Gary N. Smith

_{Senior Fellow, Walter Bradley Center for Natural and Artificial Intelligence}

Gary N. Smith is the Fletcher Jones Professor of Economics at Pomona College. His research on financial markets statistical reasoning, and artificial intelligence, often involves stock market anomalies, statistical fallacies, and the misuse of data have been widely cited. He is the author of dozens of research articles and 16 books, most recently, The Power of Modern Value Investing: Beyond Indexing, Algos, and Alpha, co-authored with Margaret Smith (Palgrave Macmillan, 2023).

Archives

A Sloppy “AI Scientist” Could Make the Science Crisis Much Worse

_{A research team claims to have developed the AI Scientist that “generates novel research ideas, writes code, executes experiments ...” Really?}_{Jeffrey Funk and Gary Smith
August 29, 2024}

It continues to astonish us how easily people can be persuaded that LLMs (chatbots) can do our thinking for us

Bad Luck Seldom Persists — But it Never Guarantees Good Luck

_{Many people embrace the fallacious law of averages in their daily lives when "regression toward the mean" is a more realistic picture}_{Gary Smith
August 21, 2024}

For example, the baseball player with the highest batting average in any season generally does not do as well the season before or the season after.

What the Luck? How Luck Matters to Olympic and Major League Wins

_{One way to think about the relative importance of skill and luck is to consider the consistency of the outcomes}_{Gary Smith
August 16, 2024}

Statistical analysis shows that in major sports, the players or teams that won most of their games benefitted to a surprising degree from good luck.

The Government-Debt Tipping Point Is Nonsense

_{There are serious problems with the economics paper by Reinhart and Rogoff, whose recommendations were widely followed}_{Gary Smith
August 9, 2024}

Reinhart/Rogoff’s dismissal of criticism of their deeply flawed study as “academic kerfuffle” is unconscionably cavalier.

How can someone who fails be faulted when everyone else is failing?

_{Market turbulence can cause endowment fund managers to travel with the herd — sacrifice returns in order to reduce annual volatility}_{Gary Smith
July 22, 2024}

The current, widely-favored 60/40 strategy has little or nothing to recommend it beyond the fact that it is what everyone else is doing.

AI Is Still a Delusion

_{Following instructions and performing fast, tireless, error-free calculations is not intelligence in any meaningful sense of the word}_{Gary Smith
June 27, 2024}

Not knowing what words mean, neither OpenAI’s ChatGPT 3.5 nor Microsoft’s Copilot nor Google’s Gemini can do a simple logic test.

LLMs Can’t Be Trusted for Financial Advice

_{The LLM responses demonstrated that they do not have the common sense needed to recognize when their answers are obviously wrong}_{Gary Smith
May 20, 2024}

It takes an experienced financial planner to distinguish between good and bad advice, so clients may as well skip the LLMs and go to the knowledgeable human.

A Man, A Boat, and a Goat — and a Chatbot!

_{Forty-five years ago, Douglas Hofstadter noted a key problem with AI: It can’t do the astonishing things our brains do, as chatbots reveal when asked to solve puzzles}_{Gary Smith
May 15, 2024}

Not understanding what words mean or how they relate to the real world, chatbots have no way of determining whether their responses are sensible, let alone true

Large Language Models Are Often Wrong, Never in Doubt

_{LLMs are statistically driven text generators, nothing more}_{Gary Smith
April 29, 2024}

As an economist, professor, and writer, I am reportedly in imminent danger of being replaced by large language models (LLMs). I’m not worried.

Universities Should Prioritize Critical Thinking Over Large Language Models

_{It is doubtful that LLMs can generate novel ways to boost productivity}_{Jeffrey Funk and Gary Smith
April 23, 2024}

No matter how many words they train on and no matter how many human trainers correct their mistakes, they will continue to lack the critical thinking skills.

Elon Musk: AI will be smarter than a human in 2025: Why he’s wrong

_{The superficial glibness of LLMs is a wonderful example of the adage that a little knowledge is a dangerous thing}_{Gary Smith
April 16, 2024}

Based on extensive training on untold amounts of text, LLMs are able to repackage superficially compelling answers that they literally do not understand.

The AI Hype Machine Just Rolls On, Living on Exhaust

_{Even chatbot enthusiasts, are starting to admit that scaling up LLMs will not create genuine artificial intelligence}_{Gary Smith
April 11, 2024}

Decades of geniuses trying to build computers that are as intelligent as they are have shown how truly remarkable our brains are—and how little we understand.

The Flea Market of the Internet: Breaking the Addiction

_{When, after a bad experience, I called Amazon the “Walmart of the Internet,” a friend pointed out that Amazon is, in fact, much worse than Walmart}_{Gary Smith
March 20, 2024}

Internet-based businesses tend to follow a life cycle in which quality deteriorates over time. Writer Cory Doctorow calls the process “enshittification.”

Sora: Life Is Not a Multiple-Choice Test

_{With Sora, as with other generative AI developments, some are quick to proclaim that artificial general intelligence has arrived. Not so fast.}_{Jeffrey Funk and Gary Smith
March 6, 2024}

The hallucinations are symptomatic of generative AI models’ core problem: they can’t identify output problems because they know nothing about the real world.

Retracted Paper Is a Compelling Case for Reform

_{The credibility of science is being undermined by misuse of the tools created by scientists. Here's an example from an economics paper I was asked to comment on}_{Gary Smith
February 23, 2024}

In my book Distrust (Oxford 2023), I recommend that journals not publish data-driven research without public access to nonconfidential data and methods used.

Why Chatbots (LLMs) Flunk Routine Grade 9 Math Tests

_{Lack of true understanding is the Achilles heel of Large Language Models (LLMs). Have a look at the excruciating results}_{Gary Smith
February 21, 2024}

Chatbots don’t understand, in any meaningful sense, what words mean and therefore do not know how the given numbers should be used.

Internet Pollution — If You Tell a Lie Long Enough…

_{Large Language Models (chatbots) can generate falsehoods faster than humans can correct them. For example, they might say that the Soviets sent bears into space...}_{Gary Smith
January 15, 2024}

Later, Copilot and other LLMs will be trained to say no bears have been sent into space but many thousands of other misstatements will fly under their radar.

Computers Still Do Not “Understand”

_{Don't be seduced into attributing human traits to computers.}_{Gary Smith
January 9, 2024}

Imagine people making decisions that are influenced by an LLM that does not understand the meaning of any of the words it inputs and outputs.

When it Comes to New Technologies Like AI, Tempers Run Hot

_{So far, the most tangible LLM successes have been in generating political disinformation and phishing scams.}_{Gary Smith and Jeffrey Funk
January 8, 2024}

LLMs often remind us of clueless students who answer essay questions by writing everything they think is relevant, hoping the right answer is in there somewhere

Let’s Dispose of Exploding Pie Charts

_{Pie charts are seldom a good idea. Here's why.}_{Gary Smith
January 2, 2024}

Points can be made without sensationalized graphs that undermine the credibility of the argument. Let’s dispose of exploding pie charts.