, and lungs is hard. Now there's a new tool that makes it possible to deliver them to muscle cells with high precision. I'm going to collect methods of delivering beyond the liver here.

New non-CRISPR gene editing system found in bacteria

Avoiding the immune system, more so than just trying to grow faster, is what cancers are mostly "trying" to dp

The evolution of the hallmarks of aging

I ask Twitter "What is the metaverse?"

Thymic involution happens with aging. But apparently it happens substantially later in naked mole rats

Organs change size through the day

Long interview with Jonathan Blow

Lenghty post argues that, all things considered, eggs are bad for you (The degree of badness discussed in the post). Includes lots of interesting discussion: Considering the baseline of the study (e.g. 1 egg to a population that is already eating a lot of fat might not do much to their cholesterol 1; 1 egg to someone eating healthy will) or pointing out sneaky things studies adjust for (e.g. if a study is adjusting for serum cholesterol when looking at the effects of eating eggs on cardiovascular disease, and serum cholesterol is a causal mediator in that relation, the relation is being controlled away!). This site has a number of articles that recapitulate what used to be common sense a few years ago, but which now seem like hot takes. I think the old common sense is true: saturated fat is bad, or that dietary guidelines are reasonable for most people to follow.

One of the reasons why large (tech) companies can have lower productivity: the way promotions work

Little did we know: What came first is not AI writing credible news articles, but humans writing nonsense ones.

Verve Therapeutics still seeing positive results for their gene therapy to permanently reduce cholesterol

My wildfires post was edited for conciseness in the a16z Future publication, you can read it here

Another explanation for the obesity epidemic: the rise of snacking and sweet beverages

Analysis of 2014 NeurIPS data shows that reviewer scores are not predictive at all of quality (as measured by citations). For rejected papers, oddly, there is a correlation, in the expected direction. For accepted papers it wasn't the score itself, but a confidence score ("how sure are you of your assessment") that is predictive of citation quantity (perhaps because the prestige of "published in NeurIPS" swamps everything else regarding citations?). I cite this same dataset in my "Does pre-grant peer review work?" post. Overall, I think that reviewer nihilism is not warranted yet (given the right structure, review can be more informative than chance; and having more than one reviewer can help, via Condorcet's jury theorem, make more robust choices). But one thing the authors say in their paper is true: Given that we’ve shown that such inconsistencies exist, we would suggest that we should be vary wary of ’top-tier’ publication counts as a measure of individual researcher quality

Blue Origin paid consultants to try to be as good as SpaceX. Then seemingly ignored all their good advice. Like most of management advice and self-help, on its face is obvious, and the devil is both in the details and actually following the advice.

Why wokeism will rule the world, by Tyler Cowen

Lenat lives, Cyc continues

Bottlenecks in reducing costs of energy in the US

GPT-3 really likes to generate fake news

Cells have massive organelles called vaults. Twitter finds out and loses its collective bio-brained mind about it

Notion interview with Alan Kay

The Cell Metabolism covers are amazing

Works in Progress, issue 5

Videos from the Metascience 2021 conference

Max Hodak just started a company to do what I would describe as NCI, Nerve-Computer Interfaces (e.g. the optical nerve is easier to use as input than electrodes in the visual cortex)

Impetus Grants: A Fast Grants for aging

Profile of Sam Zeloof, who makes chips in his garage

Age has an impact on the genetic risk for various diseases: As we grow older the impact of genetics declines. This means that in the later stages in life, the shared biology of aging overwhelms the effects of the variants that slow down aging

This year's Burning Man happened informally (the official organizers didn't organize the event). An account of how it went.

Accelerated Expertise, a review

The early history of mRNA vaccines

The evolution of the p53 protein as it relates to longevity

Cost disease socialism

Casa Gilardi

For a change from the usual links, some music I've been listening to lately: Meute and Snarky Puppy