GPT-3 Is an Advanced A.I. Language

Generative Pre-trained Transformer 3 (GPT-3) is an autoregressive language model that uses deep learning to produce human-like text.

It is the third-generation language prediction model in the GPT-n series (and the successor to GPT-2) created by OpenAI, a San Francisco-based artificial intelligence research laboratory.  GPT-3's full version has a capacity of 175 billion machine learning parameters.  GPT-3, which was introduced in May 2020, and was in beta testing as of July 2020, is part of a trend in natural language processing (NLP) systems of pre-trained language representations.

Before the release of GPT-3, the largest language model was Microsoft's Turing NLG, introduced in February 2020, with a capacity of 17 billion parameters—less than a tenth of GPT-3's.

The quality of the text generated by GPT-3 is so high that it can be difficult to determine whether or not it was written by a human, which has both benefits and risks.  Thirty-one OpenAI researchers and engineers presented the original May 28, 2020 paper introducing GPT-3. In their paper, they warned of GPT-3's potential dangers and called for research to mitigate risk.  David Chalmers, an Australian philosopher, described GPT-3 as "one of the most interesting and important AI systems ever produced."

Microsoft announced on September 22, 2020 that it had licensed "exclusive" use of GPT-3; others can still use the public API to receive output, but only Microsoft has access to GPT-3’s underlying model.

Background

According to The Economist, improved algorithms, powerful computers, and an increase in digitized data have fueled a revolution in machine learning, with new techniques in the 2010s resulting in "rapid improvements in tasks" including manipulating language.  Software models are trained to learn by using thousands or millions of examples in a "structure ... loosely based on the neural architecture of the brain".  One architecture used in natural language processing (NLP) is a neural network based on a deep learning model that was first introduced in 2017—the Transformer.  GPT-n models are based on this Transformer-based deep learning neural network architecture. There are a number of NLP systems capable of processing, mining, organizing, connecting, contrasting, understanding and generating answers to questions.

On June 11, 2018, OpenAI researchers and engineers posted their original paper on generative models—language models—artificial intelligence systems—that could be pre-trained with an enormous and diverse corpus of text via datasets, in a process they called generative pre-training (GP).  The authors described how language understanding performances in natural language processing (NLP) were improved in GPT-n through a process of "generative pre-training of a language model on a diverse corpus of unlabeled text, followed by discriminative fine-tuning on each specific task." This eliminated the need for human supervision and for time-intensive hand-labeling.

In February 2020, Microsoft introduced its Turing Natural Language Generation (T-NLG), which was then the "largest language model ever published at 17 billion parameters."  It performed better than any other language model at a variety of tasks which included summarizing texts and answering questions.

Capabilities

On May 28, 2020, an arXiv preprint by a group of 31 engineers and researchers at OpenAI described the development of GPT-3, a third-generation "state-of-the-art language model".  The team increased the capacity of GPT-3 by over two orders of magnitude from that of its predecessor, GPT-2, making GPT-3 the largest non-sparse language model to date.  Because GPT-3 is structurally similar to its predecessors, its higher level of accuracy is attributed to its increased capacity and higher number of parameters.  GPT-3's capacity is ten times larger than that of Microsoft's Turing NLG, the next largest NLP model.

Sixty percent of the weighted pre-training dataset for GPT-3 comes from a filtered version of Common Crawl consisting of 410 billion byte-pair-encoded tokens.  Other sources are 19 billion tokens from WebText2 representing 22% of the weighted total, 12 billion tokens from Books1 representing 8%, 55 billion tokens from Books2 representing 8%, and 3 billion tokens from Wikipedia representing 3%.  GPT-3 was trained on hundreds of billions of words and is capable of coding in CSS, JSX, Python, among others.  Since GPT-3's training data was all-encompassing, it does not require further training for distinct language tasks.  The training data contains occasional toxic language and GPT-3 occasionally generates toxic language as a result of mimicking its training data. A study from the University of Washington found that GPT-3 produced toxic language at a toxicity level comparable to the similar natural language processing models of GPT-2 and CTRL. GPT-3 produced less toxic language compared to its predecessor model, GPT-1, although it produced both more generations and a higher toxicity of toxic language compared to CTRL Wiki, a language model trained entirely on Wikipedia data.

On June 11, 2020, OpenAI announced that users could request access to its user-friendly GPT-3 API—a "machine learning toolset"—to help OpenAI "explore the strengths and limits" of this new technology.  The invitation described how this API had a general-purpose "text in, text out" interface that can complete almost "any English language task", instead of the usual single use-case.  According to one user, who had access to a private early release of the OpenAI GPT-3 API, GPT-3 was "eerily good" at writing "amazingly coherent text" with only a few simple prompts.  In an initial experiment 80 US subjects were asked to judge if short ~200 word articles were written by humans or GPT-3. The participants judged incorrectly 48% of the time, doing only slightly better than random guessing.

Because GPT-3 can "generate news articles which human evaluators have difficulty distinguishing from articles written by humans," GPT-3 has the "potential to advance both the beneficial and harmful applications of language models."  In their May 28, 2020 paper, the researchers described in detail the potential "harmful effects of GPT-3" which include "misinformation, spam, phishing, abuse of legal and governmental processes, fraudulent academic essay writing and social engineering pretexting".  The authors draw attention to these dangers to call for research on risk mitigation.

GPT-3 is capable of performing zero-shot, few-shot and one-shot learning.

Controversy

GPT-3's builder, OpenAI, was initially founded as a non-profit in 2015.  In 2019, OpenAI did not publicly release GPT-3's precursor model, breaking from OpenAI's previous open-source practices, citing concerns that the model would perpetuate fake news. OpenAI eventually released a version of GPT-2 that was 8% of the original model's size.  In the same year, OpenAI restructured to be a for-profit company.  In 2020, Microsoft announced the company had exclusive licensing of GPT-3 for Microsoft's products and services following a multi-billion dollar investment in OpenAI. The agreement permits OpenAI to offer a public-facing API such that users can send text to GPT-3 to receive the model's output, but only Microsoft will have access to the GPT-3's source code.

Large language models, such as GPT-3, have come under criticism from Google's AI ethics researchers for the environmental impact of training and storing the models, detailed in a paper co-authored by Timnit Gebru and Emily M. Bender in 2021.

More (including reviews) at:  https://en.wikipedia.org/wiki/GPT-3

Contrary to President Biden, Government Spending Is Never 'Paid For'

By John Tamny, RealClear Markets

October 5, 2021 – “…It’s a reminder that their [government] consumption of precious wealth doesn’t instigate economic growth; rather it’s a consequence of it. It’s loud evidence of the ferocious stupidity of a number routinely cited by politicians and economists: GDP. GDP would go down – a lot – if government spending were slashed, which just goes to show how bankrupt the calculation is. Growth happens, or is expected to happen, and politicians spend its fruits. There’s no new growth to speak of from this political consumption. To say otherwise is to engage in double counting.

“So, the Biden administration claims the trillions he would like spent are “paid for.” No, they’re not. The trillions spent are a cost. They’re an economic somnolent. Instead of trillions of precious wealth being directed to its perceived highest use in the marketplace, it will instead be allocated by people like Marco Rubio and Alexandria Ocasio-Cortez. Are Rubio and AOC as skillful at investing as Warren Buffett and Peter Thiel? Please don’t answer. It would be a waste of words.

“The simple truth is that government spending is a burdensome tax. And it’s felt right away as businesses and entrepreneurs get by with less so that politicians can spend more. And no, there’s no short-term “sugar high” as even some on the Austrian School right presume. The burden of government consumption is immediate, and it saps growth immediately.”

More at:  https://www.realclearmarkets.com/articles/2021/10/05/contra_president_biden_government_spending_is_never_paid_for_797433.html

Why Back Pain Is So Common

From: Live Science

By Tara Santora

October 5, 2021 – Back pain is incredibly common, with 26% of Americans reporting at least one full day of lower-back pain within a three-month period, according to a 2006 study in the journal Spine. It's also the leading cause of disability across the globe, according to a 2014 study in the journal Annals of the Rheumatic Diseases.

So why do humans have so much back pain?

"Because we walk on two legs," said Jeremy DeSilva, a paleoanthropologist at Dartmouth University. Before humans began walking upright, our mammal ancestors had been running around on four legs for tens of millions, or even hundreds of millions, of years, he told Live Science. Mammals with this body shape have a horizontal spine that acts as a suspension bridge, holding up their torso.

About 7 million years ago, human ancestors evolved a more upright posture, DeSilva said. Their spine became vertical, allowing them to move around on two feet. Experts don't agree on why humans evolved to become bipedal, but one of the major theories is that it helped to transition from the jungles to the savanna. Although this adaptation helped humans flourish, it came with some costs.

"Because evolution can only work with pre-existing anatomies and pre-existing forms, we have this spine that evolution has tinkered with," DeSilva said. "And it's made it good enough. I mean, we're still here. But it doesn't mean we don't have problems. Evolution leads to being just good enough to survive. It doesn't lead to your comfort."

Bruce Latimer, a physical anthropologist at Case Western Reserve University in Ohio, described the spine as a series of cups (vertebrae) and saucers (disks between the vertebrae) balanced on top of each other. Most people have 24 of these cups and 23 disks. Ligaments and muscles help stabilize the stack, but because it's vertical, the disks are prone to slippage. 

"Humans are the only mammal that we know of that as we age, we can get spontaneous fractures of our vertebrae just from having that weight on top of each successive vertebra," DeSilva said.

The natural curve of the human spine also causes issues. The spine curves to balance weight, to allow for flexibility and to avoid blocking the birth canal. But because of this bend, people are susceptible to developing more severe curves, such as kyphosis (an outward curvature of the upper spine) or scoliosis (a lateral curvature of the spine), DeSilva said. At each curve, the spine is also prone to fractures.

Modern life in industrialized countries also plays a role. Core muscles stabilize the back, but many people have weak midsections. "If you're sitting at a desk all day, slouched over, and you're not working the lower back muscles, then they're easily strained," DeSilva said. 

Although there are multiple factors, evolution is the major culprit, DeSilva said. After all, our ancient ancestors, including the famous Australopithecus Lucy, had back problems, too, according to a 1983 study in the American Journal of Physical Anthropology. 

Not all bipeds have as much back pain as humans, however. Some large terrestrial birds, such as ostriches, walk upright on two limbs without much of an issue. 

"As far as I know, ostriches don't have to go to the chiropractor very often," DeSilva said. One reason why is that the bird's spine is more diagonal than vertical, so it can act more as a suspension bridge rather than a tower of cups and saucers. The ostrich also had significantly more time to evolve a high-functioning back. "They've had a roughly 200 million-year head start on us," DeSilva said. "When it comes to a bipedal skeleton, we're kind of the new kids on the block."

https://www.realclearscience.com/articles/2021/10/05/why_do_so_many_people_have_back_pain_797494.html

How Close Is Nuclear Fusion Power?

[This is a transcription of a YouTube video embedded in the link below]

By Sabine Hossenfelder

October 2, 2021 -- Today I want to talk about nuclear fusion. I’ve been struggling with this video for some while. This is because I am really supportive of nuclear fusion, research and development. However, the potential benefits of current research on nuclear fusion have been incorrectly communicated for a long time. Scientists are confusing the public and policy makers in a way that makes their research appear more promising than it really is. And that’s what we’ll talk about today.

There is a lot to say about nuclear fusion, but today I want to focus on its most important aspect, how much energy goes into a fusion reactor, and how much comes out. Scientists quantify this with the energy gain, that’s the ratio of what comes out over what goes in and is usually denoted Q. If the energy gain is larger than 1 you create net energy. The point where Q reaches 1 is called “Break Even”.

The record for energy gain was just recently broken. You may have seen the headlines. An experiment at the National Ignition Facility in the United States reported they’d managed to get out seventy percent of the energy they put in, so a Q of 0.7. The previous record was 0.67. It was set in nineteen ninety-seven by the Joint European Torus, JET for short.

The most prominent fusion experiment that’s currently being built is ITER. You will find plenty of articles repeating that ITER, when completed, will produce ten times as much energy as goes in, so a Gain of 10. Here is an example from a 2019 article in the Guardian by Phillip Ball who writes

“[The Iter project] hopes to conduct its first experimental runs in 2025, and eventually to produce 500 megawatts (MW) of power – 10 times as much as is needed to operate it.”

Here is another example from Science Magazine where you can read “[ITER] is predicted to produce at least 500 megawatts of power from a 50 megawatt input.”

So this looks like we’re close to actually creating energy from fusion right? No, wrong.

Remember that nuclear fusion is the process by which the sun creates power. The sun forces nuclei into each other with the gravitational force created by its huge mass. We can’t do this on earth so we have to find some other way. The currently most widely used technology for nuclear fusion is heating the fuel in strong magnetic fields until it becomes a plasma. The temperature that must be reached is about 150 million Kelvin. The other popular option is shooting at a fuel pellet with lasers. There are some other methods but they haven’t gotten very far in research and development.

The confusion which you find in pretty much all popular science writing about nuclear fusion is that the energy gain which they quote is that for the energy that goes into the plasma and comes out of the plasma.

In the technical literature, this quantity is normally not just called Q but more specifically Q-plasma. This is not the ratio of the entire energy that comes out of the fusion reactor over that which goes into the reactor, which we can call Q-total. If you want to build a power plant, and that’s what we’re after in the end, it’s the Q-total that matters, not the Q-plasma. 

 

Here’s the problem. Fusion reactors take a lot of energy to run, and most of that energy never goes into the plasma. If you keep the plasma confined with a magnetic field in a vacuum, you need to run giant magnets and cool them and maintain that. And pumping a laser isn’t energy efficient either. These energies never appear in the energy gain that is normally quoted.

The Q-plasma also doesn’t take into account that if you want to operate a power plant, the heat that is created by the plasma would still have to be converted into electric energy, and that can only be done with a limited efficiency, optimistically maybe fifty percent. As a consequence, the Q total is much lower than the Q plasma.

If you didn’t know this, you’re not alone. I didn’t know this until a few years ago either. How can such a confusion even happen? I mean, this isn’t rocket science. The total energy that goes into the reactor is more than the energy that goes into the plasma. And yet, science writers and journalists constantly get this wrong. They get the most basic fact wrong on a matter that affects tens of billions of research funding.

It’s not like we are the first to point out that this is a problem. I want to read you some words from a 1988 report from the European Parliament, more specifically from the Committee for Scientific and Technological Options Assessment. They were tasked with establishing criteria for the assessment of European fusion research.

In 1988, they already warned explicitly of this very misunderstanding.

“The use of the term `Break-even’ as defining the present programme to achieve an energy balance in the Hydrogen-Deuterium plasma reaction is open to misunderstanding. IN OUR VIEW 'BREAK-EVEN' SHOULD BE USED AS DESCRIPTIVE OF THE STAGE WHEN THERE IS AN ENERGY BREAKEVEN IN THE SYSTEM AS A WHOLE. IT IS THIS ACHIEVEMENT WHICH WILL OPEN THE WAY FOR FUSION POWER TO BE USED FOR ELECTRICITY GENERATION.”

They then point out the risk:

“In our view the correct scientific criterion must dominate the programme from the earliest stages. The danger of not doing this could be that the entire programme is dedicated to pursuing performance parameters which are simply not relevant to the eventual goal. The result of doing this could, in the very worst scenario be the enormous waste of resources on a program that is simply not scientifically feasible.”

So where are we today? Well, we’re spending lots of money on increasing Q-plasma instead of increasing the relevant quantity Q-total. How big is the difference? Let us look at ITER as an example.

You have seen in the earlier quotes about ITER that the energy input is normally said to be 50 MegaWatts. But according to the head of the Electrical Engineering Division of the ITER Project, Ivone Benfatto, ITER will consume about 440 MegaWatts while it produces fusion power. That gives us an estimate for the total energy that goes in.

Though that is misleading already because 120 of those 440 MegaWatts are consumed whether or not there’s any plasma in the reactor, so using this number assumes the thing would be running permanently. But okay, let’s leave this aside.

The plan is that ITER will generate 500 MegaWatts of fusion power in heat. If we assume a 50% efficiency for converting this heat into electricity, ITER will produce about 250 MegaWatts of electric power.

That gives us a Q total of about 0.57. That’s less than a tenth of the normally stated Q plasma of 10. Even optimistically, ITER will still consume roughly twice the power it generates. What’s with the earlier claim of a Q of 0.67 for the JET experiment? Same thing.

If you look at the total energy, JET consumed more than 700 MegaWatts of electricity to get its sixteen MegaWatts of fusion power, that’s heat not electric. So if you again assume 50 percent efficiency in the heat to electricity conversion you get a Q-total of about 0.01 and not the claimed 0.67.

And those recent headlines about the NIF success? Same thing again.  It’s the Q-plasma that is 0.7. That’s calculated with the energy that the laser delivers to the plasma. But how much energy do you need to fire the laser? I don’t know for sure, but NIF is a fairly old facility, so a rough estimate would be 100 times as much. If they’d upgrade their lasers, maybe 10 times as much. Either way, the Q-total of this experiment is almost certainly well below 0.1.

Of course the people who work on this know the distinction perfectly well. But I can’t shake the impression they quite like the confusion between the two Qs. Here is for example a quote from Holtkamp who at the time was the project construction leader of ITER. He said in an interview in 2006:

“ITER will be the first fusion reactor to create more energy than it uses. Scientists measure this in terms of a simple factor—they call it Q. If ITER meets all the scientific objectives, it will create 10 times more energy than it is supplied with.”

Here is Nick Walkden from JET in a TED talk referring to ITER “ITER will produce ten times the power out from fusion energy than we put into the machine.” and “Now JET holds the record for fusion power. In 1997 it got 67 percent of the power out that we put in. Not 1 not 10 but still getting close.”

But okay, you may say, no one expects accuracy in a TED talk. Then listen to ITER Director General Dr. Bigot speaking to the House of Representatives in April 2016:

[Rep]: I look forward to learning more about the progress that ITER has made under Doctor Bigot’s leadership to address previously identified management deficiencies and to establish a more reliable path forward for the project.

[Bigot]:Okay, so ITER will have delivered in that full demonstration that we could have okay 500 Megawatt coming out of the 50 Megawatt we will put in.

What are we to make of all this?

Nuclear fusion power is a worthy research project. It could have a huge payoff for the future of our civilization. But we need to be smart about just what research to invest into because we have limited resources. For this, it is super important that we focus on the relevant question: Will it output energy into the grid.

There seem to be a lot of people in fusion research who want you to remain confused about just what the total energy gain is. I only recently read a new book about nuclear fusion “The Star Builders” which does the same thing again (review here). Only briefly mentions the total energy gain, and never gives you a number. This misinformation has to stop.

If you come across any popular science article or interview or video that does not clearly spell out what the total energy gain is, please call them out on it. Thanks for watching, see you next week.

Posted by Sabine Hossenfelder

http://backreaction.blogspot.com/2021/10/how-close-is-nuclear-fusion-power.html

Restoring Vision After a Stroke

Vision loss can be a side effect from stroke. Neurons don't regenerate, and stem cell therapy is costly, difficult, and chancy. Researchers have figured out a way to use gene therapy to recover lost vision after a stroke in a mouse model.

From: Perdue University

October 2, 2021 -- Most strokes happen when an artery in the brain becomes blocked. Blood flow to the neural tissue stops, and those tissues typically die. Because of the locations of the major arteries in the brain, many strokes affect motor function. Some affect vision, however, causing patients to lose their vision or find it compromised or diminished. A research team led by Purdue University's Alexander Chubykin, an associate professor of biological sciences in the College of Science, in collaboration with the team led by Gong Chen at Jinan University, China, has discovered a way to use gene therapy to turn glial brain cells into neurons, restoring visual function and offering hope for a way to restore motor function.

Neurons don't regenerate. The brain can sometimes remap its neural pathways enough to restore some visual function after a stroke, but that process is slow, it's inefficient, and for some patients, it never happens at all. Stem cell therapy, which can help, relies on finding an immune match and is cumbersome and difficult. This new gene therapy, as demonstrated in a mouse model, is more efficient and much more promising.

"We are directly reprogramming the local glial cells into neurons," Chubykin said. "We don't have to implant new cells, so there's no immunogenic rejection. This process is easier to do than stem cell therapy, and there's less damage to the brain. We are helping the brain heal itself. We can see the connections between the old neurons and the newly reprogrammed neurons get reestablished. We can watch the mice get their vision back."

Chubykin's research is especially important because visual function is easier than motor skills to measure accurately, using techniques including optical imaging in live mice to track the development and maturation of the newly converted neurons over the course of weeks. Perfecting and understanding this technique could lead to a similar technique reestablishing motor function. This research bridges the gap in understanding between the basic interpretation of the neurons and the function of the organs.

https://www.sciencedaily.com/releases/2021/10/211002123006.htm

Cruel Left Wing Authoritarianism Is Thriving

Left-Wing Authoritarianism Is Real And Needs To Be Taken Seriously In Political Psychology, Study Argues

By Emma Young

Authoritarianism has been well-studied by psychologists. Well, right-wing authoritarianism has. In fact, as that’s typically the only type that’s studied, you might be forgiven for thinking that’s what authoritarianism is. The very idea of left-wing authoritarianism (LWA) has received not only little academic attention, but a lot of scepticism from psychologists. “I think I have not found any authoritarians on the left because if there ever were any, most of them have dried up and blown away….” wrote Bob Altemeyer, pioneer of work on right-wing authoritarianism, in 1996.

But as Thomas Costello at Emory University and colleagues write in their new paper in the Journal of Personality and Social Psychology: Personality Processes and Individual Differences, “From Maoist China to the Khmer Rouge (and perhaps even the French Reign of Terror), history abounds with examples of LWA at the broader societal level, rendering psychology’s inability to identify left-wing authoritarians puzzling.”

Puzzling is right. Perhaps predominantly left-leaning researchers have been unwilling to even go there…. But in their new paper, Costello and his colleagues absolutely go there. They conclude that LWA does indeed exist, and they define not only its characteristics but the characteristics of the people who subscribe to it. They also reveal substantial similarities between authoritarians on the political right and the left.

First, the team used a bottom-up approach to devise a scale to get at LWA. As well as scouring research papers for items that might relate to authoritarianism, they solicited ideas for items from psychologists, political scientists and philosophers. Using multiple batches of online participants, the team gradually whittled down and revised these items. They ended up with 39 items that reflected three conceptually distinct dimensions of LWA. I’ll take the definitions directly from the paper: 

• Antihierarchical Aggression – the belief that those currently in power should be punished, the established order should be overthrown, and that extreme actions, such as political violence, are justifiable to achieve these aims.
• Anticonventionalism – the rejection of traditional values, a moral absolutism concerning progressive values and concomitant dismissal of conservatives as inherently immoral, and a need for political homogeneity in one’s social environment.
• Top-Down Censorship – preferences for the use of governmental and institutional authority to quash opposition and bar offensive and intolerant speech.

Looking over that list, I certainly know some people who I’m sure think of themselves as being extremely liberal but who would score pretty highly on the anti-conventionalism dimension, at least.

The researchers then ran studies using fresh batches of participants and an impressive array of self-report scales to look at everything from personality, to mood, to cognition. They found a few differences between left-wing and right-wing authoritarians. People who scored highly on the LWA scale reported more negative emotions and were more neurotic than average (unlike RWAs). They were also more likely to report schadenfreude. The RWAs, meanwhile, scored higher for unjustified certainty in their beliefs and confirmatory thinking (a tendency to favour information that supports your beliefs). RWA was also more strongly linked to cognitive rigidity and low openness, as well as a lower than typical belief in science.

However, there were a lot more similarities between the two groups than differences. So much so, that it seems there is a shared constellation of traits “that might be considered the ‘heart’ of authoritarianism”, the team writes. These shared traits include (and again, I’ll quote directly from the paper): a “preference for social uniformity, prejudice towards different others, willingness to wield group authority to coerce behaviour, cognitive rigidity, aggression and punitiveness towards perceived enemies, outsized concern for hierarchy and moral absolutism.”

In terms of potentially dangerous implications for others, both left- and ring-wing authoritarianism was linked to the endorsement of political violence; but for the LWAs, that was violence directed at the state (violent protests, for example), whereas for the RWAs, it was in support of the state (supporting police crackdowns, say).

As well as analysing this self-report data, the team ran a study designed to look at LWA and actual behaviour. Online participants selected the difficulty level of a set of puzzles that they believed they were giving to another participant to complete. Before they chose the puzzles, though, they were shown what they were told was the Facebook profile of this ‘partner’. Those who’d scored highly for LWA ‘punished’ partners with right-wing profiles with harder puzzles and ‘helped’ those with left-wing profiles with easier puzzles. This was the case even when participants’ political ideology was taken into account, showing that LWA can be linked to actual aggressive behaviour towards the political outgroup above and beyond any effects of political ideology, the team concluded.

None of these studies were on groups of participants who were representative of a general population, however. So the researchers ran a fresh study on 834 who were selected to be representative of the population of the US. Again, they found large correlations between LWA scores and dogmatism, as well as schadenfreude, moral disengagement and violence towards out-groups. This included actual participation in the use of force for a political cause over the preceding five years. It also included support of violent group action during protests against police brutality towards Black Americans over the summer of 2020, in the wake of the murder of George Floyd. 

All in all, then, the team present a lot of studies — and a lot of analyses. There are some limitations to the work, however. Because of the nature of the scales, it’s not possible to tell whether right-wing authoritarians are more authoritarian than the left-wing type, or vice versa, say. Also, the route to LWA may be more well-meaning than the route to RWA, the researchers write; for example, someone who wants passionately to challenge inequality in society may end up thinking violent protest is the only option. “Any similarities across LWA and RWA notwithstanding, we endorse no claims of motivational or moral equivalence (or lack thereof) across the two constructs at present,” they write.

Their body of results surely suggests, though, that arguments that the search for LWA should be abandoned should themselves be abandoned. In fact, excluding left-wing features from studies of authoritarianism “has limited the kinds of knowledge we can produce as psychological scientists,” the team argues. And given the relevance of authoritarianism to politics, globally, and to how societies respond to and fight for change, surely it’s high time for that to shift.

Reference  Clarifying the structure and nature of left-wing authoritarianism.

Emma Young (@EmmaELYoung) is a staff writer at BPS Research Digest

https://digest.bps.org.uk/2021/10/01/left-wing-authoritarianism-is-real-and-needs-to-be-taken-seriously-in-political-psychology-study-argues/

Early Tau in Brain Predicts Alzheimer’s

Researchers found that early accumulation of tau proteins in the brain as measured by PET scanner was more effective at predicting memory impairment than biomarkers in the cerebrospinal fluid or amyloid plaque in the brain.

From: Karolinska Institutet [in Sweden]

September 30, 2021 -- Over 50 million people around the world suffer from dementia. Alzheimer's disease is the most common form of dementia and is characterised by an accumulation of the proteins beta-amyloid (Ab) and tau in the brain, followed by a continuous progression in memory decline. The pathological progression can take different forms and it is difficult to predict how quickly the symptoms will develop in any particular individual. Moreover, the presence of Ab in a person's brain -- known as amyloid plaque -- does not necessarily mean that the he or she will develop Alzheimer's dementia.

"There's been a rapid development of different Alzheimer's biomarkers in recent years, enabling us to measure and detect early signs of the disease in patients," says the study's first author Marco Bucci, researcher at the Center for Alzheimer Research, part of the Department of Neurobiology, Care Sciences and Society, Karolinska Institutet. "But we still need to find tests that can predict the development of the disease with greater specificity, so that we can improve not only its diagnosis but also its prognosis and treatment."

Some biomarkers identify accumulations of A? or tau, while others are used to measure the loss of nerve function (neurodegeneration). Protein accumulation and neurodegeneration can be measured in the cerebrospinal fluid (CSF) and plasma, or through brain imaging using positron emission tomography (PET) and magnetic resonance imaging (MRI). Current guidelines for the early detection of Alzheimer's disease with biomarkers endorse the interchangeability of brain imaging methods and analyses of CSF biomarkers (pTau and Ab), but this has been mooted. There is also a lack of longitudinal studies showing how the biomarkers are linked to gradual cognitive impairment.

"Our study shows that the presence of amyloid plaque in the brain and changes in concentrations of Ab and pTau in the CSF can be detected early during the course of the disease, but they do not seem to have any correlation with later memory loss," says Dr Bucci. "However, our results show that the presence of tau in the brain measured by a PET scanner is linked to a rapid decline, especially of the episodic memory, which is often affected at an early stage of the disease. Our observation suggests that tau PET should be recommended for the clinical prognostic assessment of cognitive decline in Alzheimer's patients."

The results are based on brain imaging (PET and MRI) and CSF analyses in a group of 282 participants comprising people with mild cognitive impairment, people with Alzheimer's dementia and healthy controls. 213 of the participants were also monitored for three years with tests of episodic memory (i.e. short term memory related to daily events).

"Our findings show that the concentration of tau in the brain in Alzheimer's disease plays an important part in its pathological progression and may become a key target for future drug treatments," says principal investigator Agneta Nordberg, professor at the Center for Alzheimer Research, Karolinska Institutet.

                      https://www.sciencedaily.com/releases/2021/09/210930213652.htm