Tuesday, November 12, 2019

Eroding public trust in nutrition science

Harvard has responded to new research that red meat is harmless:
[N]utrition research is complex, and rarely do [its findings] reverse so abruptly. That's why it's so important to look beyond the headlines at the quality of the evidence behind the claims. Still, the publication of these new guidelines in such a prominent medical journal is unfortunate as it risks further harm to the credibility of nutrition science, eroding public trust in research as well as the recommendations they ultimately inform.
Funny how new research nearly always causes further harm to the credibility of nutrition science. Others say:
The misplaced low-fat craze of the 80's was the direct result of Harvard Professor Dr. Hegsted, who participated in the McGovern report that lead to dietary recommendation changes for Americans to eat more carbs in place of meat and fat, a recommendation that turned out to be based on "science" paid for by the sugar industry. Those recommendations caused an explosion of obesity, diabetes, heart disease, and cancer - all metabolic disorders caused by the insulin resistance that resulted from those recommended dietary changes.
My trust in nutrition science is nearly zero.

What do any of these people know about nutrition?

Physicians get a lot of respect for their medical opinions, and they probably deserve it most of the time. But most of them have never taken a course on nutrition, and don't know more than anyone else on the subject.

Everyone eats food, and so has opinions about food. Child-rearing is another subject where everyone has an opinion, but those opinions have almost no scientific value.

The nutrition research is so confusing that I don't know how to conclude that any food is healthier than any other food.

Sunday, November 10, 2019

Academic groupthink on paradigm shifts

Novelist Eugene Linden writes in a NY Times op-ed:
How Scientists Got Climate Change So Wrong ...

The word “upended” does not do justice to the revolution in climate science wrought by the discovery of sudden climate change. The realization that the global climate can swing between warm and cold periods in a matter of decades or even less came as a profound shock to scientists who thought those shifts took hundreds if not thousands of years. ...

In 2002, the National Academies acknowledged the reality of rapid climate change in a report, “Abrupt Climate Change: Inevitable Surprises,” which described the new consensus as a “paradigm shift.” This was a reversal of its 1975 report.
I wonder if he even realizes what these terms means. A scientific revolution or paradigm shift was famously described by Thomas Kuhn as a change in thinking that is incommensurable with previous theories. That is, there is no data to say whether the new thinking is any better or worse than the old. Kuhn described scientists jumping to the new paradigm like a big fad, and not really based on any scientific analysis.

Of course it is all Donald Trump's fault:
computer modeling in 2016 indicated that its disintegration in concert with other melting could raise sea levels up to six feet by 2100, about twice the increase described as a possible worst-case scenario just three years earlier.
Computer models change that much in 3 years? That says more about the instability of the models than anything else.

If the Trump administration has its way, even the revised worst-case scenarios may turn out to be too rosy. ... But the Trump administration has made its posture toward climate change abundantly clear: Bring it on!
Trump is one of the most pro-science presidents we have ever had. Even tho he is widely hated in academia, we hardly ever hear any criticisms of how he has funded scientific work.

Trump has also over-funded quantum computing, and yet Scott Aaronson posts a rant against him. Everyone is entitled to his opinion, of course, but it seems clear to me that academia is dominated by a groupthink mentality that makes their opinions on climate or presidential politics useless.

Wednesday, November 6, 2019

Carroll plugs many-worlds in videos

Lex Fridman interviews Sean M. Carroll on his new quantum mechanics book.

Carroll says that there are three contenders for a QM interpretation: (1) many-worlds, (2) hidden-variables, and (3) spontaneous collapse.

None of these has a shred of empirical evidence. We know that hidden variable theories have to be non-local, and no one has ever observed such a nonlocality. Spontaneous collapse theories contradict quantum mechanics.

After some questions, he admitted another: (4) theories concerned with predicting experiments!

He derided (4) as "epistemic", and complained that those theories (like textbook Copenhagen quantum mechanics) are unsatisfactory because they just predict experiments, and fail to predict what is going on in parallel universes or ghostly unobserved particles.

He also complained that under (4), two different observers of a system might collect different data, and deduce different wave functions.

Yes, of course, that is the nature of science.

Carroll's problem is that he has a warped view of what science is all about. He badmouths theories that make testable predictions, and says that we should prefer a theory that somehow tells us about "reality", but doesn't actually make any testable predictions.

He is a disgrace to science.

Update: See also this Google Talk video, where Carroll makes similar points.

He compares his 3 leading interpretations of QM to the 3 leading Democrat contenders for the White House. Maybe that is a fair analogy, and the leading Democrat contenders are all unfit for office, for different reasons.