STEM fuels thinkers; Wiley’s $23m AI windfall; Stanford stands down: Newsletter 19 June 2024

John Pellino/USMA

Newsletter 49. Why STEM education is about mind exercise, not job training. Plus, why AI is fair turnabout for journalists, ChatGPT is a bullshitter, and the Stanford Internet Observatory stops tracking the spread of lies. Plus, three people to follow and three books to read.

Training thinkers

For more than a decade, people have been advocating for more STEM (science, technology, engineering, math) education for students.

The ostensible reason is to interest more students in these fields, generating more qualified scientists, technologists, engineers, and mathematically qualified workers.

But the biggest problem that STEM education solves is not in the job market. The big problem is ignorance.

There is a huge challenge with ignorance among our citizens. People are ignorant about how science, logic, and statistics work. They are ignorant about biology, climate, health, economics, and (as became apparent in the last four years) infectious diseases.

An electorate unable to reason properly is ripe for manipulation. The rise of partisan opinion media, swallowed whole and then spread broadly by social networks, is possible because so many have ceased to be able to reason for themselves.

Fortunately, the history of science is rich with incredible stories that will capture the imagination of the most jaded observer.

I want students to understand how scientific observation changes opinions about the world. I want them to learn how Galileo was persecuted for observing the earth circling the Sun, contrary to church doctrine. How Newton explained gravity and mechanics — and then how Einstein proved his equations break down at near-light speeds, refining our understanding. Hypothesis, experimentation, proof, observation, and refinement are part of science. Learning how to think this way is beneficial in all of life: plumbers, nurses, shop supervisors, and voters all need to learn how to apply knowledge, form hypotheses, and change perspective when observations suggest improvements.

I want students to learn that what is unseen can still be tested and acted upon. They should learn how Pasteur observed and proved that unseen microbes caused infectious disease, and how Edward Jenner’s observation that milkmaids seemed immune to smallpox led to the first vaccine. They should learn how Watson, Crick, and Rosalind Franklin used X-rays to determine the hidden structure of DNA, and how viruses hijack its operation. They should learn how Darwin theorized and proved the way that evolution works, and how it continues to operate in our bloodstreams with transmission and mutation of pathogens.

I want students to learn how we connected the depletion of the ozone layer to refrigeration chemicals — and how we banned those chemicals and healed the earth. I want them to see how researchers proved that tobacco causes cancer, how the tobacco industry fought back, and how we as a society taxed and advertised cigarettes into submission.

I want students to learn how people devised clever experiments. How Ernest Rutherford bombarded gold foil with radioactive emissions and theorized the unseen structure of atoms. How Robert Millikan measured the minuscule charge on the electron by experimenting with oil drops.

And I want kids to code. I don’t want to them to write code so they can grow up to be software engineers. I want them to code so they can learn to think logically, design solutions, test them, find problems in them, and fix them. Software engineers don’t just whine about their code not working, the find what’s wrong and take responsibility for fixing it.

Finally, I want students to learn to understand and think rigorously about mathematics, probability, and statistics. They cannot understand the real world — from inflation figures to election polls to disease treatments — unless they understand statistics. I don’t care if they know the formula for the volume of a cone. I care a lot more if they understand what sample sizes are, what random variation is, and what a margin of error represents.

The objective of STEM education should be to teach people to think logically, devise experiments, solve problems, and be appropriately critical of what they read.

Ignorance is the enemy. An ignorant population expects others to solve their problems, and is susceptible to promises about easy solutions.

This stuff is actually pretty sexy. It’s not that hard to interest students in it. STEM education isn’t job training. It’s about teaching people to think. America needs more thinkers to survive and thrive in the next generation.

News for writers and others who think

Jeff Jarvis suggests a heretical notion: AI news search is just doing to news what news media have done to sources for a long, long time.

Three academics from the University of Glasgow published a paper in the journal Ethics and Information Technology that purports to show that ChatGPT generates bullshit, in the sense of the late philosopher Harry Frankfurt: that is, that is writes what it writes heedless of whether it is true. They aren’t the first to say it. The question is whether it’s a solvable problem, and from my interviews with experts, the answer appears to be: not completely.

The Washington Post’s Joseph Menn reports how Stanford’s Internet Observatory, which tracked the spread of disinformation online, is imploding under pressure from liars (gift link).

As Jane Friedman shared in her Hot Sheet, Barnes & Noble is removing some self-published books from BN.com. One problem is the effort required to distinguish legit self-published books from worthless auto-generated crap books.

The Guardian reports that authors increasingly have to pay to promote their own books. Among authors of non-fiction advice books, this is hardly news.

Publisher Wiley paid an unnamed large tech company $23 million to license its content for training purposes. According to Wiley interim president and CEO Matthew Kissner, “The onetime transaction to be recorded in Q4 includes access to previously published academic and professional book content for specific use in training LLM models.” Not mentioned: how did such a license address the rights of authors and copyright holders?

Three people to follow

Scott Bloom , a really witty guy who pokes fun at the irony of “thought leadership” ideas.

Phil Putnam, smartest coach on employee motivation you will ever meet.

Safi Bahcall , who understand how really big innovations happen, author of Loonshots.

Three books to read

On Call: A Doctor’s Journey in Public Service by Anthony Fauci (Viking 2024). What really happened with Trump and the pandemic — and dozens of other first-person stories of health crises from AIDS to Ebola.

How To Be Resilient: The Blueprint For Getting Results When Things Don’t Go To Plan by •Stacey Copas (Resilience for Results, 2015). A dynamic motivational speaker who comes with her own wheels.

A New Age of Reason: Harnessing the Power of Tech for Good by Larry Weber (Wiley, 2024). If you’re sick of technology horror stories, this stories in this book should make you feel a bit better.

Three inescapable reasons to get your author voice heard

Authors should stand up and be counted. You can do that now.

Click here to tell us about your experience in as short anonymous survey (by June 21). We’re trying to get to 600 authors by this Friday.

Here are three good reasons to click on that link.

  1. We know you’re annoyed with your publisher. Now you get a chance to tell the world — without anyone knowing who said it!
  2. It’s true. When they’re alone alone, authors whip out their . . . sales figures and compare how big they are. This way you can see how you stack up without anyone laughing at the size of your royalties.
  3. We’re doing a report on what choices make authors successful. If you contribute, you’ll be the first to see it . . . in less than a month.

This is your last chance. Click here by Friday. Costs you nothing but a few minutes of your time.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.