In a review of the year's podcast episodes, Tyler and Jeff discussed some of their favorites. Tyler highlighted the conversation with YouTuber Amy Austin as a standout, feeling it was destined to be good. He explained, "It was preordained that it was going to be very good because he's willing to talk and he's smart and he'll say what he thinks. So it just is not going to go wrong."

A key observation was that the single-subject episodes were often the best. These episodes, which focus deeply on one person's specific expertise, included conversations on Saudi Arabia, Buddhism, and neurosurgery. Tyler noted that he learns more from preparing for these focused discussions. While an episode with a familiar guest like Steven Pinker can be good with little preparation, the deep dives require much more work. For example, preparing for the episode on Buddhism took him four to five months.

When asked what he took away from that intensive research, Tyler shared several key insights that countered popular Western perceptions of the religion. He learned about its violent aspects, that it is a shrinking religion, and that what is often called "American Buddhism" is quite different from actual Buddhism. He had some prior knowledge of these points, but the deep preparation made them much more vivid and detailed.

A reflection on the year's podcast episodes highlights several underrated gems. The conversation with Gianamechi was particularly enjoyable due to his stories about life in the NBA. Though some listeners may have felt it wasn't deep enough, his intelligence and excellent podcasting voice made it a standout. Tyler suggests it's an episode that is better listened to than read. Another excellent, though not yet released, episode features Gaurav Kapadia, an investor from New York City.

I think I said once, Gaurav, you're either the most underrated unknown person or the most unknown underrated person, and I'm not sure which, or maybe both.

The episode with Cass Sunstein on liberalism was also singled out as surprisingly engrossing. Despite Cass being a frequent public voice, this conversation captured him at his very peak, as he felt challenged and rose to the occasion. The quality of the discussion was partly attributed to recording it in person in Cambridge. It was noted that this year, listener popularity was driven more by a guest's name recognition than the eclectic nature of the topic. Big names like Sam Altman drew large audiences, which differs from the experience of other podcasters whose more obscure guests often become the most popular.

A listener, Patrick McKenzie, asked about the biggest surprises in AI this year and how AI has impacted the podcast's production. Tyler's biggest surprise was the capability of the O3 model, which he half-jokingly called AGI. He believes the old definition of AGI, being as good as human experts, is now largely true. The main surprise was the significant leap in reasoning capabilities. Otherwise, the surprise was the lack of surprises, as progress has been steady and continuous.

They used to think there were two R's in Strawberry and now they know there's three. And in their spare time they win gold medals at Math Olympians.

Regarding the show's production, Tyler explained that AI has been transformational. It has enabled them to produce many more episodes. For a discussion with Buddhism scholar Donald Lopez, Tyler read about 30 books but also ran numerous GPT queries. This saved him hundreds of dollars and a lot of time, allowing him to get straight to the essential points for the conversation.

We could not have done as many episodes as we did had it not been for large language models.

On the technical production side, improvements have been more marginal. Machine learning algorithms that remove filler words like 'ums' and 'ahs' have improved. AI is also helpful for brainstorming and finding new guests, but it has limitations. It often suggests very well-known individuals or people who have already been on the show. For live events, the goal is to find someone who can attract an audience but isn't overexposed. The AI might suggest someone like Oprah Winfrey, who, while a great guest, doesn't fit that specific niche of being undiscovered.

A listener asked for an update on a potential episode with chess grandmaster Magnus Carlsen. Tyler explained that he had a great dinner with Magnus, who agreed to do the podcast. However, the birth of his child created what Tyler called "imperfect timing" for the recording. He still believes there is a good chance the episode will happen and plans to try again once he feels Magnus's family is more settled.

Regarding other potential chess guests, Tyler is selective. He would like to host Fabiano Caruana, whom he describes as very well-spoken, thoughtful, and philosophical. In contrast, he finds some other top players annoying.

Their voices, their personalities, their narrowness.

Tyler clarified this is not at all true of Magnus or Caruana. He was very impressed by how many different topics Magnus could speak about intelligently off the top of his head. While he wouldn't refuse an interview with the current world champion, Gukesh, he is uncertain how good the conversation would be. He presumes that at 19 years old, Gukesh may not have much experience in things outside of chess.

When traveling for events, Tyler's hotels are usually chosen for him. He finds these are typically better than what he would choose for himself. His main criteria for a good hotel are practical. He looks for a swimming pool, enough outlets in the room to plug everything in, and a sufficiently flat pillow. He also values being able to easily operate the shower and turn off all the lights, which he notes is becoming harder in more modern hotels.

While the availability of outlets has improved significantly over the last five years, the complexity of controls for showers and lights has increased, particularly in the fancier hotels selected by event organizers. The hotels he chooses for himself are generally not fancy enough to have these complicated features. The swimming pool is a particularly important factor. Tyler swims every day when a pool is available, considering it his equivalent of a gym. He finds the gym boring but enjoys the pool, valuing it more than five times as much as a hotel gym.

When asked if he has become less honest or more cautious with his growing influence, Tyler reflects that it feels the same to him, but acknowledges that people who fool themselves often don't realize it. He offers a recent example to suggest he might actually be more honest now. During a recording with Alison Gopnik, he accused her of "laboring under delusions" when he felt she was denying the heritability of IQ. Tyler notes, "I actually don't think I would have said that 10 years ago. So maybe in some ways I'm more honest."

Jeff observes that as a listener, he felt Tyler pressed her more than usual on the topic, rather than moving on. Tyler agrees, explaining his reasoning for being so direct.

But it just seemed so egregiously off to me. And it is her field, right. If someone else said that, whatever. But it's the main thing she works on.

This leads to a broader discussion about whether influential people have others around them to hold them accountable. Tyler believes this is a growing problem. People become hesitant to challenge influential figures, either because they want something or because they observe others keeping their distance and become afraid to speak up themselves.

A Supreme Court case is currently challenging the basis of recent US tariffs. Betting markets suggest an 87% chance the court will rule against the Trump administration's use of tariffs. The core of the administration's argument is that tariffs are a form of regulation, not a tax. However, Tyler expresses skepticism about this claim, noting that American history suggests the contrary.

Even if the court rules against the administration, it does not mean all tariffs will disappear. While the ability to unilaterally impose tariffs on countries like Canada or on products like chocolate from Madagascar might be curtailed, tariffs justified by national security arguments will likely remain. The situation is still developing, with alternative plans already being formulated.

A more productive dialogue about AI risks requires a more rigorous approach from those who are most worried. Tyler Cowen suggests that individuals concerned about AI risk should engage with the peer-review process to develop a formal academic literature. He notes that despite these arguments being present for 15 to 20 years, there has been a refusal to do so.

The burden of proof rests on those making the claims about AI risk, especially since other indicators do not support their predictions. Market prices and superforecasters, for the most part, are not aligned with their views. When asked why a traditional academic path is necessary in a fast-paced, online discourse, Tyler argues that both online discussion and formal review are valuable. However, peer review provides a level of careful scrutiny that informal commentary cannot match.

Peer review is super useful. You have things where there's a deliberate exchange of does this argument succeed or not? Where people scrutinize it very carefully, not just some weird blog commentators. And it works for every other science.

Tyler points to climate change as an example where peer review successfully distinguished between strong claims, like the reality of warming, and weaker claims, like the rising costs of that warming. This rigorous process is what's missing from the AI risk debate. He has even offered to help philanthropists support such a project and volunteer as a referee, but to no avail. The emergence of AI tools like Refine, which can provide peer-style comments, is seen as a potential step in the right direction.

When evaluating research, it is better to trust entire bodies of literature rather than individual papers. Bryan Kaplan's claim highlights this, emphasizing the "spontaneous order sort of wisdom" embedded in a field's collective work. This collective wisdom is seen as lacking in the area being discussed.

The question arises as to why this literature has not been developed. One hypothesis is that key figures in the field may not want that level of scrutiny. While they are sincere in their beliefs, their lack of trust in the peer-review process might reflect an insecurity about their arguments. They may believe their ideas will ultimately be proven right, but at some level, they seem hesitant to subject them to rigorous academic critique.

When considering a future where a Neuralink-like device could upload the world's knowledge directly into our minds, the question arises: is the journey or the destination more valuable? This specific scenario seems very distant and metaphysical. A more tangible reality is a world where Neuralink helps people with paralysis perform useful tasks.

However, the underlying question about process versus outcome is important. Using a more current example, consider writing. Even though large language models (LLMs) can write well, the process of writing is something humans should never abandon. There is a strong belief in the importance of the process for humans, not just the final outcome.

When asked why more people have not copied his interviewing style, Tyler Cowen questions the premise. He notes his style might not be good, and it requires a lifetime of broad reading and preparation, making it very difficult to replicate. He suggests the returns may not be high enough for others and expresses a belief in pluralism, stating, "Let a thousand flowers bloom."

Jeff agrees the style is not for everyone and notes that some are turned off by what can feel more like a "grilling" than a conversation. However, he observes that one element has become more common: the super-specific first question that gets right to the point. This approach works particularly well for video platforms like YouTube, where long introductions are detrimental to engagement. In this sense, Tyler was ahead of the curve as a content creator.

Tyler adds another reason his style is rare. It requires someone who is intellectually strong enough to be a guest but also deferential enough to want to be the host. He describes this as a "weird personality quirk" of his own, as many impressive people who would be great guests are not interested in being an interviewer. He also notes that an important part of his style is always having something to say when a guest turns a question back to him.

Tyler claims he has never experienced uncontrollable laughter in his entire life. He believes there is a ceiling on how funny something can be, just as there is a limit to how good food can taste. He explains this is consistent with his general disposition; he does not feel extreme emotional highs or lows, but rather maintains a steady middle state.

Things just aren't that funny. Like, how good can something taste? Take the best sushi I've ever had, which is quite good. Things can taste a bit better than that, but not much. So funniness. There's a maximum. It does not bring me to uncontrollable laughter. And that's just the equilibrium.

Tyler even questions if uncontrollable laughter is a form of displeasure, similar to being tickled when you want it to stop. He recalls seeing comedian Louis CK live, which he considers the funniest show he's ever attended. While he laughed a great deal, he was not close to laughing uncontrollably. He suspects this personal trait is heritable.

When asked how the comments section on his blog, Marginal Revolution, has changed his views on regulation, Tyler Cowen explains that it has made him very tolerant of content moderation by platform owners. He is not bothered when a large tech company like Meta removes content. He does not see it as censorship, but rather as the platform's right to curate its own space.

If they don't want to put someone up saying, lemon meringue pie is great, I'm like, yes, lemon meringue pie is not great. That's fine. You spread the truth. And it just doesn't bother me at all.

Tyler applies this same principle to his own blog, where he actively moderates and deletes comments that he feels do not improve the quality of the conversation, even if they don't violate specific guidelines. He believes this lesson generalizes, arguing that society should be very tolerant of dominant suppliers like Facebook doing the same thing. He dismisses the common complaint that there are no alternatives to major platforms, stating that this is not true and is becoming less true over time.

An experiment is proposed to see which of a series of previous listener questions an AI could answer most like Tyler Cowen. The questions recalled include topics like growing in influence, being more honest or cautious, how Tyler picks hotels, and whether he would change course on tariffs.

Tyler believes the AI would do the best job on the tariffs question. He reasons that if the AI has access to up-to-the-minute information from the internet, it should logically arrive at the same answer he did.

When they test this by asking ChatGPT to answer as Tyler, the AI first provides a disclaimer. It states that it cannot write as Tyler Cowen but will attempt to provide a "crisp economist's eye answer with a similar vibe."

A list of ten potential reasons could explain what might make the US change its course on tariffs. These include: inflation that voters can trace to tariffs, administrative bottlenecks at ports, a grand bargain tied to security, a recession, unified opposition from Corporate America outside of big tech, agriculture being boxed out, a carbon deal reframing the border, bargains with countries like Mexico, Vietnam, and India, legal or budget constraints, and a symbolic win the White House can sell.

While most of these are valid factors, many have already been in place. The key question is what will create a change, or a 'Delta'. Two additional factors that could account for this change are potential Supreme Court decisions and the possibility of the Republican Congress standing up to Trump more, motivated by recent election results where Democrats performed better than expected.

Tyler runs through a list of films. He considers Clint Eastwood's "American Sniper" to be a good movie, but not one of the director's most important works. He notes that "Gett: The Trial of Viviane Amsalem" hasn't stuck with him. He initially felt "Ex Machina" was a bit thin, but acknowledges its evolution into a seminal film, though he still views it as good rather than great.

Discussing Pixar's "Inside Out," Tyler remarks that it seemed splendid at the time but also hasn't stuck with him. He extends this to many Pixar classics.

I never sit around and think, 'Oh, I'd love to see Toy Story again.' Though at the time I loved it.

Jeff suggests this is unsurprising, given an earlier discussion, because a movie about inner turmoil and competing emotions is unlikely to resonate deeply with Tyler. In contrast, Tyler loved "Red Army," a 70-minute documentary about the Soviet hockey team, calling it the best movie of that year. He also praises "Meru," a documentary about mountain climbing, describing it as "behavioral economics on the screen." He mentions that it deserved a "Cass Sunstein Award," which he explains was an award Cass Sunstein used to give for the best movie on behavioral economics.

A review of films from a specific year reveals a general sense of disappointment. While some movies are acknowledged as good, they don't reach the level of greatness. For instance, "The Martian" is considered a good movie, but not as good as many other science fiction films. Similarly, the Iranian movie "About Elly" has aged well, but it is not the director's best work; "A Separation" holds that title.

When discussing the film adaptation of "Macbeth" starring Michael Fassbender, Tyler expresses a personal preference for reading Shakespeare rather than watching it on screen. He feels he processes the material better when it's on the page. His favorite Shakespearean movie is Orson Welles's "Chimes at Midnight," which notably takes great liberties with the source material.

A central theme of the discussion is the stark contrast between the quality of films and books from that year. Tyler points out the release of significant literary works like Michel Houellebecq's "Submission" and the final volume of Elena Ferrante's Neapolitan Novels, questioning if any film from the same period could match their stature. He notes that even a well-regarded film like "Sicario" feels like it explores a tired theme of drug gangs. While it has some very good sequences, it's ultimately just a good movie, not a great one.

Other films mentioned include "Mustang," a Turkish film about five orphan sisters, which is remembered as a good movie. In contrast, the Charlie Kaufman movie "Anomalisa" is largely forgotten and hasn't aged well for the speakers. The final verdict is clear: it was a fantastic year for books but a miserable one for movies.

Among a list of non-fiction books from a particular year, Robert Toombs's book on English history, "The English and Their History," stands out as the clear winner. Its impact was significant, prompting many people to ask for similar recommendations. It is described as an utterly compelling book that everyone should read.

Several economics books were also highlighted. Richard Thaler's "Misbehaving: The Making of Behavioral Economics" is praised as a good example in a category that needs more entries: books about economics and economists. Garrett Jones's book, "The Hive Mind," has held up very well and is now considered a seminal work, often cited in publications like The Economist. Another notable work is Scott Sumner's "The Midas Paradox." This book, which Sumner worked on for about 20 years, correctly argues that deflationary pressures were the biggest problem during the Great Depression. The book "Mastering the Path from Cause to Effect" by Angrist and Pischke is mentioned as an incredible achievement, though not a book one reads in a traditional sense.

A list of exceptional non-fiction books from a particular year is discussed. Notable mentions include Ashley Vance's biography of Elon Musk, which is considered very important. Another standout is Charles Moore's second volume on Margaret Thatcher, titled "Margaret Thatcher at Her Zenith in London, Washington and Moscow." It is described as one of the best biographies of the last ten years, though its high level of detail may be too much for some readers. While a book on Genghis Khan was on the list, a more contemporary book on Han Sen is suggested as a possible replacement. That year was particularly strong for non-fiction, with 28 books making the list without difficulty, prompting concern that the quality of subsequent years may not measure up.

A recent year was described as stellar for fiction. One standout book is Michel Houellebecq's "Submission," which Tyler considers the most classic work of fiction of the last two decades. He sees it not just as a novel, but also as a work of philosophy that is a unique blend of predictive and normative ideas. It's a short book that is widely referenced and rewards rereading.

You can think of it as a work of nonfiction. In a sense, it's a work of philosophy. Not quite predictive, not quite normative. A weird combination of them both. Everyone refers to it, it rewards rereads, and it's just incredible.

Jeff was in such a rush to read it that he first read the German translation, as it was released before the English version. Tyler confirmed he reread it in English and it holds up. Other notable books from that year were also discussed. While "The American People" by Larry Kramer and "The Seventh Day" by Yu Hua did not leave a lasting impression, "The Widower" from Singapore was a great book. Kamel Daoud's "The Meursault Investigation," a retelling of Camus's work, has also become a classic. The year produced three classics, with two being exceptionally high-quality and important.

A recent period featuring artists like Kanye West, Frank Ocean, Kendrick Lamar, Drake, and D'Angelo represented a second golden age for rap and R&B. However, that era is now considered over. D'Angelo has passed away, Kanye West's career has taken a different turn, and while Kendrick Lamar's work is still high-quality, it no longer feels as fresh.

Other notable music from that time includes Kamasi Washington's jazz album "The Epic" and the box set "Hula Land: The Golden Age of Hawaiian Music," which is praised for being fun, accessible, and different. Recommendations also extend to Syrian music, such as "Dabke: Sounds of the Syrian Hawran" and Omar Souleyman.

In classical music, Matthew Bengtson's recordings of the Scriabin piano sonatas are highlighted as the best versions. Tyler mentions corresponding with Bengtson, who is also an avid Magnus Carlsen fan. He recounts seeing Bengtson perform live in a piano store in Michigan, describing it as one of his best musical experiences.

When preparing a DJ set of choral music for Rick Rubin, Tyler re-evaluated the entire genre. He realized how good the 20th century was for choral music, which might be counterintuitive for some. A great deal of religious and classical music was composed during this time, suggesting that the genre never actually declined. The process also offered insight into Rick's taste. Rick loves the purity of composers like Palestrina and Arvo Pärt. This contrasts with Tyler's preference for more rhythmic or unusual-sounding choral music.

That blend between the two of us, he wanting the pure, I'm always wanting something weirder sounding, I thought made that a very good podcast.

Tyler believes this combination of tastes made their collaboration successful, and he considers it one of the projects from the past year that he is most happy about.

When traveling by plane, Jeff Holmes keeps his routine simple: he primarily reads. He estimates that 80-85% of his time on a plane is spent reading books. If Wi-Fi is available, he might also use some of that time to catch up on email. He considers train travel to be identical to being on a plane, another opportunity for reading.

His approach to driving is different. On shorter drives, such as his commute to school, he will listen to satellite radio or simply talk if he has a passenger. He specifically avoids listening to podcasts or other audio content while driving, preferring music instead. He notes that he has not been a passenger on a long road trip in over ten years.

Oman is a significant and historically rich country in the Gulf region. Unlike many other Gulf states, it was a civilization in the proper sense with a grand and glorious past, having controlled parts of Zanzibar in the 19th century. The National Museum is beautiful and very impressive. It's an important place that people don't often consider visiting.

As a travel destination, Oman is comfortable, safe, and hassle-free, with many people speaking decent English. The weather and scenery are gorgeous. However, it is not primarily a food destination. While some good food can be found, visitors should come to see the country and its landscape rather than for the cuisine.