Mind Matters Natural and Artificial Intelligence News and Analysis
Seeking Grants for an Association, a Small Business or for Research
Image licensed via Adobe Stock

Musk and LeCun Have a Superficial Debate About Science

What would have been a better debate?

Elon Musk tweeted the following:

“Join xAI if you believe in our mission of understanding the universe, which requires maximally rigorous pursuit of the truth, without regard to popularity or political correctness.”

Yann LeCun, chief scientist at tech giant Meta, could not resist responding. Musk claims to “want a maximally rigorous pursuit of the truth but spews crazy-ass conspiracy theories on his own social platform.”

It escalated quickly, with Musk questioning what science LeCun had done in the past five years, and LeCun replying: “Over 80 technical papers published since January 2022. What about you?” LeCun then said: “If you do research and don’t publish, it’s not science”.

So the most successful engineer over the last ten years criticizes academic science, and one of today’s most respected scientists, winner of the Turing Award, falls back on the number of papers he published? Haven’t most intelligent people agreed that the number of papers published, h-indexes, or journal impact factors are pretty much useless?

Nobel Laureates in physics, chemistry, and medicine agree. One says: “Most impact factors are measure of yearly average number of citations for papers published in previous two-year period, yet truly good papers often take many years to be cited, far too long to be reflected in impact factor,” or h-indexes. Another argues institutions should judge scientists based on intimate knowledge of their research, and not rely on journals they have published in.  

Unfortunately, most universities, funding agencies, and even professors still fall back on papers as their metric of performance. This determines promotion, tenure, and even funding. For instance, the research by the 2023 Nobel Laureate Katalin Kariko was shunned by the University of Pennsylvania because she couldn’t publish. So, she couldn’t get funding.

Science magazine tells the story of the rejection of her main paper with Drew Weismann in 2005. After being rejected by Nature within 24 hours, here’s what happened: “It was similarly rejected by Science and by Cell, and the word incremental kept cropping up in the editorial staff comments.”

Incremental? There are more than two million papers published each year and this research, for which Kariko and Weismann won a Nobel Prize, was deemed incremental? Because I argued that we should focus on ideas more than counting papers in a previous Mind Matters article, I will return to Yann LeCun because his comments are so indicative of how low our system of research has fallen.

Talk About Progress Made, Not Number of Papers Written

LeCun should have discussed the specific advances he made in artificial intelligence, not the number of papers. Those advances should have been the key issue in his debate with Musk. Just as Musk has successfully implemented his ideas for electric vehicles, satellite Internet, and many other technologies over the last decade, LeCun has done the same, and this should have been part of the debate.

Ironically, Musk’s electric vehicles are directly descended from research that won the 2019 Nobel Prize in chemistry, research that was done by two corporate and one academic researcher. That research was begun in the 1970s and led to the use of Li-ion batteries in laptop computers, mobile phones, and most recently in automobiles.

That is the flow of research to commercialization that Musk, LeCun, and others (such as the Wall Street Journal) should be focusing on, because we need more of that. For instance, why haven’t ultra-thin materials such as graphene, carbon nanotubes, superconductors for energy applications, bio-electronics of the Theranos kind, and quantum computers been successfully commercialized despite considerable hype about them over the last 20 years?

What science-based technologies did we get over the last 20 years? Organic light emitting diodes (OLEDs) and solar cells are somewhat successful, but they are a long way from the successes of the 1940s to 1970s such as transistors, polymers, fiber optics, radar, nuclear power, LEDs, and lasers.

Today’s Scientists Have It Backward

What’s different? Seventy years ago, scientists cared about developing something useful, a product, process, or just a good explanation of the world, much more than publishing papers. Today’s scientists have everything backward. Most could care less about products and processes that make the world a better place and more about increasing their publication count so they and their university can brag about those numbers.

Today’s leading scientists realize that things have changed, and Musk and LeCun should have added to this debate. Many scientists argue that the discovery of the DNA helix and other breakthroughs are no longer possible in today’s bureaucratic, grant-writing, administrative-burdened university research system. The idea of scientists following their hunches to find better explanations and thus better products and services has been replaced with huge labs pursuing grants to keep staff employed.

Even Nobel Laureates say such things, and not just Katalin Kariko. A biochemist, molecular biologist, and a physicist claim they could not get funding for their research in today’s emphasis on less risky projects; the physicist claims he could not get a job today. Another scientist turned policymaker argues that in today’s climate every project must succeed and thus people study only marginal, incremental topics where the path forward is clear, and you can virtually guarantee a positive result.

These scientists remember a time when no one counted the number of papers. Many of them knew what good research is when they saw it or a good idea when they heard it, partly because they grew up in a world in which problems were being addressed by kids on farms or kids fixing things in their houses. A history of Bell Labs claimed it hired mid-western farm boys because they had these skills.

I benefited from these types of professors in the 1970s when I was an undergraduate physics major. I remember talking with professors about many new technologies and their chances of success. Looking back, many of them were basically right, even though they were talking outside of their specialty, because they had a great deal of general knowledge, something that few of today’s professors have. But why have general knowledge when you are evaluated by the number of papers you publish in a narrow, esoteric journal?

That general knowledge meant that many scientists knew they had something when they made their breakthrough. For instance, Shockley, Brattain, and Bardeen knew they had something when they constructed the first workable transistor. So did Alexander Fleming when he discovered penicillin. The same holds with polymers, LEDs, and lasers.

This is not to say there weren’t surprises in the past. This 2018 article identified eight papers that were rejected before they eventually won their authors Nobel Prizes. But how many second chances would those papers and researchers get today? Research careers become stalled when papers are rejected. Just ask Katalin Kariko.

This is what Musk and LeCun should have talked about: How to turn advances in science into successfully commercialized technologies. How to survive in an environment of counting papers and calculating h-indexes. It appears, however, that Musk and LeCun aren’t much different from other scientists. Instead of rising above the superficial debates about science, they remain caught up in their own petty battles.

Jeffrey Funk

Fellow, Walter Bradley Center for Natural and Artificial Intelligence
Jeff Funk is a retired professor and a Fellow of Discovery Institute’s Walter Bradley Center for Natural and Artificial Intelligence. His book, Competing in the Age of Bubbles, is forthcoming from Harriman House.

Musk and LeCun Have a Superficial Debate About Science