Brown University Lecture Notes Part 2

The following is PART TWO of a writeup of my notes I prepared for a talk given to a multidisciplinary group of students and professors in hard sciences at Brown University. The original ask was to give a masterclass on some topic in music and math. However, there is an A.I. malaise hanging over both fields, I figured it would be good to address the topic. It was a long talk and so I’m splitting these lecture notes in to 2 parts. Part 1 gives a heuristic look at the the music industry in the 1900s and 2000s and how technology and law have shaped some of the underlying incentives. Part 2 takes the lessons of the influence of tech in Music and suggests how to understand our current mathematical moment. 

Thank you to Dr. Stephon Alexander Of The Physics Department for inviting me to speak.

Math

Ok what does all this have to do with math and science? Im sure you know this better than I do, but the discourse I’m hearing on the internet and from math friends feels alarmingly like the discourse I hear from disaffected musicians: 

“Math is about the joy and play of finding things out for which the human element is essential”, 

“Our work isn’t supposed to be practical; discovering the edges of knowledge and reason is a virtue unto itself” 

“Humans have been doing this forever so AI shouldn’t take it away from us” 

“You never know what math will be useful in time so its fine that we do a lot of ‘useless math’”

“We should just become interpreters of AI” 

“Human math is an indispensable part of human flourishing” 

“We should teach to use AI.”

“The pursuit of truth is its own reward whether it comes from AI or the human mind”

To use the lesson of the music industry, we would want know the motivations of whomever is cutting checks. A realignment would require knowing who the “record label” of basic research academia, (pure math, physics, chemistry, and biology) is. Unfortunately its not so simple as science research is funded at many different universities by a variety of public and private institutions. However there is one large organization that sets the science agenda for America and that is the National Science Foundation. The NSF provides 25% of funding of basic research in the US. While we are not looking at one incentive structure, my point will still stand using the NSF as an analogy for the rights holders in the music industry

So first lets learn about how the NSF started. The NSF was established in 1950 after a 1947 recommendation by Vannevar Bush, an MIT trained chemical engineer who submitted a report called The Endless Frontier to president Truman. Bush concluded that World War 2 was won by advantages in basic research including the physics of the atom bomb as well as medical advances the prevented soldiers in the field from dying of disease. He added that basic research would be key to maintaining an edge in national security, public health, and even the growth of private industry, which would take results from basic research and apply as needed. Finally, he was concerned that the major physicists involved in the Manhattan project all came from Europe and believed that the government should spend to train great scientists at home. 

His recommendation was the establishment of the National Science Foundation with a permanent endowment used to support basic research. Funds would be distributed through universities, rather than government appointed commissions, to grow universities as research hubs around the country. Scientists would be given the freedom to identify important areas of inquiry, rather than follow strict government priorities. In addition, money would be apportioned to scholarships and fellowships for talented students without the means to pay for advanced training in science. Lastly, effort would be made to find talented students from non-traditional backgrounds and outside academic circles.

The goals of the NSF have varied overtime given other national priorities or presidential whim. In the 70s there was a push to focus more on social problems than basic research and it has gone back and forth about whether engineering is a priority. The NSF has changed its relationship with congress and science policy making throughout its history. As such, we don’t have as clean of an incentive structure to reason with as we did in the music example. But we know advancing basic research for military, medical, and industrial edge is a top priority. 

Now consider the following scenario. Suppose you are the director of the NSF and a machine gets invented which generates the results of your math research priorities for a fraction of the cost at a fraction of the time. You don’t necessarily want mathematicians to lose their jobs. In fact you were one at earlier in your career, but the efficiency gains from the machine cannot be ignored. So you sit down to negotiate in good faith with the community, hoping to save as many jobs as possible. However, in your meeting, they try to convince you to continue funding using phrases like:

“Math is about the joy and play of finding things out for which the human element is essential”, 

“Our work isn’t supposed to be practical; discovering the edges of knowledge and reason is a virtue unto itself” 

“Humans have been doing this forever so AI shouldn’t take it away from us” 

“You never know what math will be useful in time so its fine that we do a lot of ‘useless math’”

“We should just become interpreters of AI” 

“Human math is an indispensable part of human flourishing” 

“We should teach to use AI.”

“The pursuit of truth is its own reward whether it comes from AI or the human mind”

Are you going to continue funding these people to the same level? Given your mandate and how many lives depend on it being met, are these the arguments you submit to? 

Unlike replaceable musicians justifying their price to a hard uncaring music industry, basic researchers have some negotiating power under the Forever Frontier Pact. In particular, if the mathematical community came to you and said, “Inumeracy/Stupidity is a national threat. We are a needed as a line of defense against brainrot because our future needs really smart people and tech is a poor teacher. In an increasingly complex world we need more subtle and sophisticated ways of building and thinking and preparing young people to do basic research is the best way we have to effectuate that.” You might find a few more millions in the budget. 

But what about frontier mathematics? These researches have worn their ‘uselessness’ as a humblebrag for years. If AI is getting their useless results fast I might just slash their budget and redirect to biology and chemistry which still need wet labs and humans to design experiments. But what if they came back to you and said “Frontier mathematics is not useless, just a little unfocused because of how vast it is. Our top thinkers have outlined a set of research programs that live at the cutting edge of the field. These programs comprise collections of problems across a few areas of mathematics whose resolution informs some issues in our sister sciences. And they are so attractive that we will continue to bring the best minds in the world to study in the USA. In designing and executing these programs, we are not just training mathematicians to solve problems and build theories, we are training TASTE in the most meaningful areas of research. Taste in knowing what problems to work on, how the web of mathematics fits together. Taste for now is still a human faculty and the one most connected to the NSF priorities.” 

Or what if they came to you and said “We have long had a problem with the number of unread papers in print and online. We actually may have more ideas than we need, they are just disorganized. Why don’t you grant us some money to survey, centralize, and classify the journals and the arXiv to make the corpus of knowledge more accessible to trained minds? We can call this the spring cleaning project.” Are these more compelling reasons for funding than “play” “exploration” “history” and “community”? 

SUMMARY

The point is that part of why AI as induced an existential crisis in the musical and mathematical communities is that we may not have had a clear enough understanding about how we were funded. It wasn’t necessary because the incentives were aligned (tolerated) in music for about 30 or 40 years, when musicians themselves were really important to making music. And the picture in math has been relatively stable since WWII since basic research became a federal priority. Incentives might have intentionally been obscured as both large labels and the government have reason keep musicians who are invested in the art over the commerce, or basic researchers who also have political and social opinions, on a need to know basis. In addition, both rode the wave of American prosperity and geopolitical dominance from the postwar to now

But as the paymasters’ incentives change with the technology, the structure of careers are likely to change as well. Both mathematicians and musicians have been hit pretty hard by AI. And understanding the incentives of the entities writing the checks is an awfully good place to start figuring out survive in the change. 

Mathematicians aren’t as misaligned to their payers incentives. This is where there is likely to be some leverage for the field, supposing the mathematical community can organize itself. Re-emphasize the need for humans to be trained in mathematics at a high level, even if we have a strong theorem proving machine. Bring research programs to the forefront of the discussion and talk to the other sciences to discover where the most important math is. Get good at sniffing out great questions that have broad implications and work with AI to solve them. Careers in mathematics don’t have to become another historical accident

Previous
Previous

Brown University STEM JAZZ Lecture Notes Part 1