I first met Tom Kalil when he was Deputy Director of the White House Office of Science and Technology Policy for the Obama Administration. Tom is one of those people who immediately strikes you as both immensely knowledgeable and incredibly practical – a thinker in constant search for the levers that will create the greatest positive impact for society through advancements in science and technology.
After a long career in government (he served in both the Clinton and Obama administrations) and academia, Tom joined Schmidt Futures as Chief Innovation Officer, where he leads initiatives to harness technology for societal challenges, improve science policy, and identify and pursue 21st–century moonshots.
In this first part of our interview, Tom and I discussed what Schmidt Futures is working on in the education and workforce development space, and more importantly Tom’s thoughts on what government innovation needs to happen to rapidly advance workforce development in the same way DARPA has advanced defense research. In Part II of our conversation, which will be published next month, we discuss how Schmidt Futures is introducing new ways to fund and advance scientific research.
Our conversation has been edited for length and clarity.
Tyton Partners (TP):
Can you give us a quick background on Schmidt Futures?
Tom Kalil (TK):
Schmidt Futures is a philanthropic initiative supported by Eric and Wendy Schmidt with a major focus on talent, identifying extraordinary individuals who are seeking to make the world a better place and making important advances in science and society.
Schmidt Futures is bringing ‘learning engineering’ to bear on how we teach math and science
You work on many different initiatives and our focus at Tyton Partners is really on education and workforce development. I want to hear about the work you are doing in that sector and how you are galvanizing people around different initiatives that you have put together.
There is an interesting initiative led by my colleague, Kumar Garg who I had an opportunity to work with when I was in the White House Office of Science and Technology Policy. He is leading an initiative called LEVI, which stands for the Learning Engineering Virtual Institute, which is supporting several teams seeking to double the rate of improvement in eighth-grade math for students that are not doing well. It’s also advancing the field of learning engineering.
One way to think about learning engineering is, “How would you combine computer science and the science of learning?” We are particularly interested in creating an infrastructure for continuous improvement. We see in K-12 education that productivity is negative. We doubled real pupil expenditures on K-12 education, and there have not been significant improvements in learning outcomes.
A question we’re interested in is, “Do we have the building blocks for an infrastructure for continuous improvement that includes:
- Instrumented blended learning environments
- Dozens of theories from the science of learning that are specific enough to inform instructional design
- More rapid feedback loops that involve the learner, the instructor, the course designer, and the science of learning
- Advances in data science and machine learning
- The ability to do rapid low-cost experimentation using things like A/B testing?”
For example, imagine creating a learning environment that gets better at not only identifying student misconceptions, but figures out what the right pedagogical strategy to address those misconceptions is? That way, online and blended learning environments would get better as more students use them and as researchers had the ability to do more A/B testing.
The untapped potential of science and technology to advance innovation in education and workforce development
You also do work beyond K-12, in getting people into high-paying jobs, especially those who do not have a college degree. Schmidt Futures has funded things like the $1 Billion Wage Gap Challenge, and I know this is a topic you’ve been thinking about for many years. I’d love to hear your thoughts around what’s next in this space, what needs to happen, and where do you see innovation playing a role?
One of the things that hit me over the head when I was in the government was that I would interact with some agencies like DARPA, where they had a budget of over $3.5 billion. The program managers used to joke that if you came to DARPA and served as a Program Manager for four years, and you didn’t do something like the equivalent of inventing the internet, then you got a B. They were working on developing aircraft that could go from New York to LA in 11 minutes and 20 seconds. They began to fund Moderna to work on vaccines for infectious disease in 2011, so when the pandemic hit, we were able to dramatically reduce the time to go from a new pathogen to an effective vaccine.
Other agencies have either little or no capacity to imagine and invent the future. The research budget of the Department of Labor is $0, and to the extent that they do anything they would call research it would be evaluating existing programs to see if they work.
I think that in the same way that in areas like national security, biomedical research, clean energy, space, and basic science we have well-funded efforts for mobilizing scientists, engineers, and entrepreneurs to solve problems in a new way, we should be doing the same thing with respect to economic and social mobility.
The thought experiment I’m interested in is, “Imagine you woke up and you were the Chief Technology Officer for the Department of Labor. And you had a budget, a team, and the ability to fund research and partner with the private sector and offer incentive prizes—the full suite of tools you would need to accelerate innovation. What goals would you set and what are examples of concrete projects that you would support to achieve those goals?”
One of the things we’ve seen is declining real wages for non-college-educated workers and research which suggests that workers that participate in our largest workforce development programs are net worse off as a result of having participated in them. Ideally, we would have workforce interventions that dramatically increased someone’s wages in a short period of time. An example would be, “I want to take a non-college-educated worker who is in a near minimum-wage job and give them a family-supporting wage and be able to do that in months rather than years.”
Why do we believe that’s possible? We’ll, there was a DARPA program, which had the name, “Education Dominance”. They were able to use AI-based digital tutors to model the interaction between an expert and a novice and take in new Navy recruits. Within six months, they were outperforming people who’ve been with the Navy for nine years in technical skills. So, you could imagine applying learning science and learning technology to reduce the time required to give someone a skill that is a ticket to the middle class.
The second thing that you could do is make it easier for employers to move away from a four-year degree in terms of making a hiring decision. The analogy that I use here is the flight simulator. Before we allow a pilot to fly a plane with 300 passengers, we have them interact with a flight simulator. Imagine creating task-based simulations that are predictive of on-the-job performance. I think it is great that some employers have been willing to move away from a four-year degree, we just saw the Governor of Pennsylvania do that for the entire state. My question is, “Clearly, employers are using the presence or absence of a degree for aspects of workforce quality that may be difficult to observe directly, so what do we replace that with, and can science and technology play a role?”
I’ll give you one more example. There is an idea where research is getting done in a university context. You have a distribution of workforce performance in a particular skill. Let’s say you run a hospital and you observe there are some neonatal nurses that just dramatically outperform their peers, such that there are much lower levels of mortality on the shifts that they are responsible for. Presumably, as an employer, what you’d like to know is what that nurse knows and is able to do better than the median or the lagging performance nurse. I would love to see a collaboration between the people who do cognitive task analysis and the CFO of the company, to make the economic case of the importance of those skills and the material impact they would have on the bottom line of the firm for upgrading the skills of the existing workforce.
Those are some examples of ideas I think would be interesting to explore if I woke up and I were the CTO of the Department of Labor and I wanted to have innovation in workforce development, in the same way that NIH advances biomedical research, and DARPA advances innovation in technologies that are important for the military.
I love these ideas. In the first Voices of Impact interview we published, I spoke with Marc Spencer from OneTen. What I like about this is that similarly, you’re talking about innovations on the supply side – training for the skills that are needed. But also on the demand side – actually solving for employers’ issues that arise from asking, as you pointed out, “If I get rid of the college degree as a requirement, what do I replace that with? How do I evaluate potential? Replacing something that for decades I’ve been using as a placeholder as a kind of proxy for that?” It excites me to see how both operators like OneTen and funders like Schmidt Futures are approaching the problem simultaneously from both directions.