The Context Window with David Deming Podcast

 This is part three of David Deming's three-part series on A.I. in education. In part one, David - labor economist and Dean of Harvard College - talked about how A.I. has changed college classrooms and what he and others are doing to protect the integrity of teaching and learning at Harvard. That was the strategy on defense. Part two was offense.

Today he pulls on one final thread that was left hanging in the two previous discussions, and it's the most controversial and contentious application of A.I. in the classroom: writing. Should we use A.I. to write for us? Should we use A.I. to improve our own writing? And if writing is thinking, will A.I. writing lead to cognitive offloading? Will it shrink our brains?

David found more than a dozen studies, mostly randomized controlled trials, that specifically study A.I. and writing. A.I. makes writing better. It's especially good at turning weak writing into competent conventional prose. That's not the same as making you better at writing.

He wants to be very clear: the evidence does not support, and he would not encourage you to use A.I. to write for you, at least if you want to retain any learning from what you've written. However, A.I. does make some parts of the writing process easier and more available. You can get individualized feedback, hear good counterarguments, consider alternative framings, and receive help with grammar and sentence construction.

Students deserve a thoughtful approach that is guided by evidence, not the extremes of a total ban or surrender. Rather than giving up on the term paper or banning A.I. and hoping for the best, we should intentionally incorporate it into the parts of the writing process where it can be helpful. For example, in giving honest, critical feedback. Acceptance or encouragement isn't enough.

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

In Part 2 of his series on A.I. in education, David Deming asks: Can we use A.I. to make learning better? And if so, how?

Last week, David discussed how A.I. has changed college classrooms at Harvard and around the country, and how we can protect the integrity of the learning process. That was defense against A.I. This week, he talks about offense.

When David dug into the research on A.I. and education almost a year ago, the evidence was pretty clear that A.I. without any guidelines, or guardrails, was harmful for learning because it makes things too easy. But since then, more than a dozen randomized controlled trials or other studies with good research designs have been published on the topic. Based on that evidence, he has become somewhat optimistic that A.I. really can improve learning, provided we can get beyond the current chaos of unfettered A.I. use and integrate it more thoughtfully into teaching.

David draws three key lessons from those studies: “think first, ChatGPT later”; use A.I. to personalize productive struggle; and build in scaffolding that makes it easy for students to use A.I. the right way and hard to use it the wrong way.

The status quo is a Wild West of A.I. use, often happening in secret and rarely with any supervision. But it doesn’t have to be that way, and the research I’ve read suggests we may have a path forward.

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

 Today’s topic is A.I. in the classroom, and it's a personal one for David. As Dean of Harvard College, he's been thinking and writing and talking with students a lot about A.I.: how it’s changed teaching and learning in college classrooms at Harvard and around the country, how we should adapt as teachers, and what students can do to prepare themselves for a workplace and a society where A.I. is everywhere - and they are expected to use it more and more just to keep pace.

His basic view is that we should pursue a mix of what he calls defensive and offensive strategies for dealing with A.I. The defensive strategy starts with the idea that the learning process and the struggle that sometimes comes with it need to be protected. That means A.I.-proofing assessments so that good performance truly demonstrates mastery of the material.

Unenforceable A.I. bans make it hard to do the right thing and easy to do the wrong thing, and that’s not fair to students. David argues that teachers need to design assignments and assessments better to protect the thinking that good writing requires. There’s a hopeful sense, in his view, in which A.I. will push educators to make the kinds of classroom improvements that they probably always should have been making anyway.

Next week, David will turn to offense: how to use A.I. to make learning better.

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

 David's guest today is ⁠Emily Oster.⁠ Emily is an economist at Brown University, author of four New York Times bestselling books, and founder of the parenting information platform, ParentData. Here books ⁠Expecting Better⁠ and ⁠Cribsheet⁠, which looked at how to interpret research on topics like alcohol consumption during pregnancy, breastfeeding, and sleep training, were genuine cultural phenomena, especially for anxious parents wondering where to turn for good advice.

In October 2020, when the world was still deep in the pandemic, she wrote an influential article in The Atlantic entitled "⁠Schools Aren't Super-Spreaders⁠," which was driven by her read of the early research on COVID-19's spread in schools. The CDC and then-Education Secretary Betsy DeVos cited her work as a reason to open schools earlier. Emily also built a data platform to try and understand the impacts of pandemic learning loss.

David wanted to talk with Emily about her road from wunderkind economist to bestselling author and public voice in health and parenting. They discuss the bumps along that road, what she learned from her mother about kindness and directness, and why she cares more about finding the answer than being right.

Parent Data: ⁠https://parentdata.org/⁠

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

 What happens when A.I. starts improving itself faster than humans can keep up? The researchers building these models can't fully explain how they work. If those models begin designing their successors, the things we don't understand could get baked into each new generation. How do we keep control?

In part two of his series on recursive self-improvement, David Deming focuses on alignment, the challenge of getting A.I. to do what we want. Human morality can't be reduced to a set of rules, and getting a model to behave well today doesn't guarantee it will ehave well as it becomes more capable.

He then turns to control: limiting what models can do, keeping the ability to shut them down, and building safeguards that assume failures will happen. From factory safety to A.I. agents that find ways around their instructions, David asks what it would take to contain the consequences when things go wrong.

He argues that A.I. labs should have to meet much higher safety standards before they let models improve themselves. Powerful machines need engineering safeguards and people who can be held accountable when those safeguards fail.

Part 1 of this series: https://thecontextwindowpodcast.substack.com/p/can-ai-get-better-at-making-itself

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

 Recursive self-improvement (R.S.I.) has gone from a nerdy, obscure concept to front-page news. That’s Bay Area slang for artificial intelligence using artificial intelligence to improve… artificial intelligence. Why does it have everyone so freaked out? And why is it linked so closely with fears of out-of-control superintelligent A.I.?

In part one of a two-part series about R.S.I., David Deming explains what R.S.I means, how A.I. models get trained, and what autonomous R&D looks like on the ground. He walks through the evidence from the frontier labs and academic researchers on how close we are to A.I. running its own research process.

He then also turns to the economics of R.S.I. A recent paper asks whether recursive self-improvement could cause A.I. capabilities to go vertical - what some people would call “takeoff.” David works through this argument and and tries to put together a speculative estimate of how close we might be. Spoiler alert, we're not there yet.

Part two will focus on the relationship between R.S.I. and alignment, which is the general challenge of getting the A.I. models to do what we want them to do.

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

 David's guest today is Jason Furman, an economist and professor at the Harvard Kennedy School, where they've been colleagues for the past several years. Before coming to Harvard, Jason had a distinguished career in government. He worked in the White House during the entire Obama administration, including as the chair of the President’s Council of Economic Advisers from 2013 to 2017, and he was one of the chief architects of the American Recovery and Reinvestment Act of 2009, which helped the country climb out of the depths of the Great Recession. Jason is now a contributing writer at The New York Times and the faculty director of the Mossavar-Rahmani Center for Business and Government here at Harvard, and he co-teaches EC 10, Introductory Economics, one of Harvard's largest courses.

David and Jason both love economics and have chosen careers in it, but they've come to it from totally different directions. Jason's experience is much more practical and hands-on than that of most other academic economists. David asks him about his career path into policymaking and politics, what he learned from his time in Washington, and what advice he would give to people who want to work in government, too.

David also ask him for a crash course in macroeconomics. Why is inflation bad? Why has it been so much higher over the past few years than everybody expected? And what can we do about it? What happens when governments try to inflate away their debt? And of course, I ask him about the economic impacts of AI and what we might expect in the years ahead.

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

How seriously should we take the risks of superintelligent A.I.? ⁠Matthew Yglesias⁠⁠ joins David Deming for an A.I. safety guide for normal people who are just paying attention.

Matt is a journalist and founder of the widely read newsletter ⁠Slow Boring⁠, which he started in 2020 after leaving the news site ⁠Vox⁠, which he co-founded with Ezra Klein and Melissa Bell in 2014. He is one of the masters of what some people call “⁠explanatory journalism⁠,” meaning going beyond just the facts and trying to help readers really understand the context around an issue. Matt has been writing about A.I. and related issues for a long time, but as he acknowledges in his column, A.I. safety has finally gotten the public’s attention, and the politics and policy of A.I. might move fast from here onward.

In this conversation, David asked Matt why A.I. safety is suddenly getting so much attention, how seriously we should be taking the risks of superintelligent A.I., and what it might realistically look like along the way to go from friendly chatbots on a computer screen to serious safety issues and even human extinction. They also talk about the politics of A.I. safety, on how to realistically slow things down with China also racing for the lead and whether Matt thinks a frontier A.I. lab could be nationalized in the next few years.

Read Matt's column, "A simple plan to save the world from rogue A.I.": ⁠https://www.slowboring.com/p/a-simple-plan-to-save-the-world-from ⁠

For show notes and full episodes, subscribe to The Context Window with David Deming: ⁠https://thecontextwindowpodcast.substack.com

David Deming's guest today is former U.S. Secretary of Commerce and Rhode Island Governor Gina Raimondo. Secretary Raimondo grew up in Rhode Island, where her father worked at the Bulova watch factory in Providence until it closed down, leaving him unemployed in his mid-50s. Today, she is trying to help workers avoid a similar fate as we transition to an AI economy. In June, she and former Governor of Indiana Eric Holcomb launched RAISE US, a national organization working with employers and states to prepare the American workforce move into new jobs.

David asked Secretary Raimondo how sure she is that widespread job losses are coming due to AI, how she plans to measure which initiatives are working, whether training programs can create jobs, and what obligations big companies and frontier AI labs have to help. The two also talk about wage insurance, job sharing, how to prepare for an uncertain future, including the advice she gives to her own children.

Then Secretary Raimondo gave David homework and challenged him to design a tax incentive for companies to invest in people. He took her up on it - and shares his proposal after the interview.

For show notes and full episodes, subscribe to The Context Window with David Deming: https://thecontextwindowpodcast.substack.com

In today's episode, David talks to Cate Hall about her book You Can Just Do Things: How High-Agency People Get What They Want Out of Life, written with Sasha Chapin. Agency is all the rage these days. Tech CEOs like Sam Altman have said that high agency is the skill they are hiring for in the age of AI. High agency people know what they want and they just get things done. But Cate's own life shows how difficult it can be to figure out what you want in the first place.

Cate was an incredibly high achiever from a young age. She double-majored in biochemistry and philosophy, graduated from Yale Law School, and was on the fast track with prestigious clerkships and a top job at a fancy law firm. But she was miserable. She quit law and became a very successful poker player, then struggled with drug addiction for about three years and lost everything. After rebuilding her life, she became co-CEO of a biotech company and later the CEO of the Astera Institute, a multibillion-dollar private foundation.

How did she do it? Her answer is agency. Cate shares advice on how to be more agentic in your own life, like staring into the abyss, communicating directly instead of expecting people to read your mind, and learning to love the moat of low status. She and David also discuss purpose, love, religion, and why willingly accepting constraints on your life can itself be an exercise of agency.

You Can Just Do Things, written with Sasha Chapin: https://www.catehall.com/the-book

Cate's newsletter, Useful Fictions: https://usefulfictions.substack.com/

For show notes and full episodes, subscribe to The Context Window with David Deming: https://thecontextwindowpodcast.substack.com

_

If you or someone you know is struggling with substance use, SAMHSA’s National Helpline offers free, confidential help and treatment referrals, and is available 24 hours a day. Call 1-800-662-HELP (4357) or visit FindTreatment.gov.