412 ‒ The evolving science of cancer: emerging approaches to prevention and early detection, & more
In a Nutshell
Cancer is shifting from a "bad luck" disease to one driven by inflammation, insulin resistance, and genetic risk interactions that convert dormant "sleeper cells" into tumors. Immunotherapy (CAR-T, checkpoint inhibitors) now cures certain leukemias/lymphomas, while GLP-1 agonists and metabolic therapies targeting hyperinsulinemia show promise for prevention and treatment. Longevity requires addressing regenerative failure (thymus, cartilage, immune system) alongside eliminating cardiovascular disease and cancer through biomarkers and systematic prevention.
These notes were generated by AI and may contain inaccuracies.
Peter Attia welcomes Siddhartha Mukherjee, noting 15 years since publication of The Emperor of All Maladies. Attia recalls reading it immediately upon release.
Mukherjee divides progress into three categories: prevention, early detection, and treatment.
Prevention
The good news includes expanded understanding of carcinogenesis beyond traditional carcinogens (x-rays, benzene, tobacco). Inflammatory states are now recognized as significant contributors to cancer development. The relationship is not simple gun-and-trigger; inflammatory states interact with mutations in complex ways.
The bad news is the absence of reliable biomarkers. Mukherjee states: "I don't have a high cholesterol for cancer." No equivalent to hypertension or elevated cholesterol exists for identifying pro-carcinogenic inflammatory states in individual patients.
Genetics and Risk
Knowledge of hereditary cancer risk has expanded beyond BRCA1 and BRCA2 to include many additional genes affecting breast, prostate, and other cancers. In some cases, risk arises from multiplicity of genes rather than single mutations. Genetic sequencing can now quantify relative risk for breast, ovarian, and increasingly other cancers.
Obesity is confirmed as a risk factor for certain cancers, notably endometrial cancer. Overall cancer mortality in the United States is declining, contradicting the notion that cancer rates are exploding. However, concerning trends include:
- Rising colorectal cancer in young people
- Plateauing breast cancer mortality gains from the previous 20 years
- Rising esophageal cancer
- Other cancers increasing in younger populations
The causes remain unclear for most of these increases. Endometrial cancer rise likely relates to obesity. Esophageal cancer is probably linked to combined alcohol and tobacco exposure, though drinking rates have declined, suggesting a 5-10 year lag effect.
Early Detection
Blood tests exist for early cancer detection, but applying them broadly to general populations without risk stratification leads to excessive false positives, unnecessary biopsies, and follow-up testing without reducing mortality. These tests show promise when applied to high-risk groups: individuals with genetic mutations predisposing to cancer or those with previous cancer diagnoses.
New imaging technologies are emerging, including continuous rather than discrete imaging. The traditional model of annual mammograms represents discrete testing with relatively poor image quality. Continuous monitoring approaches may yield different insights.
Treatment Advances
Immunotherapy has proven effective against some cancers. Mukherjee notes that 15 years ago he believed he would not live to see a medicine extending lifespan for smokers with lung cancer; this is now possible when immunotherapy is given to appropriate populations. These drugs "unleash the immune system to recognize cancers as foreign cells."
CAR-T cells—genetically engineered T cells—can cure liquid cancers including leukemias and lymphomas. "Cure" here means 15-20 year survival rather than 5-year survival. Patients treated with CAR-T therapy for refractory leukemias and lymphomas are walking around 15 years post-treatment.
Metabolic therapies exploiting cancer's metabolic differences from normal cells represent another treatment avenue.
Second- and third-line therapies have expanded survival in multiple cancers:
- Enhertu in breast cancer works after many prior lines fail
- Three or four new medicines for myeloma
- Two or three new medicines for lymphoma
- Two or three new medicines for prostate cancer
Surgical treatments have become more tissue-sparing while maintaining cancer removal efficacy, improving functional recovery.
Novel agents on the horizon include Daxon Rasib, a medicine targeting the Ras gene in pancreatic cancer. Patients with metastatic pancreatic cancer are now living months to years longer than would have been possible 10-15 years ago.
Attia references work by Brett Vogelstein and colleagues. A 2015 Science paper, revisited in 2017, analyzed mutations across cancer types:
- Approximately 5% were hereditary
- Approximately 30% arose from environmental triggers or carcinogens
- Approximately 66% were purely random—"bad luck"
This led to the concept of cancer as a "bad luck disease," where even individuals doing everything correctly could develop cancer through random mutations.
Mukherjee notes that the 5% figure is outdated and likely higher. Vogelstein's analysis focused on single-gene mutations (BRCA1, BRCA2, P53). The field has identified additional single-gene mutations increasing hereditary cancer risk.
In breast cancer, even accounting for all known single genes, only about one-third to one-fourth of hereditary breast cancer is explained. The remaining fraction involves multiple genes—potentially hundreds—each contributing small risk increases.
Mukherjee introduces the shove versus nudge mutations framework:
- Shove mutations (BRCA1, BRCA2) dramatically elevate risk—ninefold or tenfold above population baseline
- Nudge mutations each contribute small risk increments; accumulated nudges plus environmental factors eventually push individuals over the threshold
Mukherjee argues cancer is not purely bad luck but involves bad luck plus additional pathological processes, specifically inflammation. Work from Charlie Swanton's laboratory and others demonstrates that healthy individuals harbor cells bearing all genetic hallmarks of cancer without developing the disease.
Mukherjee references 1950s Oxford experiments using the carcinogen DMBA painted on mouse backs, producing few cancers. Adding the inflammatory agent croton oil increased cancer incidence dramatically to approximately 80 cancers. Later experiments by Allan Balmain at UCSF used genetically predisposed mice; introducing inflammatory wounds via staples produced massive cancers at wound sites.
Mukherjee proposes that individuals carry sleeper cells—mutated cells that remain dormant. An inflammatory response converts these into active cancer cells. The "real bad luck" is experiencing the particular inflammatory state that awakens sleeper cells.
Mukherjee emphasizes that "inflammation" remains an imprecise term requiring further mechanistic specification.
The shove/nudge framework has clinical utility. Quantifying where patients fall on the risk cascade enables risk stratification, enrollment in prevention trials, and personalized screening. Categories of mutation origin are not discrete but overlap.
Obesity confers enormous risk for endometrial cancer, though the mechanism remains unclear. Possibilities include inflammation or hormonal effects on sleeper cells. Obesity itself is unlikely to be directly mutagenic but represents a physiological state capable of converting dormant cells into tumors.
Attia questions whether the risk stems from adiposity itself versus insulin resistance, hyperinsulinemia, and the metabolic milieu accompanying obesity. Mukherjee notes that human data are lacking but references a 2017 Nature paper co-authored with Lou Cantley.
Lou Cantley co-authored a paper in Nature examining the relationship between insulin and cancer treatment outcomes. The research demonstrated that in animal models, obesity itself is not required to trigger the cascade that sustains cancer. The critical factor is insulin and the subsequent cascade of events that follows the insulin response, which appears responsible for cancer cell resistance to therapy.
The findings suggest that chronic insulinemic tone - the sustained elevated insulin levels rather than acute spikes - has the capacity to promote cancer growth and maintain resistance to certain therapies. Removing this chronic insulin tone makes cancers more vulnerable to treatment. This represents established findings rather than speculation.
Mitch Lazar at Penn conducted research approximately 15 years ago examining cardiometabolic disease incidence through a two-by-two analysis framework. The study distinguished between individuals who were obese versus metabolically unhealthy, recognizing that lean individuals can still be metabolically unhealthy while some obese individuals remain metabolically healthy.
The research revealed that disease prevalence tracked with hyperinsulinemia rather than adiposity itself. While this study focused on cardiovascular disease rather than oncology, it establishes plausibility that the same mechanism could apply to cancer, though the diseases differ in their pathology.
GLP-1 agonists initially raise insulin levels transiently to facilitate glucose disposal, but ultimately tend to reduce hyperinsulinemia. The mechanism remains unclear, though it may relate to weight loss and improved energy balance.
A proposed study design would involve simple randomization of standard cancer care versus standard care plus GLP-1 agonist treatment. The ideal cancers for such investigation include endometrial, ovarian, and breast cancers. The study would need to carefully examine patients' insulinemic response versus their weight loss response separately.
With AI capabilities, researchers could potentially dissociate these two variables to determine which drives outcomes. A broader epidemiological study examining whether GLP-1 users have lower endometrial cancer risk could be conducted within 5-10 years, though experimental approaches are preferred over observational studies due to confounding factors.
Osteoarthritis research has identified that the condition is not simply an obesity-related disease. Non-obese individuals develop osteoarthritis, including in non-weight-bearing joints like the hands, where massive inflammatory nodes develop. This challenges the traditional view that osteoarthritis results primarily from mechanical stress on weight-bearing joints.
The condition appears to result from an imbalance between cartilage generation and destruction. During normal aging, cartilage undergoes continuous cycles of generation and destruction. When this balance is disrupted - specifically when stem cells responsible for cartilage formation decrease while destruction continues - osteoarthritis develops.
Research demonstrated that injecting an inflammatory response, such as LPS (lipopolysaccharide) or bacterial mimics that activate the immune system, into joints of young animals causes stem cells to stop producing cartilage. This inflammatory activation mimics osteoarthritis exactly, with destruction increasing and genesis decreasing.
This finding applies across multiple stem cell types, including skin, bone marrow, and blood stem cells, where certain immunological responses reduce regenerative capacity. The teleological explanation for why immune responses would suppress stem cell function remains unknown and may not be scientifically addressable.
Approximately 80% of solid organ tumors contain novel neoantigens that can be recognized by the immune system as non-self. This includes the most lethal cancers: breast, colon, lung, and pancreas. However, the immune system often fails to mount sufficient responses despite this recognition capability.
The traditional view that new antigens arise from the 30-100 mutations typically found in cancer cells has been challenged. Instead, research from Tyler Jacks' laboratory at MIT demonstrates that cryptic antigens - pieces of genes that should not normally be expressed - emerge when cancer cells deregulate their genetic decoding machinery. These cryptic proteins represent the actual neoantigens that should trigger immune recognition.
Cancer cells survive by expressing multiple checkpoint molecules - flags on their surface that signal to T cells and B cells "I'm normal. Don't attack me." While normal cells use these checkpoints to prevent autoimmunity, cancer cells exploit this system to evade immune detection despite expressing neoantigens.
There are dozens of such checkpoints, not just the few currently targeted by checkpoint inhibitor therapies. This explains why even cancers with numerous neoantigens often escape immune attack - they have evolved multiple mechanisms to suppress immune recognition.
AI analysis of CAT scans revealed a correlation between thymus presence and health outcomes. The thymus serves as the developmental site for T cells - essentially their "kindergarten." Individuals with intact thymuses showed lower rates of autoimmune disease and cancer, while those whose thymuses had involuted with age demonstrated higher disease incidence.
This represents a correlative rather than causal finding, but suggests the thymus could serve as a biomarker similar to how Lipitor is used for cardiovascular disease risk. Experimental approaches would involve transplanting new thymuses into older animals to test whether this restores immune function.
Previous attempts at thymic regeneration used combinations of growth hormone, metformin, and other compounds in open-label studies without randomization. These studies suggested thymus regrowth but lacked rigorous controls. The rationale for these specific compounds remains unclear.
Current understanding favors closed circular research approaches: starting with human epidemiological observations (like thymus retention correlating with fewer diseases), then moving to mechanistic studies. AI analysis of medical imaging has identified thymus status as a potentially significant marker for wellness, longevity, and disease risk.
AI analysis of tens of thousands of CT scans has revealed thymus involution patterns. Researchers can conduct experimental thymus implantation in animal models by opening the chest and inserting a new thymus to study growth factors and mechanisms. This systematic approach allows detailed investigation of why some individuals retain their thymus while others experience involution. The goal is to identify thymic growth factors that could be used therapeutically to maintain thymus function and prevent diseases including cancer.
CAR T therapy represents a major immunotherapy success, particularly for B-cell lymphomas and leukemias, achieving durable cures rather than merely extending median survival. A chimeric antigen receptor (CAR) T-cell is created by taking T-cells from a patient and genetically modifying them using a virus-based system to insert a gene that enables recognition and killing of cancer cells. These modified T-cells are then grown and reinjected, effectively weaponizing them to target cancer while overriding normal inhibitory signals.
One CAR T company in India has successfully treated nearly 150-200 patients at 1/20th the cost of US therapy.
CAR T cells that target B-cell leukemias and lymphomas kill all B cells, not just malignant ones. This works because humans can survive without B cells. However, this reveals that CAR T therapy is lineage-specific and tissue-specific rather than truly cancer-specific. The therapy cannot be applied to brain cancer because CAR T cells would destroy essential brain tissue along with cancer cells.
Tissues that can potentially be targeted include breast tissue, ovaries, endometrium, pancreas, and prostate, as humans can survive without these organs.
Ovarian cancer presents a promising target because ovarian cancer cells (actually fallopian tube cells) express well-known antigens suitable for CAR T development. Multiple attempts are underway to create CAR T cells against ovarian antigens.
Pancreatic cancer presents greater challenges because KRAS mutations, present in most pancreatic cancers, produce antigens that are hidden inside cells rather than on the surface where CAR T cells can target them. KRAS is better suited for small molecule inhibition rather than immunotherapy approaches.
Pancreatic cancer outcomes depend significantly on anatomical location. The pancreas resembles a leaf, with head tumors near the intestines causing pain and jaundice, while tail tumors are often detected incidentally. Tail tumors have better surgical outcomes due to less involvement with surrounding organs and vessels. Some patients have survived 10 years with incidentally discovered tail pancreatic cancers treated surgically.
The pancreatic head represents the most complex anatomical region below the neck, where intestines, blood vessels, and other structures create a "traffic intersection" making surgical intervention extremely difficult.
Cancer therapy typically progresses from treating diffuse metastatic disease to surgical removal of bulk tumor followed by adjuvant therapy to eliminate remaining cancer cells. Clinical trials are being planned to test whether KRAS inhibitors like Dacogen Rasib (Daxsonrasib) can serve as effective adjuvant therapy following Whipple procedures for pancreatic cancer patients with KRAS mutations.
Lower tumor burden significantly improves therapy success probability. Reducing from 1 billion to 1 million cancer cells decreases the likelihood of resistant cancer cells by a factor of 1,000. While exceptions exist (papillary thyroid cancer, some checkpoint inhibitors), the principle generally holds that fewer tumor cells correlate with better outcomes.
Colorectal cancer rates are increasing in younger populations, prompting recommendations to begin screening at age 40 for average-risk adults rather than 45. Potential contributing factors include increased exposure to forever chemicals in food that demonstrate pro-inflammatory effects in animal models, disruption of gut microbiome through antibiotics and dietary changes, and reduced fiber intake leading to less frequent bowel movements compared to high-fiber diets that historically protected populations.
Recommendations include doubling or tripling fiber intake, reducing forever chemical exposure, and consuming live yogurt at least weekly for probiotic benefits.
AI applications span multiple areas: epidemiological scanning for disease correlations, clinical trial recruitment and monitoring, and drug discovery through molecular design. Every drug functions through lock-and-key binding to proteins, though proteins are dynamic structures requiring keys that accommodate molecular movement.
Two primary approaches exist: search-and-fit molecular design (testing libraries of potential keys) and generative molecular design (creating keys from scratch based on lock structure). Search-and-fit requires testing enormous libraries (10^20 possibilities), making it computationally prohibitive without quantum computers.
Generative chemistry uses AI to learn the rules of synthetic molecular chemistry and physics, scaling one chemist's expertise to 50,000 virtual chemists. This approach examines protein pocket geometry, electrostatics, and molecular principles to design drugs from scratch rather than searching existing libraries. This represents AI's most valuable current application in creating medicines for human diseases rather than chatbots or image generation.
The complexity of teaching AI biological rules for simulating actual experiments remains significantly greater than teaching chemistry and physics rules.
The complexity of teaching AI biological rules for simulating actual experiments is significantly greater than teaching chemistry and physics rules, as the rules in biology are not as neat. There aren't really any rules. The biology of a kidney cell is very particular—its job is to filter out salts, and this is well-known and discrete. A brain cell has no resemblance to a kidney cell, though metabolically there is some resemblance. The brain cell's job is completely different: to transmit information, sequester information, store it. A skin cell is a completely different kind of cell.
A muscle cell has a very mechanical and relatively straightforward functional determination of what it can do. What's happening in the neuron is not as mechanical as what's happening in the myocyte or the nephron, where you can see an ion moving through a channel against a gradient, and when the gradient switches, it does something specific. There is some use to building biological models of cells, but it will not be easy, and it will not work until you get down to every cell-by-cell level.
In 15 years, new antigens and new theories of how cancers arise because of inflammatory states or other states will become clear. The role of obesity will also become clear. The answer has always been prevention. The question is whether prevention will be general prevention through habit change, behavioral change, and lifestyle change, or whether it will be chemo-prevention.
Just as Lipitor can be given long before a heart attack, and an ACE inhibitor can be given if blood pressure is high to prevent a heart attack, the question is whether there will be molecules that will prevent cancer. There will be chemo-prevention, but it is not clear that chemo-prevention is going to be a simple relationship like that. Lipitor decreases LDL levels, and Losartan or an ACE inhibitor or ARB inhibits the capacity to have high blood pressure by acting on the kidneys' hormonal system.
There is an analogy in cardiovascular disease where patients are now starting to be checked for CHIP. While therapies are still limited, colchicine is available. There is evidence that if you have clonal hematopoiesis, you can treat that preemptively, and that is the first immune-based therapy for heart disease. This presents a model that could be applied to oncology in 15 years.
GLP-1s are available in obesity and potentially in inflammation, and this is a trial to watch carefully. Breast and endometrial cancer can be used as models. There are a variety of anti-inflammatories of which some are known.
The prevention of pathological disease—cancer and heart disease, the two biggest killers—is not the answer to the prolongation of longevity. One current model of longevity is the prevention, elimination, treatment, or cure of diseases that kill us, of which the two big ones are cardiovascular disease and cancer.
The orthogonal way to look at longevity is to not be interested in the diseases, cancer and heart disease, but instead be interested in a physiological state that promotes the elimination or decrease of not just those diseases, but all diseases, including diseases of degeneration, osteoarthritis, immune failure, bone marrow failure, and chronic unopposed inflammation.
If the physiological state of the body could be changed to reduce the inflammatory state and reduce the chronic insulinemic tone, that may be the path to longevity. Wellness instead of illness may be the path to longevity.
Atherosclerotic cardiovascular disease is the number one killer globally. There is a world in which this can be turned into an orphan disease because the causal drivers are well known and treatments exist for all of them. If you don't smoke, take whatever drugs are necessary to stay insulin sensitive, obliterate apoB and LDL-C down into the physiological levels you were born with, and immediately treat hypertension if it develops, you cannot get heart disease. The pathophysiology inside the artery wall cannot happen.
There is not the same degree of confidence that anything could be done to eliminate all cancers. Even if completely insulin sensitive and never smoked, there are still enough genes and random mutations to require reliance on early detection and treatment.
People who are super centenarians—more than 110 years old—do not die of cancer. They are dying of failure of skin regeneration—their skin falls off. They are dying of bones being broken. They are dying of regenerative failure. They still die of atherosclerosis and infections like immune cells breaking down. Infections represent immune failure. They are not dying of cancer. They are dying of bone breaks, skin failure, immune failure, blood failure—all regenerative failures.
There are two worlds. One world in which humans die at age 70, 80, 90 of the usual suspects—the three major suspects being neurological disease, cancer, and heart disease. In the opposite world, once not dying of those diseases, you can get into your hundreds, and the killers are not those diseases. The killers are immune failure, blood system failure, bone breakage—basically regenerative failure.
Elie Metchnikoff, the Russian Parisian scientist who discovered the innate immune system, in the 1890s was the first person to propose that death be studied systematically. He created the word thanatology—the study of death, which is the opposite of the study of longevity. Metchnikoff argued that if the proximal causes of death could be removed—cancer, heart disease, neurological disease, accidents, and infections—human beings could live to 100 or 102, and at some point would naturally want to die just like trees at a certain age after hundreds of years finish their sap and die.
Rather than picking off the buds one by one—atherosclerosis, cancer—something is needed that goes after the foundation of aging. The degree of optimism about AI curing aging and making people immortal within 10 or 15 years is low. Regenerative failure is a real phenomenon, and treating regenerative failure is not easy. It will be a huge new frontier in biology.
What is regenerative failure of cartilage? What is regenerative failure of the brain? What is regenerative failure of the immune system of the thymus—there is no thymus, so how can a new one be made? The new frontiers of longevity will have to deal with systematically addressing regenerative failures, separate from the diseases that kill at 70 and 80.
There is a hack to longevity, but it is not a simple hack. It is not going to be peptides. There are 20 amino acids per individual molecule, so the combinatorics of injecting yourself with a random group of peptides means the chances of finding the right peptide at the right concentration are basically zero. That is a mathematical truth.
The hack is finding out systematically how to address regenerative failures and separating that out from the diseases that kill at 70 and 80. Both need to happen in parallel. Remove atherosclerotic heart disease using all mechanisms and medicines available now, including behavioral change—most importantly, behavioral change. Then the question left is regenerative failure.
Physical longevity requires systematic study of how to remove regenerative failure in people in their 70s, 80s, and 90s, including regenerative failure of the brain. Virtual longevity is a different universe—Matrix red pill, blue pill—where the physical body is no longer needed, or robotics takes over, or you are a brain in a box.
The NIH is not putting resources into studying the hallmarks of aging and gerontology. If fundamental questions are to be answered, they must be done systematically. This is very big science—a Manhattan project of science. It nearly lies outside the current NIH mandate. A new NIH would be needed to do this, with a new mandate.
The problem with gerontology is that history is important. What changed cardiovascular science was finding biomarkers for disease—proxies for disease. You can measure blood pressure, you can measure cholesterol. Very soon it was said that heart attacks do not need to be waited for. If a drug reduces blood pressure or cholesterol, that drug will be made available. In post-hoc studies, those medicines were found to actually reduce heart attacks.
There are no such biomarkers for cancer, and there are none for aging or longevity. Biological clocks are talked about, but a reliable biomarker is needed.
Finding a reliable biomarker for aging is big science. The biomarker study itself will take 15, 20, 30 years because it cannot be done retrospectively, though it can be done sort of retrospectively by looking at what is in the blood of people who live to 100. The interventions testing program could be used to look in the metabolome and proteome of all the ITP winners versus losers to get some clues. Finding the biomarker is very big science, and it is reliable science. The only problem is that it cannot be screwed up because it sets the trajectory indefinitely.
David Sabatini has argued that a Manhattan Project around aging is needed, and it is not going to be the NIH stepping up. It will be a group of philanthropists that decide to put money into it. The effort must have a significant engine in basic science. It cannot go straight to drug discovery, which is the temptation. When these things are started as for-profit companies, there must be a return to shareholders quickly, so the path goes straight to drug discovery and whack-a-mole. The view is that there must be a continuum, and basic science must be present.
This is not a drug discovery project. This is a cholesterol discovery project. This is a hypertension discovery project. Unless cholesterol, hypertension, etc. for aging are discovered, the effort will be lost. It requires a 30-year Manhattan project—maybe 15 years if good with AI. The NIH has no mandate for this right now. Maybe a new NIH will, but the current NIH does not have a mandate.
It is unlikely the NIH will ever take this on. Therefore, a dinner is overdue to brainstorm ideas with Lou, David, and others.
There is a National Institute of Aging. They could change some of their focus, and there is a focus on well aging. The problem is the NIH's mandate is illness, not wellness. That is a great thing—the NIH has done great things with it, producing incredible things from vaccines to Lipitor. It is not the current mandate.
A metabolic trial was initially proposed to Lou Cantley and others. Lou Cantley and the founder sat down for dinner in a restaurant on the Upper East Side and on the back of a napkin sketched out a potential strategy based on the idea that cancer cells must depend on particular metabolic inputs, which make them different from normal cells. This is known from Otto Warburg's work in the 1900s—the metabolism of a cancer cell is not the metabolism of a normal cell. Every time something is different about a cancer cell than a normal cell, that makes it possible. It is an opportunity.
That opportunity was used to make new kinds of drugs that would particularly target cancer metabolism and less so normal metabolism. That led to a company called Five Therapeutics. It went public recently. It is in the middle of trials, and early trials are very promising. The last public disclosures indicated that targeting cancer metabolism in endometrial and breast cancer was working.
A book on death and dying and longevity is being written—not the usual kind, but the kind of stuff discussed: how to really systematically think about longevity, how to systematically think about death.
Sid Mukherjee emphasizes that before writing any words, extensive preparation is essential. He describes this as "dream time" – a period of deep thinking about the book's fundamental elements including who drives the narrative, geographic and temporal placement, mood, target reader, structure, and anatomy of chapters and sections. This preparatory phase should last months, weeks, or even years before putting the first word on the page.
Mukherjee compares the book writing process to music, noting that musicians understand the importance of knowing the mood of a piece before playing it. He explains that once the skeleton or mood of the book is established, the actual writing flows more quickly because the foundation is already set. The notes serve as mechanisms to drive the mood, just as pages and words hang on the established structure.
Mukherjee distinguishes between the writing and editing phases, noting that his actual writing period is relatively short once the conceptual work is complete. However, the editing process is significantly longer and more deliberate. He credits his editor Nan Graham, who has been his editor from the beginning and has also worked with authors including Thomas Pynchon, Arundhati Roy, and Stephen King.
The conversation concludes with mutual appreciation between the hosts, with Mukherjee expressing thanks for the discussion and mentioning he owes a dinner to his interviewer.
Keep Peter Attia MD in your library
Save the videos and channels worth coming back to, and find them again in one place.
