Category: Methodology

The Most Powerful A.I. Systems Still Do Not Understand, Have No Common Sense, and Cannot Explain Their Decisions

(p. B1) David Ferrucci, who led the team that built IBM’s famed Watson computer, was elated when it beat the best-ever human “Jeopardy!” players in 2011, in a televised triumph for artificial intelligence.

But Dr. Ferrucci understood Watson’s limitations. The system could mine oceans of text, identify word patterns and predict likely answers at lightning speed. Yet the technology had no semblance of understanding, no human-style common sense, no path of reasoning to explain why it reached a decision.

Eleven years later, despite enormous advances, the most powerful A.I. systems still have those limitations.

. . .

(p. B7) The big, so-called deep learning programs have conquered tasks like image and speech recognition, and new versions can even pen speeches, write computer programs and have conversations.

They are also deeply flawed. They can generate biased or toxic screeds against women, minorities and others. Or occasionally stumble on questions that any child could answer. (“Which is heavier, a toaster or a pencil? A pencil is heavier.”)

“The depth of the pattern matching is exceptional, but that’s what it is,” said Kristian Hammond, an A.I. researcher at Northwestern University. “It’s not reasoning.”

Elemental Cognition is trying to address that gap.

. . .

Eventually, Dr. Ferrucci and his team made progress with the technology. In the past few years, they have presented some of their hybrid techniques at conferences and they now have demonstration projects and a couple of initial customers.

. . .

The Elemental Cognition technology is largely an automated system. But that system must be trained. For example, the rules and options for a global airline ticket are spelled out in many pages of documents, which are scanned.

Dr. Ferrucci and his team use machine learning algorithms to convert them into suggested statements in a form a computer can interpret. Those statements can be facts, concepts, rules or relationships: Qantas is an airline, for example. When a person says “go to” a city, that means add a flight to that city. If a traveler adds four more destinations, that adds a certain amount to the cost of the ticket.

In training the round-the-world ticket assistant, an airline expert reviews the computer-generated statements, as a final check. The process eliminates most of the need for hand coding knowledge into a computer, a crippling handicap of the old expert systems.

Dr. Ferrucci concedes that advanced machine learning — the dominant path pursued by the big tech companies and well-funded research centers — may one day overcome its shortcomings. But he is skeptical from an engineering perspective. Those systems, he said, are not made with the goals of transparency and generating rational decisions that can be explained.

“The big question is how do we design the A.I. that we want,” Dr. Ferrucci said. “To do that, I think we need to step out of the machine-learning box.”

For the full story, see:

Steve Lohr. “You Can Lead A.I. to Answers, but Can You Make It Think?” The New York Times (Monday, August 29, 2022): B1 & B7.

(Note: ellipses added.)

(Note: the online version of the story was updated Sept. 8, 2022, and has the title “One Man’s Dream of Fusing A.I. With Common Sense.”)

“You Will Do Your Best Creative Work by Yourself”

(p. A12) The value of gathering to swap loosely formed thoughts is highly suspect, despite being a major reason many companies want workers back in offices.

“You do not get your best ideas out of these freewheeling brainstorming sessions,” says Sheena Iyengar, a professor at Columbia Business School. “You will do your best creative work by yourself.”

Iyengar has compiled academic research on idea generation, including a decade of her own interviews with more than a thousand people, into a book called “Think Bigger.” It concludes that group brainstorming is usually a waste of time.

Pitfalls include blabbermouths with mediocre suggestions and introverts with brilliant ones that they keep to themselves.

. . .

Plenty of people have always bemoaned brainstorming. Longtime Wall Street Journal readers may recall a 2006 “Cubicle Culture” column that skewered the popular practice, and Harvard Business Review published a research-based case against the usefulness of brainstorming in 2015.

. . .

Sometimes leaders bring employees together to create the illusion of wide-open input, says Erika Hall, co-founder of Mule Design Studio, a management consulting firm in San Francisco. In-person brainstorming is part of the back-to-office rationale for many of her clients, and she generally advises the ones that truly want to improve collaboration to first carve out some alone time for their workers.

When Hall needs inspiration, she goes for a run.

“It’s freaky,” she says. “I will go run on a problem, and things will happen in my head that do not happen under any other circumstance.”

Others might find “Aha!” moments in the shower or while listening to music. Leaving breakthroughs to private serendipity can feel, to bosses, like losing control, she acknowledges, but it might be more effective than trying to schedule magic in a conference room.

For the full commentary, see:

Callum Borchers. “ON THE CLOCK; Switch Off Brainstorming If You Want Brighter Ideas.” The Wall Street Journal (Thursday, May 18, 2023): A12.

(Note: ellipses added.)

(Note: the online version of the commentary was updated May 18, 2023, and has the title “ON THE CLOCK; Office Brainstorms Are a Waste of Time.”)

The book by Iyengar mentioned above is:

Iyengar, Sheena. Think Bigger: How to Innovate. New York: Columbia Business School Publishing, 2023.

“Cochrane Reviews Are Often Referred to as Gold Standard Evidence in Medicine”

The credibility of Cochrane reviews matters. One of their most important reviews, that I cite in my in-progress work on clinical trials, suggests that results of randomized double-blind clinical trials, usually agree with results of observational studies on the same topic. This matters a lot, because observational studies can give us more and quicker actionable results, saving lives.

(p. A23) Cochrane reviews are often referred to as gold standard evidence in medicine because they aggregate results from many randomized trials to reach an overall conclusion — a great method for evaluating drugs, for example, which often are subjected to rigorous but small trials. Combining their results can lead to more confident conclusions.

. . .

. . . what we learn from the Cochrane review is that, especially before the pandemic, distributing masks didn’t lead people to wear them, which is why their effect on transmission couldn’t be confidently evaluated.

For the full commentary, see:

Zeynep Tufekci. “In Fact, the Science Is Clear That Masks Work.” The New York Times (Saturday, March 11, 2023): A23.

(Note: ellipses added.)

(Note: the online version of the commentary has the date March 10, 2023, and has the title “Here’s Why the Science Is Clear That Masks Work.”)

In His Bathysphere Beebe “Maintained a Sense of Childlike Optimism”

(p. 28) Beautifully written and beautifully made, “The Bathysphere Book” is a piece of poetic nonfiction that strives to conjure up the crushing blackness of the midnight zone. Full color, overflowing with stunning illustrations of the uncanny creatures that live beyond the sun, it raises questions of exploration and wonder, of nature and humanity, and lets readers find answers on their own.

. . .

As he slipped deeper and deeper beneath the waves, Beebe bore witness to “a black so black it called his very existence into question,” and saw creatures that could be recorded only by describing them to Else Bostelmann, a painter who worked like a police sketch artist to render animals she would never see in colors like “bittersweet orange, metallic opaline green, orange rufous and orange chrome.”

. . .

. . . he maintained a sense of childlike optimism that pervades the book, cutting through the limitless cold of the sea: “Having traveled the world from the depths of the sea to the highest mountains, tramped through jungles and flown across continents, Beebe was more and more adamant that wonder was not produced by swashbuckling adventures — it was a way of seeing, an attitude toward experience that was always available. At every turn, the world’s marvels were right before our eyes.”

For the full review, see:

W. M. Akers. “Under the Sea.” The New York Times Book Review (Sunday, June 4, 2023): 28-29.

(Note: ellipses added.)

(Note: the online version of the review was updated May 31, 2023, and has the title “Deep-Sea Creatures of Bittersweet Orange and Metallic Opaline Green.”)

The book under review is:

Fox, Brad. The Bathysphere Book: Effects of the Luminous Ocean Depths. New York: Astra Publishing House, 2023.

Organisms Differ Based on How DNA’s Fluid Instructions Are Implemented by Different Cells

(p. 13) Martinez Arias, a developmental biologist, has lived and breathed the cell’s struggle to be heard over a career spanning 40-odd years. His story is one of DNA elites against hardworking, blue-collar cells. Cells, not DNA, Martinez Arias points out, determine the ripples of our fingerprints and the texture of our irises.

Martinez Arias builds his argument against the supremacy of DNA around Frankenstein-like experiments that involve borrowing a gene from one organism and dropping it into another. Take, for instance, the fruit fly PAX6 gene. When this gene is mutated, flies develop without eyes. Yet when a human version of PAX6 is swapped in for the fly gene, it makes a fly with fly eyes, not a fly with human eyes.

This is because fly cells are doing the work. Living things are much more fluid, Martinez Arias argues, than the concept of a DNA instruction manual would have us believe. An organism is less like a car, built according to a precise blueprint, he suggests, than a hobbyist’s renovation project, where the cells who live there build a deck or replace a light fixture based on the tools that happen to be lying around in the garage and whatever lumber is on sale at the store. Many of the differences between you and me are the result of accidents in time, enacted by our cells.

For the full review, see:

Alex Johnson. “Going Viral.” The New York Times Book Review (Sunday, September 17, 2023): 13.

(Note: the online version of the review has the date Aug. 15, 2023, and has the title “A Reason to Cheer for Cells and the Viruses That Feed on Them.”)

The book under review is:

Arias, Alfonso Martinez. The Master Builder: How the New Science of the Cell Is Rewriting the Story of Life. New York: Basic Books, 2023.

European Farmers Want Climate Protected by More Innovation, Not by Less Agriculture

(p. 4) To meet climate goals, some European countries are asking farmers to reduce livestock, relocate or shut down — and an angry backlash has begun reshaping the political landscape before national elections in the fall.

. . .

Those like Helma Breunissen, who runs a dairy farm in the Netherlands with her husband, say that too much of the burden is falling on them, threatening both their livelihoods and their way of life.

For almost 20 years, Ms. Breunissen has provided the Dutch with a staple product, cow’s milk, and she felt that her work was valued by society, she said. The dairy sector in the Netherlands, which also produces cheeses like Gouda and Edam, is celebrated as a cornerstone of national pride.

But the sector also produces almost half the Netherlands’ emissions of nitrogen, a surplus of which is bad for biodiversity. Ms. Breunissen and thousands of other farmers bridle that they are now labeled peak emitters.

“I was confused, sad and angry,” said Ms. Breunissen, who manages a farm of 100 cows in the middle of the country. “We are doing our best. We try to follow the rules. And suddenly, it’s like you are a criminal.”

. . .

In the Netherlands, the government has asked thousands of farmers to scale back, move or close. The authorities set aside about 24 billion euros, about $26 billion, to help farmers put in place more sustainable solutions — or to buy them out.

. . .

For Ms. Breunissen, who is 48 and works as a veterinarian in addition to her duties on the farm, none of the government-proposed options seem feasible. She is too young to quit and too old to uproot her life, she said, and the authorities have not provided enough support and information on how to change what she now does.

“There are so many questions,” she said. “The trust in the government is completely gone.”

. . .

A host of new groups are vying to displace traditional parties. They include the Farmer Citizen Movement, known by its Dutch acronym BBB, which was established four years ago.

. . .

Caroline van der Plas, the party’s co-founder, used to be a journalist in The Hague covering the meat industry, and she has never worked in farming. But she grew up in a small city in a rural area, and she said in an interview that she wanted to be “the voice of the people in rural regions who are not seen or heard” by policymakers.

She and her party have talked down the need for drastic steps to cut emissions, saying the reductions can be achieved through technological innovation. Policies should be based on “common sense,” she said, while offering no concrete solutions.

“It’s not like science says this or that,” Ms. van der Plas said, referring to how theories can change. “Science is always asking questions.”

For the full story, see:

Monika Pronczuk and Claire Moses. “New Climate Standards Have Farmers in Europe Bristling.” The New York Times, First Section (Sunday, Aug. 27, 2023): 4.

(Note: ellipses added.)

(Note: the online version of the story was updated Aug. 28, 2023, and has the title “Labeled Climate Culprits, European Farmers Rebel Over New Standards.”)

The “Woke-Mind” Is “Anti-Science, Anti-Merit and Anti-Human”

(p. 9) At various moments in “Elon Musk,” Walter Isaacson’s new biography of the world’s richest person, the author tries to make sense of the billionaire entrepreneur he has shadowed for two years — sitting in on meetings, getting a peek at emails and texts, engaging in “scores of interviews and late-night conversations.” Musk is a mercurial “man-child,” Isaacson writes, who was bullied relentlessly as a kid in South Africa until he grew big enough to beat up his bullies. Musk talks about having Asperger’s, which makes him “bad at picking up social cues.”

. . .

At one point, Isaacson asks why Musk is so offended by anything he deems politically correct, and Musk, as usual, has to dial it up to 11. “Unless the woke-mind virus, which is fundamentally anti-science, anti-merit and anti-human in general, is stopped,” he declares, “civilization will never become multiplanetary.”

. . .

The musician Grimes, the mother of three of Musk’s children (. . .), calls his roiling anger “demon mode” — a mind-set that “causes a lot of chaos.” She also insists that it allows him to get stuff done.

. . .

He is mostly preoccupied with his businesses, where he expects his staff to abide by “the algorithm,” his workplace creed, which commands them to “question every requirement” from a department, including “the legal department” and “the safety department”; and to “delete any part or process” they can. “Comradery is dangerous,” is one of the corollaries. So is this: “The only rules are the ones dictated by the laws of physics. Everything else is a recommendation.”

Still, Musk has accrued enough power to dictate his own rules. In one of the book’s biggest scoops, Isaacson describes Musk secretly instructing his engineers to “turn off” Starlink satellite internet coverage to prevent Ukraine from launching a surprise drone attack on Russian forces in Crimea. (Isaacson has since posted on X that contrary to what he writes in the book, Musk didn’t shut down coverage but denied a request to extend the network’s range.)

. . .

Isaacson believes that Musk wanted to buy Twitter because he had been so bullied as a kid and “now he could own the playground.”  . . .  Owning a playground won’t stop you from getting bullied.

For the full review, see:

Jennifer Szalai. “Self-Driving Czar.” The New York Times Book Review (Sunday, September 24, 2023): 9.

(Note: ellipses added.)

(Note: the online version of the review was updated Sept. 11, 2023, and has the title “Elon Musk Wants to Save Humanity. The Only Problem: People.”)

The book under review is:

Isaacson, Walter. Elon Musk. New York: Simon & Schuster, 2023.

Weight Loss Drugs Discovered Through “Tedious Trial and Error”

The first sentence quoted below implies that weight loss drugs are an exception in being discovered through trial and error rather than “through a logical process.” But I believe that drug discoveries in recent decades for cancer, heart disease, and Alzheimer’s also owe a lot to trial and error processes.

(p. A1) While other drugs discovered in recent decades for diseases like cancer, heart disease and Alzheimer’s were found through a logical process that led to clear targets for drug designers, the path that led to the obesity drugs was not like that. In fact, much about the drugs remains shrouded in mystery. Researchers discovered by accident that exposing the brain to a natural hormone at levels never seen in nature elicited weight loss. They really don’t know why, or if the drugs may have any long-term side effects.

“Everyone would like to say there must be some logical explanation or order in this that would allow predictions about what will work,” said Dr. David D’Alessio, chief of endocrinology at Duke, who consults for Eli Lilly among others. “So far there is not.”

. . .

(p. A16) . . . results from a clinical trial reported last week indicate that Wegovy can do more than help people lose weight — it also can protect against cardiac complications, like heart attacks and strokes.

But why that happens remains poorly understood.

“Companies don’t like the term trial and error,” said Dr. Daniel Drucker, who studies diabetes and obesity at the Lunenfeld-Tanenbaum Research Institute in Toronto and who consults for Novo Nordisk and other companies. “They like to say, ‘We were extremely clever in the way we designed the molecule,” Dr. Drucker said.

But, he said, “They did get lucky.”

. . .

After tedious trial and error, Novo Nordisk produced liraglutide, a GLP-1 drug that lasted long enough for daily injections. They named it Victoza, and the F.D.A. approved it as a treatment for diabetes in 2010.

It had an unexpected side effect: slight weight loss.

. . .

Finally, after liraglutide was approved in 2010 for diabetes, Dr. Knudsen’s proposal to study the drug for weight loss moved forward. After clinical trials, the F.D.A. approved it as Saxenda for obesity in 2014. The dose was about twice the diabetes dose. Patients lost about 5 percent of their weight, a modest amount.

. . .

Despite the progress on weight loss, Novo Nordisk continued to focus on diabetes, trying to find ways to make a longer-lasting GLP-1 so patients would not have to inject themselves every day.

The result was a different GLP-1 drug, semaglutide, that lasted long enough that patients had to inject themselves only once a week. It was approved in 2017 and is now marketed as Ozempic.

It also caused weight loss — 15 percent, which is three times the loss with Saxenda, the once-a-day drug, although there was no obvious reason for that. Suddenly, the company had what looked like a revolutionary treatment for obesity.

. . .

Researchers continue to marvel at these biochemical mysteries. But doctors and patients have their own takeaway: The drugs work. People lose weight.

For the full story, see:

Gina Kolata. “Medical Mystery Shrouds Drugs for Weight Loss.” The New York Times (Friday, August 18, 2023): A1 & A16.

(Note: ellipses added.)

(Note: the online version of the story has the date Aug. 17, 2023, and has the title “We Know Where New Weight Loss Drugs Came From, but Not Why They Work.”)

Improved AI Models Do Worse at Identifying Prime Numbers

(p. A2) . . . new research released this week reveals a fundamental challenge of developing artificial intelligence: ChatGPT has become worse at performing certain basic math operations.

The researchers at Stanford University and the University of California, Berkeley said the deterioration is an example of a phenomenon known to AI developers as drift, where attempts to improve one part of the enormously complex AI models make other parts of the models perform worse.

“Changing it in one direction can worsen it in other directions,” said James Zou, a Stanford professor who is affiliated with the school’s AI lab and is one of the authors of the new research. “It makes it very challenging to consistently improve.”

. . .

The goal of the team of researchers, consisting of Lingjiao Chen, a computer-science Ph.D. student at Stanford, along with Zou and Berkeley’s Matei Zaharia, is to systematically and repeatedly see how the models perform over time at a range of tasks.

Thus far, they have tested two versions of ChatGPT: version 3.5, available free online to anyone, and version 4.0, available via a premium subscription.

The results aren’t entirely promising. They gave the chatbot a basic task: identify whether a particular number is a prime number. This is the sort of math problem that is complicated for people but simple for computers.

Is 17,077 prime? Is 17,947 prime? Unless you are a savant you can’t work this out in your head, but it is easy for computers to evaluate. A computer can just brute force the problem—try dividing by two, three, five, etc., and see if anything works.

To track performance, the researchers fed ChatGPT 1,000 different numbers. In March, the premium GPT-4, correctly identified whether 84% of the numbers were prime or not. (Pretty mediocre performance for a computer, frankly.) By June its success rate had dropped to 51%.

. . .

The phenomenon of unpredictable drift is known to researchers who study machine learning and AI, Zou said. “We had the suspicion it could happen here, but we were very surprised at how fast the drift is happening.”

For the full commentary, see:

Josh Zumbrun. “THE NUMBERS; AI Surprise: It’s Unlearning Basic Math.” The Wall Street Journal (Saturday, Aug. 5, 2023): A2.

(Note: ellipses added.)

(Note: the online version of the commentary has the date August 4, 2023, and has the title “THE NUMBERS; Why ChatGPT Is Getting Dumber at Basic Math.”)

Bullshit Is Worse Than a Lie

(p. A17) Professor Frankfurt became best known for a single, irreverent paper largely unrelated to his life’s main work.

The paper, written in the mid-1980s under the same title as his eventual book, discussed what to his mind was a pervasive but underanalyzed feature of our culture: a form of dishonesty akin to lying but even less considerate of reality. Whereas the liar is at least mindful of the truth (if only to avoid it), the “bullshitter,” Professor Frankfurt wrote, is distinguished by his complete indifference to how things are.

Whether its purveyor is an advertiser, a political spin doctor or a cocktail-party blowhard, he argued, this form of dishonesty is rooted in a desire to make an impression on the listener, with no real interest in the underlying facts. “By virtue of this,” Professor Frankfurt concluded, “bullshit is the greater enemy of truth than lies are.”

. . .

For all this sang-froid, Professor Frankfurt was heartfelt in his philosophical pursuits. Throughout his career, he was drawn to lines of inquiry — about freedom, love, selfhood and purpose — that he said appealed to him not only as an academic but also “as a human being trying to cope in a modestly systematic manner with the ordinary difficulties of a thoughtful life.”

For the full obituary, see:

James Ryerson. “Harry G. Frankfurt, a Philosopher Eager to Cut the Bull, Dies at 94.” The New York Times (Tuesday, July 18, 2023): A17.

(Note: ellipsis added.)

(Note: the online version of the obituary has the date July 17, 2023, and has the title “Harry G. Frankfurt, Philosopher With a Surprise Best Seller, Dies at 94.”)

Frankfurt’s best-known book is:

Frankfurt, Harry G. On Bullshit. Princeton, NJ: Princeton University Press, 2005.

Scientists Had Political Motives for Dismissing Wuhan Lab-Based Covid Origin

(p. A17) On March 17, 2020, the journal Nature Medicine published a paper by five scientists, “The Proximal Origin of SARS-CoV-2,” that dismissed “any type of laboratory based scenario” for the origin of the pandemic. It was cited by thousands of news outlets to claim that the virus emerged naturally. But Slack messages and emails subpoenaed and released by the House Oversight Select Subcommittee on the Coronavirus Pandemic suggest that some of the authors didn’t believe their own conclusions. Before, during and even after the publication of their paper, they worried privately that Covid-19 was caused by a laboratory escape, perhaps even of a genetically engineered virus.

. . .

On April 16, a month after publication, Mr. Andersen wrote that “I’m still not fully convinced that no culture was involved” and “we also can’t fully rule out engineering”—i.e., that the virus not only was released from the lab but had been genetically manipulated there. He worried about the Wuhan lab’s research on live SARS-like viruses from bats at low biosafety levels: “it’s definitely concerning work, no question about it.”

So why did they publish a paper denying that laboratory origin was plausible? The answer may lie in their messages. In early February 2020, Mr. Rambaut wrote: “Given the s— show that would happen if anyone serious accused the Chinese of even accidental release, my feeling is we should say that given there is no evidence of a specifically engineered virus, we cannot possibly distinguish between natural evolution and escape so we are content to ascribing it to natural processes.”

Mr. Andersen replied: “I totally agree that that’s a very reasonable conclusion. Although I hate when politics is injected into science—but it’s impossible not to.”

. . .

To adjust the conclusions in a scientific paper for political reasons isn’t part of the scientific process. The world was misled with serious consequences.

For the full commentary, see:

Matt Ridley and Alina Chan. “The Covid Lab-Leak Deception.” The Wall Street Journal (Thursday, July 27, 2023): A17.

(Note: ellipses added.)

(Note: the online version of the commentary has the date July 26, 2023, and has the same title as the print version.)