In the popular imagination, the AI alignment debate is between those who say everything is hopeless, and others who tell us there is nothing to worry about.Leopold Aschenbrenner graduated valedictorian from Columbia in 2021 when he was 19 years old. He is currently a research affiliate at the Global Priorities Institute at Oxford, and previously helped run Future Fund, which works on philanthropy in AI and biosecurity. He contends that, contrary to popular perceptions, there aren’t that many people working on the alignment issue. Not only that, but he argues that the problem is actually solvable. In this podcast, he discusses what he believes some of the most promising paths forward are. Even if there is only a small probability that AI is dangerous, a small chance of existential risk is something to take seriously.AI is not all potential downsides. Near the end, the discussion turns to the possibility that it may supercharge a new era of economic growth. Aschebrenner and Hanania discuss fundamental questions of how well GDP numbers still capture what we want to measure, the possibility that regulation strangles AI to death, and whether the changes we see in the coming decades will be on the same scale as the internet or more important. Listen in podcast form here, or watch on YouTube.Links:* Leopold Aschenbrenner, “Nobody’s on the Ball on AGI Alignment.” * Collin Burns, Haotian Ye, Dan Klein, and Jacob Steinhardt, “Discovering Latent Knowledge in Language Models Without Supervision.” * Kevin Meng, David Bau, Alex Andonian, and Yonatan Belinkov, “Locating and Editing Factual Associations in GPT.”* Leopold’s Tweets: * Using GPT4 to interpret GPT2 .* What a model says is not necessarily what’s it’s“thinking” internally. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Bryan Caplan joins the podcast to talk about his new book, Voters as Mad Scientists: Essays on Political Irrationality.Bryan begins by explaining why he hates politics. Much of the conversation then centers around Caplan’s simplistic theory of the right and left. This is compared and contrasted with Scott Alexander’s thrive/survive theory of the political spectrum, Robin Hanson’s theory of farmers and foragers, and Hanania’s “Liberals Read, Conservatives Watch TV.”Near the end, the discussion turns to the political climate at GMU, and whether the intellectual community that has been built can survive the trend towards DEI. Caplan emphasizes that he has noticed a difference since Glenn Youngkin came to power in Virginia, showing that politics actually matters for determining the future of free speech and intellectual freedom.For previous Bryan appearances on the podcast, see: May 2021, September 2022, and May 2022.Listen in podcast form or watch on YouTube. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

This week we’re rereleasing a previous episode with Marc Andreessen, originally released on August 16, 2021. He is co-founder and general partner of Andreessen Horowitz. Earlier in life, he was the co-founder of Opsware, Ning, and Netscape. Marc joins the podcast to talk about what’s gone wrong with science, the prerequisites for progress, and how tech has changed our lives and has the potential to disrupt stagnant institutions. Topics also include how the internet has influenced dating, what venture capitalists actually do, and whether there is too much – or too little – money in politics.For a transcript of the conversation, see here. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Robin Hanson joins the podcast to talk about the AI debate. He explains his reasons for being skeptical about “foom,” or the idea that there will emerge a sudden superintelligence that will be able to improve itself quickly and potentially destroy humanity in the service of its goals. Among his arguments are:* We should start with a very low prior about something like this happening, given the history of the world. We already have “superintelligences” in the form of firms, for example, and they only improve slowly and incrementally * There are different levels of abstraction with regards to intelligence and knowledge. A machine that can reason very fast may not have the specific knowledge necessary to know how to do important things.* We may be erring in thinking of intelligence as a general quality, rather than as more domain-specific.Hanania presents various arguments made by AI doomers, and Hanson responds to each in kind, eventually giving a less than 1% chance that something like the scenario imagined by Eliezer Yudkowsky and others will come to pass. He also discusses why he thinks it is a waste of time to worry about the control problem before we know what any supposed superintelligence will even look like. The conversation includes a discussion about why so many smart people seem drawn to AI doomerism, and why you shouldn’t worry all that much about the principal-agent problem in this area.Listen in podcast form or watch on YouTube. You can also read a transcript of the conversation here.Links:* The Hanson-Yudkowsky AI-Foom Debate* Previous Hanson appearance on CSPI podcast, audio and transcript* Eric Drexler, Engines of Creation* Eric Drexler, Nanosystems* Robin Hanson, “Explain the Sacred”* Robin Hanson, “We See the Sacred from Afar, to See It the Same.”* Articles by Robin on AI alignment:* “Prefer Law to Values” (October 10, 2009)* “The Betterness Explosion” (June 21, 2011)* “Foom Debate, Again” (February 8, 2013)* “How Lumpy AI Services?” (February 14, 2019)* “Agency Failure AI Apocalypse?” (April 10, 2019)* “Foom Update” (May 6, 2022)* “Why Not Wait?” (June 30, 2022) This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Nicholas Bagley is a professor of law at the University of Michigan, former Chief Legal Counsel to Governor Gretchen Whitmer, and a former attorney in the US Department of Justice. He joins the podcast to talk about his article, “The Procedure Fetish,” in which he calls for liberals to embrace reforms to make federal government agencies less sclerotic and more capable of addressing social problems. Richard presents Bagley with questions surrounding issues such as why we should trust government agencies with more power, the role of cost-benefit analysis, the performance of the FDA during Covid-19, and civil service reform, including President Trump’s executive order that would have made it easier to fire more officials. The two discuss whether there can be a synthesis between the right and left on major issues surrounding government regulation.Listen to the podcast here, or watch on YouTube.Links: * Nicholas Bagley, “The Procedure Fetish”* Bagley on The Ezra Klein Show* Bagley on Twitter* Michael Lewis, The Fifth Risk* Matt Yglesias on Operation Warp Speed and the blowback to it* Cass Sunstein on the role of OIRA* Derek Thompson, “The Abundance Agenda” This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Tim Miller is a former political operative who has worked for Jeb Bush and John Huntsman, and is currently a writer for The Bulwark and an MSNBC analyst. He joins the podcast to talk about his political memoir, Why We Did It: A Travelogue from the Republican Road to Hell. With a former insider’s perspective, Miller discusses* Where the Republican Party went wrong* The importance of character in politics* Mistakes made by Clinton and George W. Bush that led us to this point* To what extent right-wing populists have legitimate grievances* The effect of the changing media environment on our fractious politics* Why only Chris Christie could have derailed Trump in 2016* Whether, to stop Trump, other candidates should get out of the way and support DeSantisThe discussion closes on whether there are reasons to be hopeful about the future of the Republican Party. Listen here or watch the video on YouTube. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Jobst Landgrebe is a German scientist and entrepreneur. He began his career as a Fellow at the Max Planck Institute of Psychiatry, then moved on to become a Senior Research Fellow at the University of Göttingen, working in cell biology and biomathematics. In April 2013, he founded Cognotekt, an AI based language technology company.Barry Smith is Professor of Philosophy at the University at Buffalo, with joint appointments in the Departments of Biomedical Informatics, Neurology, and Computer Science and Engineering. He is also Director of the National Center for Ontological Research and Visiting Professor in the Università della Svizzera italiana (USI) in Lugano, Switzerland. Landgrebe and Smith join the podcast to talk about their book Why Machines Will Never Rule the World: Artificial Intelligence without Fear. As the title indicates, the authors are skeptical towards claims made by Nick Bostrom, Elon Musk, and others about a coming superintelligence that will be able to dominate humanity. Landgrebe and Smith do not only think that such an outcome is beyond our current levels of technology, but that it is for all practical purposes impossible. Among the topics discussed are* The limits of mathematical modeling* The relevance of chaos theory* Our tendency to overestimate human intelligence and underestimate the power of evolution* Why the authors don’t believe that the achievements of Deep Mind, DALL-E, and ChatGPT indicate that general intelligence is imminent * Where Langrebe and Smith think that believers in the Singularity go wrong. Listen in podcast form or watch on YouTube.Links:* The Feynman Lectures on Physics* Landgrebe on Galactica and ChatGPT.* Rodney Brooks, “Intelligence without Representation.”* Nick Bostrom, Superintelligence. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Joe Henrich is the Ruth Moore Professor of Biological Anthropology and Professor of Human Evolutionary Biology at Harvard University. He is the author of Why Humans Cooperate, The Secret of Our Success, and The WEIRDest People in the World. He joins the podcast to talk about his work. Topics include:* The implications of Henrich’s theories for the debate over AI alignment* The nature of intelligence* Whether genetic differences between populations explain societal outcomes* If the Ancient Greeks and Romans were already WEIRD* How to understand the group selection debate* Why Islamic familial practices may have stunted economic development and growth* The political and ideological reaction to his last bookListen in podcast form or watch on YouTube. A transcript of the podcast can be found at the Richard Hanania newsletter.Links:* Joe Henrich, “The WEIRDest People in the World.”* Joe Henrich, “The Secrets of Our Success.”* Richard Hanania, “How Monogamy and Incest Taboos Made the West.”* David Epstein, “The Sports Gene.” * Seth Stephens-Davidowitz, “Don’t Trust Your Gut.” * Elizabeth Shim, “North Korea finishes fourth at International Mathematical Olympiad.” * Minnesota Transracial Adoption Study.* Bryan Caplan, “The Wonder of International Adoption: Adult IQ in Sweden.” This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Garett Jones is a Professor of Economics at George Mason University. He joins the podcast to talk about his new book, The Culture Transplant. Richard asks whether IQ is superior to other measures used to predict prosperity, and the relationship between Garett’s new book and Hive Mind. He also presses the author on whether there is a selection effect in data showing that people preserve the traits of their original culture over time. The conversation then gets into issues of causal inference, namely whether we should focus more on American history or cross-national trends to inform our understanding of US policy. Richard suggests that while immigration might in some contexts lead to larger government, in the US it is arguably the case that diversity has been a hindrance to the expansion of the welfare state. And how important is trust, actually? It correlates with a lot of good things, but how much is that relationship simply driven by observations from Scandinavia? Garett makes the case for trust having an important causal role. This leads to a discussion of whether trust is simply a proxy for trustworthiness, and whether the latter trait is more important. Garett also explains why Chinese migration could be a key force in lifting the third world out of poverty. Near the end, he discusses what he thinks America would look like after his preferred immigration policy, and what he’s working on next.Listen to the podcast here or watch on YouTube. Links:* Garett Jones on the Institutionalized podcast* Previous Jones appearance on the CSPI podcast* Alex Nowrasteh, critiques of The Culture Transplant, Part 1 and Part 2* Bryan Caplan review This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com

Alexander Young is a researcher at the UCLA Anderson School of Management Genomics Department and School of Medicine’s Human Genetics Department, working with the Social Science Genetic Association Consortium (SSGAC). He studies the genetics of cognitive ability and educational attainment, with a particular focus on developing methods to uncover true measures of heritability for important traits.Richard and Alexander talk about why siblings are so useful for this purpose, in the midst of a larger overview of the history of behavioral genetics and modern methods. Twin and adoption studies show much higher levels of heritability than genome wide association studies (GWAS). Why might this be the case? Different theories are discussed, along with ways to solve seeming discrepancies.The conversation goes on to cover the societal relevance of Alexander’s work, and attempts to isolate research on genes and cognitive ability within the academy.Listen in podcast form or watch on YouTube:Links:* Alexander’s Twitter account.* Alexander Young, “Solving the Missing Heritability Problem.”* Alexander Young and co-authors, “Deconstructing the Sources of Genotype-Phenotype Associations in Humans.”* James Lee, “Don’t Even Go There.” This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.cspicenter.com