Manifold

Steve and Corey talk to Jason about a fundamental question of neuroscience: Do humans grow new neurons as adults? The dogma that humans do not, gave way to the dogma that they do, which is now being questioned. Adult neurogenesis has been associated with learning, better cognitive function and resistance to depression. Jason suggests that a simple error of treating young mice as models for adult humans led to excessive optimism regarding the potential for later neuronal growth. Recent findings suggest that adults grow few, if any, new neurons but that what little neurogenesis occurs can probably be enhanced by exercise.

Resources

• Transcript
• The Synder Lab
• Warren Sturgis McCulloch Interview (1969)

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Steve and Corey talk to Tim Searchinger about the unintended consequences of biofuels policies. Searchinger argues that these policies do not consider the opportunity costs of using plants for fuel rather than food. Combined with crazy carbon accounting principles, existing rules make cutting down trees in the US, shipping them to Europe and burning them in power plants count as carbon neutral under the Kyoto protocol. The three also discuss how eating less beef in the developed world along with educating women, family planning, and reducing child mortality in the developing world can decrease stress on land use and emissions.

Resources

• Transcript
• Creating a Sustainable Food Future: A Menu of Solutions to Feed Nearly 10 Billion People by 2050
• Timothy Searchinger

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Jamie Metzl joins Corey and Steve to discuss his new book, Hacking Darwin. They discuss detailed predictions for the progress in genomic technology, particularly in human reproduction, over the coming decade: genetic screening of embryos will become commonplace, gene-editing may become practical and more widely accepted, stem cell technology may allow creation of unlimited numbers of eggs and embryos. Metzl is a Technology Futurist, Geopolitics Expert, and Sci-Fi Novelist. He was appointed to the World Health Organization expert advisory committee governance and oversight of human genome editing. Jamie previously served in the U.S. National Security Council, State Department, Senate Foreign Relations Committee and as a Human Rights Officer for the United Nations in Cambodia. He holds a Ph.D. in Southeast Asian history from Oxford University and a J.D. from Harvard Law School.

Resources

• Transcript
• Hacking Darwin: Genetic Engineering and the Future of Humanity
• Jamie Metzl’s Personal Website

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Polymath and economist Tyler Cowen (Holbert L. Harris Professor at GMU) joins Steve and Corey for a wide-ranging discussion. Are books just for advertising? Have blogs peaked? Are podcasts the future or just a bubble? Is technological change slowing? Is there less political correctness in China than the US? Tyler’s new book, an apologia for big business, inspires a discussion of CEO pay and changing public attitudes toward socialism. They investigate connections between populism, stagnant wage growth, income inequality and immigration. Finally, they discuss the future global order and trajectories of the US, EU, China, and Russia.

Resources

• Transcript
• Personal Website
• Marginal Revolution [Blog]
• Conversations with Tyler [Podcast]
• Tyler Cowen | Bloomberg Opinion Columnist
• Big Business: A Love Letter to an American Anti-Hero

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Steve and Corey talk to Betsy McKay, senior writer on U.S. and global public health at The Wall Street Journal, about her recent articles on heart disease. Betsy describes how background reporting led to her article linking the recent drop in life expectancy in the United States, often attributed to the opioid crisis or increases in middle age suicides due to economic despair, to the increasing prevalence of heart disease, driven by the rise in obesity. The three also discuss current public health recommendations on how to reduce heart disease risk and on the use of calcium scans to assess arterial plaque buildup. Steve describes boutique medical programs available to the super-rich that include full body scans to search for early signs of disease. Betsy elaborates on how she approached reporting on a new study linking egg consumption to higher cholesterol and increased risk of death, a result at odds with other recent findings and national recommendations that two eggs a day eggs is safe and healthy. Finally, they consider whether people are wasting money on buying fish oil supplements.

Resources

• Transcript
• Death Rates Rising for Young, Middle-Aged U.S. Adults
• How to Reduce Your Risk of Heart Disease
• New BP guidelines that set elevated BP as above 120mmHG/80 and Stage 1 hypertension is 120-130/80-90, Stage 2 140/90 or above.
• New ACC/AHA High Blood Pressure Guidelines Lower Definition of Hypertension
• Heart Attack at 49—America’s Biggest Killer Makes a Deadly Comeback
• Study Links Eggs to Higher Cholesterol and Risk of Heart Disease
• Fish Oil: Hunting for Evidence to Tip the Scales
• Don’t Use Bootleg or Street Vaping Products, C.D.C. Warns

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Steve and Corey speak with Ted Chiang about his recent story collection “Exhalation” and his inaugural essay for the New York Times series, Op-Eds from the Future. Chiang has won Nebula and Hugo awards for his widely influential science fiction writing. His short story “Story of Your Life,” was the basis of the film Arrival (2016). Their discussion explores the scientific and philosophical ideas in Ted’s work, including whether free will is possible, and implications of AI, neuroscience, and time travel. Ted explains why his skepticism about whether the US is truly a meritocracy leads him to believe that the government-funded genetic modification he envisages in his Op-Ed would not solve the problem of inequality.

Resources

• Exhalation by Ted Chiang
• Stories of Your Life and Others by Ted Chiang
• Ted Chiang’s New York Times Op-Ed From the Future
• Transcript

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Dr. Rebecca Campbell is a Professor of Psychology at Michigan State University, whose research focuses on violence against women and children with an emphasis on sexual assault. Steve and Corey discuss her recent National Institute of Justice-funded project to study Detroit’s untested rape kits. Dr. Campbell describes the problem of untested kits and her work with police departments around the country to reduce the backlog. She explains how the use of the national CODIS database has led to sharply higher estimates of the proportion of rapes committed by serial perpetrators and how many rapists appear to be criminal “generalists”, committing a wide range of offenses. She describes the dynamics of sexual assault investigations, the factors that lead police to put more effort into investigating certain cases over others, and how common ways of questioning women can lead them to disengage from the process. Other topics include the incentives at work in law enforcement, the slow pace at which new research in DNA testing and treatment of victims is incorporated into police training, and Dr. Campbell’s efforts to engage with law enforcement agencies to improve investigative practices.

RESOURCES

Papers

• The National Problem of Untested Sexual Assault Kits (SAKs): Scope, Causes, and Future Directions for Research, Policy, and Practice.
• Forgotten evidence: A mixed methods study of why sexual assault kits (SAKs) are not submitted for DNA forensic testing.
• Why Police “Couldn’t or Wouldn’t” Submit Sexual Assault Kits for Forensic DNA Testing: A Focal Concerns Theory Analysis of Untested Rape Kits
• Tested at Last: How DNA Evidence in Untested Rape Kits Can Identify Offenders and Serial Sexual Assaults.
• Comparing Standard and Selective Degradation DNA Extraction Methods: Results from a Field Experiment with Sexual Assault Kits
• Evaluation of a Victim-Centered, Trauma-Informed Victim Notification Protocol for Untested Sexual Assault Kits (SAKs)
• Do You Wish to Prosecute the Person Who Assaulted You?: Untested Sexual Assault Kits and Victim Notification of Rape Survivors Assaulted as Adolescents
• Will history repeat itself? Growth mixture modeling of suspected serial sexual offending using forensic DNA evidence

Other

• Why Don’t Police Catch Serial Rapists? – The Atlantic
• MSU Appoints Presidential Advisers on Sexual Misconduct Issues
• Transcript

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Steve and Corey talk with Mark Moffett, Photographer and Research Fellow at the Smithsonian Institute, about his new book The Human Swarm: How our Societies Arise, Thrive and Fall. They discuss Mark’s view that being able walk into a cafe filled with others and not be attacked illustrates what makes human societies distinct and so successful. Mark explains why he is far more interested in questions about when war and other events occur than with traditional issues such as the genetic origins of human behavior. The three discuss Dehumanization and its Chimp equivalent, Dechimpanizeeization, and how they lead to the division of societies, friend turning against friend, and genocide. They discuss the conditions under which foreigners are embraced and whether the US might ever enter into a post-racial society where group differences don’t matter and immigrants are more easily accepted.

Resources

• Mark Moffett’s Bio
• Mark Moffett’s Photography
• The Human Swarm: How Our Societies Arise, Thrive, and Fall
• Transcript

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

John Schulman is a research scientist at OpenAI. He co-leads the Reinforcement Learning group and works on agent learning in virtual game worlds (e.g., Dota) as well as in robotics. John, Corey, and Steve talk about AI, AGI (Artifical General Intelligence), the Singularity (self-reinforcing advances in AI which lead to runaway behavior that is incomprehensible to humans), and the creation and goals of OpenAI. They discuss recent advances in language models (GPT-2) and whether these results raise doubts about the usefulness of linguistic research over the past 60 years. Does GPT-2 imply that neural networks trained using large amounts of human-generated text can encode “common sense” knowledge about the world? They also discuss what humans are better at than current AI systems, and near term examples of what is already feasible: for example, using AI drones to kill people.

Resources

• John Schulman
• OpenAI
• Better Language Models and Their Implications (GPT-2)
• Transcript

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.

Daniel Max, staff writer at The New Yorker and author of Every Love Story is A Ghost Story, a biography of David Foster Wallace, speaks with Corey and Steve about his first book, The Family that Couldn’t Sleep. The discussion covers the emerging genre of literary non-fiction, Daniel’s process of writing The Family that Couldn’t Sleep, and how he approached and gained the trust of the family at the heart of the story. Corey probes Daniel about how he handled the complex scientific characters, Carl Gajdusek and Stanley Prusiner, who led research into prion disease for 40 years. Daniel recounts how Shirley Glasse (now Lindenbaum) discovered how prions were transmitted through ritual cannibalism in Papua New, a critical step in solving the mystery of what causes of the disease, but how credit was given to Gajdusek. The three discuss the painfully slow pace of research and the inspiring story of a young couple, Eric Minikel and Sonia Vallabh, who have changed careers to dedicate their lives to finding a cure.

Resources

• Max’s New Yorker Page
• Max’s initial 2001 article for the New York Times Magazine on the Italian Family with FFI
• Max’s 2013 New Yorker story on Minikel and Vallabh
• The Family that Couldn’t Sleep
• Every Love Story is A Ghost Story
• Transcript

Announcing this for some friends at Mechanize - a startup that builds environments for training and evaluating frontier LLMs. Its customers include the top AI labs, and it has contributed to the breakthrough in coding capabilities of frontier models. 

Mechanize is hiring! 

https://mechanize.work/b/hsu 

Compensation is extremely competitive. For technical roles, $300-500k. They are also seeking smart generalists. 

For example: 

Research Engineer, Alignment: Build evals that test for misaligned model behaviors  $500K salary 

Puzzle Maker: Design interesting and original puzzles that LLMs can’t yet solve  $300K salary 

Mechanize understands that my readership is highly selected. There is a VERY GOOD CHANCE you will be interviewed if you apply via the link above.