Nick Cassimatis: fear not AI, this too shall pass

  On this episode of Unsupervised Learning Razib talks to George Washington University archaeologist Eric Cline. The author of 1177 B.C. - The Year Civilization Collapsed, Cline has a new book out, After 1177 B.C. - The Survival of Civilizations. While 1177 B.C. closed with the end of the first global civilization, that of the Eastern Mediterranean at the end of the Bronze Age, After 1177 B.C. tells the story of those who picked up the pieces. But first Cline and Razib talk about the popular appetite for ancient history, and how 1177 B.C. became a surprise bestseller. Cline’s training is in archaeology and they discuss how new technologies like ancient DNA and isotope analysis are now contributing to transforming our understanding of the past. Then they turn to he organization of After 1177 B.C., how Cline decided to build on regional geographically focused histories rather than constructing a tightly integrated single narrative thread. This gets to the reality that the period covered in After 1177 B.C. is one of disintegration and isolation, as the networks binding together ancient Near Eastern kingdoms collapsed, with some states like that of the Hittites disappearing, and others like Egypt re-emerging sharply restructured.  Cline and Razib also discuss the lacunae in our understanding of the past, and the possibility that civilization may have gone through more cycles than we yet understand, with perhaps some social and technological complexity in the Pleistocene that we had not previously anticipated. Cline points out that Göbekli Tepe certainly must have had precursors in the Pleistocene, as local people could not have constructed such a site without skill and know-how accumulated over generations. If you have a sibling with autism, your future child’s risk for an autism diagnosis is increased by 2 - 3.5x. Orchid’s whole genome embryo reports can help mitigate your child’s risk by screening for 200+ genetic variants definitively linked to autism and other neurodevelopmental disorders. Discuss your situation with a genetics expert.  Special note: I’ve partnered with Roundtable to create a unique space dedicated to genetics and history enthusiasts. Our exclusive space is going to be organized into small, intimate roundtables of 4-6 people to ensure deep discussions and personal engagement. Every week, I'll provide an agenda for these discussions, and twice a month, I host an AMA for all members. If this opportunity excites you, apply here.  

On this episode of Unsupervised Learning Razib talks to Kristian Kristiansen, an archaeologist at the University of Gothenburg and affiliate professor at the Lundbech Center for Geogenetics, Copenhagen University. A past guest on this podcast, Kristiansen has recently contributed to an astonishing lineup of landmark papers published in Nature just in the last few months, Population genomics of post-glacial western Eurasia, Elevated genetic risk for multiple sclerosis emerged in steppe pastoralist populations, 100 ancient genomes show repeated population turnovers in Neolithic Denmark and The selection landscape and genetic legacy of ancient Eurasians. They also discuss his chapter in the 2023 book The Indo-European Puzzle Revisited: Integrating Archaeology, Genetics, and Linguistics. Razib and Kristiansen discuss the state of the emerging synthesis between archaeology, genetics and historical linguistics. Though himself an archaeologist, Kristiansen admits that in many ways historical linguists were correct, with models of mass migration now overturning those of cultural diffusion. He also gives a high-level summary of soon-to-be-published work on the spread of plague in Europe 5,000 years ago, and its role in the collapse of Neolithic civilization and the rise of steppe Indo-Europeans. Kristiansen gives a summary of recent developments in understanding the archaeology of Bronze Age Northern Europe, and in particular the expansion of the Corded Ware people. Razib and Kristiansen also discuss the role of distinct migration streams of the steppe people and their contribution to various Indo-European populations. Is it time to wonder if the Greeks descended from Corded Ware or Yamnaya?

  Today on Unsupervised Learning, Razib talks to long-time podcast favorite Samo Burja. Burja is the founder of Bismarck Analysis and Bismarck Brief, a Research Fellow at the Long Now Foundation and The Foresight Institute. He is also now the chair of the editorial board of Palladium Magazine. Already a four-time guest on Unsupervised Learning (he has previously shared his views on China's future, Russia's present and archaeology's past, his role at Bismarck Analysis and geopolitical uncertainty, reflected on his piece in Palladium on Finding "lost civilizations" and covered his ideas on "social technology," China, and the foreign view of America), the Slovenian-born Burja is one of the most original and incisive public intellectuals writing in America today. His 2021 piece, Why Civilization is Older than We Thought, brings a level of depth and rigor to historical heterodoxy that you rarely find anymore. Burja has also forwarded the “great founder theory” of historical change and formulated the idea of “live players” in social analysis. In this episode, Razib asks Burja for his sense of the world landscape in early 2024, revisiting conversations that delve into logistical details of the Russian invasion of Ukraine and the future of Chinese power. Burja continues to be pessimistic about the long-term prospects of European and Ukrainian resistance to a Russian war-machine that is geared toward grinding its way through lengthy battles of attrition. He also asserts that the current bearish attitude toward Chinese power is short-sighted, arguing that Western media in particular understates the technological and economic achievements of the PRC over the last generation. Burja believes that even if the “China bulls” were overly optimistic, the “China bears” go to excess in the opposite direction. Finally, he touches upon his vision for Palladium Magazine, a publication he has long contributed to, and which he now helms.

  On this episode of Unsupervised Learning Razib talks to Steve Hsu, physicist, entrepreneur and public intellectual. Hsu is an Iowan who earned his undergraduate degree from Caltech and his Ph.D. from Berkeley. Later he was a Harvard Junior Fellow, before moving on to professorships at Yale and the University of Oregon, and finally settling down at Michigan State University in 2012. Hsu is founder of Safeweb and Genomic Prediction, and his current focus is on a new AI startup. Between 2012 and 2020, he was vice president for research and graduate studies at Michigan State. Hsu also has a blog, Information Processing, and a podcast, Manifold. Razib asks Hsu about where cognitive and behavioral genomics are in 2024, and where they are going. They discuss the reality that while study of educational attainment (EDU) has proceeded relatively far, the study of intelligence itself has been neglected. Hsu outlines the case for why cognitive phenotypes should be studied, even if the topic remains controversial and fraught. They then address the current fad for artificial intelligence, and how emerging companies in the space are going to transform workplace productivity and culture. Hsu contrasts the rapid pace in the advancement in AI with the torpidity of behavioral genomics. Finally, Razib and Hsu discuss changes in academia in the wake of the “Great Awokening” that led to his ouster in 2020 from his role as vice president of research and graduate studies at Michigan State. Hsu talks about how his work in IQ and genetics became weaponized on social media by left-wing graduate students during the George Floyd protests.

Today Razib talks to Murtaza Hussain about the social, cultural and political context of recent fissures in the US around the conflict in Israel and Gaza. Hussain is a reporter at The Intercept and has his own Substack. They begin their conversation talking about Hussain’s response to the 10/7 Hamas attacks on Israel, and Israel’s subsequent invasion of Gaza. Hussain discusses his bewilderment and disappointment at some commentators who he saw being knee-jerk and tribalistic in their response. He also talks about the generational divides on Israel that have become apparent: while American Boomers and Gen-X tend to support the Jewish state overwhelmingly, while Millennials, and especially Zoomers, are more divided, or perhaps even a pro-Palestinian, perspective. Hussain, a Pakistani Canadian Millennial, though now a naturalized US citizen, does not approach the subject of the Israel/Palestine conflict with a reflexive sympathy for any particular side, but does believe that many Americans are unaware of the broad support that Palestinian nationalism attracts worldwide and especially among the youth. Razib and Hussain then discuss the intellectual history that led up the conflict between Hamas, an Islamist movement, and the state of Israel, and how the Palestinian national struggle is positioned within the global Left. They discuss the various connections between Palestinian nationalism in the 1960’s and 1970’s and the Soviet Bloc and Left-wing national liberation movements like the IRA, and how that might impact sympathies of activists. Hussain contends that in many ways the Palestinian liberation movement is a leftover of 20th-century struggles, with the end of apartheid South Africa and the peace accords in Ireland. He outlines the multi-decade relationship both the Palestinians and Israelis have had with the non-aligned movement and postcolonialism. Though today Israel is coded as a Western nation (they participate in Eurovision), Hussain notes that as Prime Minister, David Ben Gurion attempted to send Israeli delegates to the non-aligned Bandung Conference in 1955, a move that was blocked by Egypt. 

On this episode of Unsupervised Learning Razib welcomes back paleoanthropologist Chris Stringer. Affiliated with the Natural History Museum in London, Stringer is the author of African Exodus. The Origins of Modern Humanity, Lone Survivors: How We Came to Be the Only Humans on Earth and  Homo Britannicus - The Incredible Story of Human Life in Britain. A proponent since the 1970’s of the recent African origin of modern humans, he has also for decades been at the center of debates around our species’ relationship to Neanderthals. In the 1980’s, with the rise to prominence of the molecular model of “mtDNA Eve,” Stringer came to the fore as a paleoanthropological voice lending support to the genetic insights that pointed to our African origins. Trained as an anatomist, Stringer asserted that the fossil evidence was in alignment with the mtDNA phylogenies, a contention that has been broadly confirmed over the last five decades.  But in 2010, Stringer and other proponents of an “out of Africa” “with replacement” model of recent human origins began to modify their views in response to the mounting evidence of archaic admixture, the introgression of Neanderthal and Denisovan genes into the modern human genome. On this episode, Razib queries Stringer on the state of human evolution from the fossil’s-eye view in 2024. They discuss “Dragon Man,” and whether this is just a fossil of one of the Denisovan populations. Razib also presss Stringer about the diversity of human species in Southeast Asia, and just how many Denisovan populations or “races” might have existed. They also touch on Homo naledi, and the ensuing controversies around naledi-related publications. Razib seeks Stringer’s opinion on different models of African origins for our lineage, from extensive archaic admixture to “African multi-regionalism.” On a more speculative note, they mull over the possibilities for complex societies in the Pleistocene in light of the finds at Göbekli Tepe. With Stinger’s over five decades in the discipline, very few rival his qualifications or capability to provide a bird’s-eye view of where we are in understanding human evolution in 2024.

On this episode of Unsupervised Learning, Razib talks to Zoe Booth and Iona Italia. Booth is community engagement officer and Italia is managing editor at Quillette. An Australian, Booth has degrees in French, Politics and Law from the University of Newcastle. Italia is an erstwhile academic of British nationality and mixed Parsi and Scottish heritage, with a Ph.D. in English literature from Cambridge University. She is the author of Our Tango World, former editor-in-chief of Areo Magazine and the host of the Two for Tea Podcast. Razib discusses both of their trajectories into the heterodox intellectual sphere, Booth, from her starting point as a younger Millennial and Italia as a member of Generation X. While Booth recounts she had typical generational views on social justice and left-inflected politics, Italia admits despite being very left-wing most of her life she was never very well disposed to the identitarian trend that has crystallized into “woke” politics in the 2020s. Booth also addresses the reality that even if the existence of Quillette, a female-led bastion of free thought, with founding editor Claire Lehmann and now managing editor Italia might seem to suggest otherwise, it is not always easy to be a heterodox woman. Booth and Italia discuss how female personality orientation tends more toward making people feel comfortable and included rather than confrontations over truth claims that might hurt feelings. Italia and Razib also address her unique personality quirk of very high disagreeability, which might explain both her rejection of group-think and her earnest quest for the truth as she understands it. Booth and Italia talk about how the recent events around the Gaza war between Israel and Hamas, have resulted in changes in their social life due to political polarization. Overall Quillette has taken a pro-Israel position across the editorial staff, which has resulted in some blowback among their readership. Italia also talks about her own change from solidarity to the global left because of their Hamas-friendly stance, and her continued rejection of conservative social movements, including Islam. Booth and Italia also address Quillette’s consistent trend of touching cultural and political third rails, but in the service of classical liberal values. Italia believes any blowback toward her and the magazine comes disproportionately from a small group of malcontents, and that broadly liberal values are much more popular than most people realize.

On this episode of Unsupervised Learning Razib  talks to Nick Cassimatis, erstwhile artificial intelligence researcher and currently an entrepreneur. Cassimatis has undergraduate and doctoral degrees in cognitive and computer science from Massachusetts Institute of Technology, and a master’s degree in child psychology from Stanford. He studied for his Ph.D. under Marvin Minsky, arguably the most prominent and influential artificial intelligence researcher of the second half of the 20th century. Later, Cassimatis was a professor at Rensselaer Polytechnic Institute, founder of a successful startup, and a researcher at Yahoo and Samsung. Because of the explosion of large language models as implemented in OpenAI’s ChatGPT and Google’s Gemini, we are now living through an artificial intelligence “hype cycle.” But Cassimatis observes that this is not the first time this has occurred. The 1960’s saw enthusiasm triggered by the ELIZA therapist chatbot. Then, in the 1990s another wave of interest crested because of the mastery of chess by Deep Blue. Finally, there was a boom of excitement around artificial intelligence after Watson’s victory in Jeopardy in the early 2010s. But these hype cycles also have their equivalent troughs; Cassimatis recounts that when he went to study artificial intelligence in the early 2000s, many people discouraged him because the field’s allure had cooled considerably. And yet, under Minsky he developed an interest in how computers could learn, writing papers like A cognitive substrate for achieve human-level intelligence. This background makes Cassimatis a particularly well-informed and trenchant observer and analyst of the current arguments about the possible emergence of artificial general intelligence in the next decade, and what it means for the future of humanity. But first, Cassimatis and Razib step back and address some basics. What is machine learning? How does this relate to deep learning and natural language processing? What are transformers and what is a neural network? These are terms that are thrown around casually in the technology press, but these concepts emerge from over fifty years of research in computer science. With those preliminaries out of the way, Razib probes Cassimatis’s opinions about the past and future of large language model-driven artificial intelligence, and the probability of Ray Kurzweil’s “intelligence explosion” soon. Cassimatis believes it is likely that this hype cycle will eventually fade and suspects that large language models may run up against their limits very soon. He suggests that since ChatGPT’s release in the fall of 2022, the massive transformations predicted in our lives have not come to pass more than a year later. It has not, for example, replaced search on the web, nor has it revolutionized software engineering. And it is the last issue, the impact of artificial intelligence and advances in computing that underpin Cassimatis’ current start-up, Dry.Ai, a platform for developing applications in a no and low-code framework. The enablement of faster and more productive programming frameworks like GitHub Copilot over the last few years has prompted some to wonder if  a crash in demand for engineers is in the offing, with a smaller number of far more productive workers. Cassimatis reminds us that in the early days of high-level programming languages, like Perl, Python or Java, the same argument was mooted. And yet, on the contrary, the demand for developers has remained high. Cassimatis expects  in the near future to see artificial intelligence hitched up to platforms like Dry.Ai which will make programming easier, reducing the time from conception to final release of an application. Overall, he sees a future that is more technologically advanced, but he does not anticipate that the next generation will bring the revolutionary transformation of all life as we know it. For the first time ever, parents going through IVF can use whole genome sequencing to screen their embryos for hundreds of conditions. Harness the power of genetics to keep your family safe, with Orchid. Check them out at orchidhealth.com.

  On this episode of Unsupervised Learning Razib talks about AI, the singularity and the post-human future, with James D. Miller, a Smith College economist, host of the podcast Future Strategist and the author of Singularity Rising: Surviving and Thriving in a Smarter, Richer, and More Dangerous World. Miller and Razib first met at 2008’s “Singularity Summit” in San Jose, and though Singularity Rising was published in 2012, some of the ideas were already presented in earlier talks, including at that conference. More than 15 years since Miller began formulating his ideas, Razib asks him how the theses and predictions in his book have held up, and how they compared to Ray Kurzweil’s The Singularity is Coming. On this last point, Miller is very bullish on Kurzweil’s prediction that artificial intelligence will surpass that of humans by 2030. He also believes that the “intelligence explosion,” Kurzweil’s “technological singularity” when AI transforms the earth in unimaginable ways through exponential rates of change will in fact come to pass. But while Kurzweil predicts that the singularity will usher in an era of immortality for our species, Miller has a more measured take. He believes AI will drive massive gains in economic productivity, from cultural creativity to new drug development regimes (one of the original rationales behind IBM’s AI program). But while Kurzweil anticipates exaltation of conscious human life into an almost divine state, Miller suspects that AI may eventually lead to our demise. He estimates a 10% probability that Kurzweil is correct that we will become immortal, and a 90% probability that AI will simply shove us aside on this planet as it begins to consume all available resources. Overall, Miller is satisfied with the predictions in the first third of Singularity Rising. Computational technology has become far more powerful than it was in the late aughts, with a supercomputer in everyone’s pocket. Though the advances in AI seem to exhibit discontinuities, in particular with the recent seminal inventions of transformers and large language models coming to the fore, the smoothed curve aligns with Kurzweil’s 2030 target for human-level intelligence. On the other hand, where Miller has been disappointed is the merely modest advances in biological human engineering, with far fewer leaps forward than he had anticipated. Razib and Miller discuss whether this is due to limitations in the science, or issues of governance and ethics. Miller closes making the case for a program of cloning the great 20th-century genius John von Nuemann and the statesman Lee Kuan Yew. While the computational innovation driving AI seems to have advanced on schedule, and the biological revolution has not taken off, the last section of Miller’s book focused on the economic impacts of the impending singularity. He still believes the next 10-20 years will be incredible, as our economy and way of life are both transformed for the good. Until that is, humans become obsolete in the face of the nearly god-like forms of AI that will emerge around 2050. Until then, Miller anticipates the next generation will see rapid changes as people make career shifts every half a decade or so as jobs become redundant or automated. If Singularity Rising proves correct, the next generation will be defined by what the economist Joseph Schumpeter termed “creative destruction.” If Miller is correct, it may be the last human generation. For the first time ever, parents going through IVF can use whole genome sequencing to screen their embryos for hundreds of conditions. Harness the power of genetics to keep your family safe, with Orchid. Check them out at orchidhealth.com.  

    On this episode of Unsupervised Learning Razib talks to Rob Henderson, author of the new book Troubled: A Memoir of Foster Care, Family, and Social Class. Henderson is a commentator known for coining the term "luxury beliefs," a tendency among elites to use their beliefs to signal social status, with real-life costs of those beliefs born by non-elites alone. Henderson grew up in California foster homes, before being adopted into a working-class family in Redding, CA. After an academically undistinguished period in high school, he joined the U.S. Air Force straight out of high school, eventually serving a short stint in Germany. While in the military, he was identified as intellectually gifted via standardized tests, and it was during this time that Henderson developed habits that equipped him to become an exemplary airman, and eventually a public intellectual. Along the way, Henderson earned a bachelor’s degree from Yale and a Ph.D. in psychology from Cambridge University.  Henderson first came to public prominence with his 2018 New York Times op-ed, Why Being a Foster Child Made Me a Conservative. Later he outlined the concept of luxury beliefs in Quillette, and moved his popular newsletter to Substack.  Troubled fleshes out the working-class life experiences that made Henderson who he is today, and how and what sets him apart from other members of the elite-educated “professional-managerial class.” While he was an indifferent student who barely graduated high school, parental expectations prepped Henderson’s classmates for the Ivy League. At the time Henderson was studying at Yale, the median family income of his fellow students was $192,000. In Redding, CA, where he grew up, the median family income was about $65,000. About ten times as many Yale students hailed from the top 1% (19%) than the bottom 20% (2.1%) of family income. Troubled would shock many of Henderson’s Yale classmates, because the economic, social and cultural deprivation and domestic volatility he describes would be so alien and unrelatable. Among the most striking illustrations of how he grew up was Henderson’s perplexity upon being expected to be excited for his first birthday party in his adoptive home. As a former foster child with a winter-break birthday, not only had he never received presents, had a cake or a party. Henderson had literally never been sung “Happy Birthday.” Troubled beings with Henderson’s primary memories of his Korean immigrant mother when he was a toddler. After she was deported, Henderson’s formative years in childhood were spent as a ward of the state, shuffled from foster placement to placement. Razib also touches on something that Henderson has discovered in the last few years with consumer genetics. Not knowing who his father was, and clearly not being fully Asian, he had always trusted he was of mixed white and Korean heritage. But a 23andMe test makes it plain that his father was genetically Mexican. Not entirely shocking as Henderson was conceived in Southern California, the genetic test turned him overnight into a “BIPOC” individual, with nearly 20% indigenous American ancestry. For the first time ever, parents going through IVF can use whole genome sequencing to screen their embryos for hundreds of conditions. Harness the power of genetics to keep your family safe, with Orchid. Check them out at orchidhealth.com.