Jeff Hirsh Read: January 2025
Inspiration: Read Jamie Metzl’s prior book Hacking Darwin and interested to read his latest book on the future of technology
Summary
Written with the help of ChatGPT, below is a brief summary to understand what is covered in the book.
“Superconvergence”, published in 2024 by author Jamie Metzl, explores the rapid technological advancements and their convergence, reshaping every aspect of human life. He analyzes how the merging of artificial intelligence, biotechnology, nanotechnology, and other innovations is transforming industries, societies, and individuals at an unprecedented pace. The book discusses the profound implications of this convergence on politics, economy, ethics, and human evolution. Metzl emphasizes the need for new forms of global governance to navigate these capabilities safely and ensure an equitable future for all. Through meticulous research and compelling storytelling, Metzl provides a thought-provoking analysis of the transformative power of technological convergence.
Unedited Notes
Direct from my original book log, below are my unedited notes (abbreviations and misspellings included) to show how I take notes as I read.
The rapid innovation and expansion of knowledge among humans of last several centuries is a function of interactions/networking/collaboration, individual person’s brain not noticeably improved/more powerful vs millenia ago, but big change is degree to which humans are connected and can iterate/connect dots, enables more time to imagine and solve new problems, from evolutionary pov we are wired to be conservative but modern comforts allow for more risks relatively but have to be conscious about it, modern evolution science and genetics come from blend Darwin and Mendel’s work of 1850s/60s, Darwin travel world and track species and evolution (ie one species does not change into another, accumulation of small changes, Mendel studied 10,000 pea plants to figure how how genes passed on, Mendel’s work not really noticed until 1900 (published 1866), european botanists sparked insights from darwin and mendel combo to deduce genetic heredity dynamics, 1902 “genetics” coined by english biologist, 1953 Watson and Crick discover DNA double helix, first genome sequenced 1980 as computing power enable, first human genome sequenced 2003 after 13 years and $2.7bn, now can do much better for <$100, “robot” coined 1920 by czech playwright, 1956 “artificial intelligence” coined, 20k essential proteins in the human body and 230mm proteins known to science—human body proteins made of 20 amino acids, “protein folding problem” as key challenge in understanding intricate physical structure of proteins (3d structure), DeepMind created AlphaFold program 2018 to figure out protein folding—given amino acid sequences of specific proteins and build models of 3d structures, struggled at first but by 2020 “predicting protein structures based on amino acids essentiall solved”, 2021 deepmind predicted 350k protein structures, made data base public and 2022 catalogue reach 214mm proteins, example: alphafold predictions enable one team to understand specific protein antibiotic resistance in 30 mins when been studying for a decade, real important impact is how developments like alphafold free up time of scientists to look forward at new innovations/theories, mathematically calc’d as millions of years “freed up”, per DeepMind founder Demis Hassabis: biology is fundamentally an info processing system—AI is the perfect description language, like math is for physics but simple math cannot explain complex/emergent nature of biology so need AI, GMO = living organisms whose genetic code has been transformed by human intervention to give them traits not otherwise expected in nature, over 90% of all soybeans corn and cotton grown in the US are gmo, 2012 Jennifer Doudna paper on crispr-cas9 made waves as new tool for gene editing—consists of an RNA that can be engineered to seek out specific site in genome where code of rna instructions for the location being sought matches the specific code in the genome, when matches then the crispr system docks in exact spot and cas9 enzyme makes a cut across the two legs of double helix ladder then pastes in new dna sequence there (think of crispr cas-9 as “find cut paste”), 2020 MIT and Harvard develop base editing and prime editing as “find edit” gene editing (no cut needed), Moderna was young biotech firm when covid started and got access to digital genome of the virus jan 11 2020, moderna vaccine developed only via computers and algos (not a wet lab), within 2 days of receiving genome they had recipe for moderna vaccine without a single wet lab experiment then 2 months later to human trials and 9 months after that first dose to public, whole genome sequencing is the key to unlock personalized medicine vs generalized medicine of today, China has made a huge centralized push to sequence citizens genomes, epigenomics = system of regulating gene expression, transciptomics = way msgs are delivered by rnas from the genomes to the ribosomes, protenomics=way our proteins are expressed, need all of these and more, single cell multinomics = understand the functioning of many diff systems within individual cells via AI, 37 trillion diverse human cells in avg person, use data to understand cancer—US national cancer institute Cancer Genome Atlas, Columbia Univ has began sequencing 100k newborns to test for genetic conditions, UK working similarly on newborn testing, “polygenic risk scoring” = advanced stats and machine learning to identify and continually weigh the influence of diff data patterns potentially associated with known disease outcomes (ie see things humans cannot)—Illumina is one company leading and DeepMind working on it too, a system of prediction of risk but in early stages, ex vivo gene therapy = removing a persons cell and genetically altering to correct/enhance then return cell to body (eg CAR-T therapies), ex vivo has risk that body may reject modified cells or attack unexpected other cells, next frontier is in vivo gene therapy = use tools for editing genome directly inside the body with precise instructions (eg crispr-cas9, base editing, prime editing), regenerative medicine via stem cells discovered 1981 in UK via mice then 2006 japanese scientisit Yamanaka show could induce differentiated cell to become undifferentiated (spark push in regenerative stem cell therapies to “trick” the body), pharmacogenomics = predict which drugs work best for individual people vs current general approach—rests in data and genomics, push to sequence full genomes of newborns led in UK and China to track/catch mutations/diseases, good today with single gene mutation conditions (“Mendelian”) but cancer and many others far more complex, “polygenic risk scoring” = uses advanced stats and ML to identify and continually weigh influence of diff patterns potentially assoc with known disease outcomes (eg Illumina leading in new AI on this, DeepMind followed), gene therapies been around since 80/90s but boom/bust—1999 man died during trial in US killed hype/FDA cracked down, ex vivo therapy example: CAR-T therapies (remove person’s cell, genetically alter, return cells), CAR-T shown promise but very expensive, in vivo being explored for range of neuro diseases, cancer, blood, etc, stem cells = drive regeneration of cells (numerous when younger especially) as can be undifferentiated then differentiated—discovered 1981 in UK and US at same time, 2006 Japanese scientist accounced “Yamanaka Factors”: found way to use 4 genes to induce differentiated adult cells to become undifferentiated (ie stem cell could manipulate), health care today is really sick care vs future want to be proactive/ongoing monitoring—also driven as hospitals/doctors paid to deliver services vs prevent future care, Israel Estonia Finland Singapore leading integrated/preventive care push for population—Finland has law requiring all healthcare data integrated with national data server so can analyze individual and whole, Finland has 8 biobanks to collect and analyze all citizens, 88% of all corn is GMO in the US—ie all changed by scientists transferring genes from one species to another (most commom being for insect and herbicide resistance), Norman Borlaugh was iowa scientist sent to Mexico post ww2 to see if could develop wheat resistant to fungal rust disease that was decimating wheat in mexico, Bourlag succeeded and then crossed this wheat with Gaines winter wheat (also dev post ww2 at Washington State College, sourced from japanese seeds), this cross breed created wheat that could be grown winter and spring and rust resistant—took world by storm, mid 1960s mexico transformed with this new wheat—huge export/surplus, IR8 rice dev at Int’l Rice Research Inst in Phillipines helped quadruple rice yields per acre across asia from 1965-2010, total human livestock consumpt from 8bn slaughtered animals 1950 to 73bn today, population est to cross 10bn before 2080 so need to grow food supply 70%, 12bn acres of land allocated to agriculture today (50% all land on earth, up 5x vs half century ago), by 2050 need 50% more land if continue as trending, from Earth’s pov it doesn’t care if temp rises really (flux greatly over history of time anyway), but concern is re humans and the ecosystems/infra we have shaped, C3 plants = wheat rice soy, transform sunlight into 3 carbon molecules (1% sunlight absorption efficiency) vs c4 plants at 50% (such as corn), papayas in hawaii were decimayer 40s thru 80s by PRV infection so early 90s scientist use recombinant dna new tech to boost papaya plant defense and worked to save papaya industry, companies innovating agriculture—Joyn Bio (Gingko Bioworks), Indigo Ag, public perception of gmo’s vs scientists is stark contrast (90% scientists approve vs <30% public comfortable) due to anti-gmo activists, failure of governments, ag companies, scientists to explain benefits/risks and general anti big corporation sentiment, even non-GMO labeled foods are modified to some extent (cheese, corn based sugar), Europe much more restrictive on gmo vs US, China keenly focused given population vs lack of natural resources, 2017 china state owned enterprise acquired swiss Syngenta for gmo innovation, china focused on gmo corn and soybean varieties, meat consumption climbing worldwide—US 6x since 1900, Europe 2x since 1960, china 3x since 1979, wild animals plummeting vs domesticated and pets, 500mm dogs on earth vs 200k gray wolves closest living relative, emissions from domesticated livestock are 14.5% of all human generated emissions—driven by cattle farming (more than trucks, cars, ships, planes combined), cows consume 25 cals of feed for every cal of beef—inefficient, only 3% of our oceans are truly protected from fishing/other damaging activities—industrialized overfishing is decimating, 30×30 plan = 30% of entire surface of earth designated as protected area by 2030 (published 2019, some progress at UN)—hard to enforce like paris climate accords, domestication of chickens led to avg size 4x from 1957 (905g) to 2005 (4202g) as we industrialize and selectively breed, 80% of salmon consumed today is from salmon farms not wild salmon—farms use large nets to enclose coasts/waterways, effort to use biotech to modify salmon, pigs, cattle such that require less environmentally damaging industrialization, ways to hack cattle so grow larger/more beef avail so need fewer in total (ie more meat per animal), Recombinetics as one company engineering cattle better able to endure hot weather, 100k people on US organ donation waitlist—push to genetically alter pig organs so compatible (xenotransplantation), 1954 Ray Kroc was milkshake salesman who founded mcdonalds, Impossible Foods founded 2011 by scientist Pat Brown—genetically engineer plant proteins to look/taste like meat, gas chromatography mass spectrometry = developed 1950s to understand multiple components of gasses—these machines underpin Impossible Foods ability to mimic burgers smell, Beyond Meat licensed tech from UofMissouri, still plant based meat only expect to make up 5% of meat consumption 2030, leading dutch vascular pulmonologist Mark Post begin tinkering with specific sten cells from cow muscle biopsies and gave nutrients to make it expand and continue to treat like a fetus by using serum extracted from cow so would grow, aug 2013 Post introduce this first cultured meat burger which cost $325k (funded by Sergey Brin), Post then founded Mosa Meats in 2015, used Fetal Bovine Serum from cows to grow—FBS from pregnant cows is expensive, takes large quantity and hard to get so defeats purpose of reducing reliance of huge number of cows, so needed alternate way to mimic FBS impact to these stem cells, Singapore ans Israel leading innovation, investment ramping across private and public sectors, big hurdle is public sentiment on cultured meats, Aleph is Israel leading 3d printing of stem cells, industry beginning with higher end food like wagyu, caviar then more affordable meats over time as cost comes down, still hard to scale this industry, other cos: Future Meat Tech (israel), Good Meat (part of Eat Just), partnering with biotech firms to create new bioreactor designs to scale meat production, industry will need steady supply of biopsies from cattle—but need fewer cattle so can live more free and less cost to farmers to care for, Post argue if went to 100% cultured meat then need 1650 cattle for all beef global needs today, need to keep farmers economics aligned/viable in this move (ie share profits), more realistic scenario could reduce cow needs 50% (from 1bn today)—this has huge impact to water and land needs too (ie benefit to environment), estimate that replacing 20% of meat sold globally with cultured could free up 5% of earths total land mass (3.1mm sq miles), first modern commercial oil well drilled in Titusville PA 1859, fossil fuels have to be heated to unlock their energy, climate change is bad for life as currently configured (not really for earth as a celestial body), biofuels = fuel from corn and other crops (ethanol made from corn), boom and bust biofuel cycles as only made sense when oil price high, also try with seaweed and algae, now using genetically engineered plants to make more efficient use of full plant biomass—progress made with algae and yeast but still too costly, global plastics industry derived from crude oil, nat gas, and coal, 1907 new yorker Leo Baekeland produced worlds first synthetic plastic as byproduct of petroleum, after crude oil extracted it is heated in refinery distillation toward so oil separates into heavier and lighter parts then make plastics and polymers via mixing components with specialty chemicals (“bakelite plastic” new vs natural plastic before and much easier to produce/mold), huge boost come world war via nylon and plexiglass needs then take off from there, today produce 400mm tons of plastic per year (2% of all human co2 emissions) but also <10% recycled (ie waste), expected plastic needs 2x by 2050, much of waste into ocean, future is plastics made cia renewable resources vs fossil fuels (85% of current commercial plastic could be converted), seaweed is one resource trying to tap (sargassum), companies Loliware, Seaweed Generation, Algopack, Carbonwave, concrete = cement mixed with sand and gravel, cement made via dig up/extract from blasting large quarries then transport to crushing plants to be ground to pwoder then burned in massive kiln then ground again—huge use of energy/emissions, 8% of all human greenhouse emissions from cement production (3x aviation), spider silk is the strongest polymer on earth but hard to extract vs silkworms—now attempt via bioengineering, data centers are 1.5% of world energy use and emit equal carbon dioxide to all US commercial air travel, data growing exponentially—by 2040 under current methods lack requisite wafer grade silicon for storage needs, DNA data storage = revolutionary vs current silicon computer chips, dna is 1mm times denser than silicon, estimate a refrigerator box of dna could store all the worlds data today, need to translate a/g/c/t to 1s and 0s and vice versa, leverage genome sequencers as used in labs today where read letters and AI algo translate to 1s and 0s so computer can recreate original data, but today dna storage multiples more expensive but can fall exponentially, key companies illumina, msft, western digital, twist bioscience, cortical labs, covid outbreak just happened to occur 1k miles from natural habitat of horseshoe bats in a city with worlds largest collect of bat coronaviruses with a spotty safety record and secrecy was exploring how viruses infect human cells, and 1.5 years pre pandemic wuhan institute of virology apply with consortium of global scientists for 14mm to collect sars-like viruses in remote areas then see how better infect human receptors via reengineering, 1977 viral flu outbreak in china spread to ussr killing 700k people globally, 2019 10k people died in china after vaccine factory accident that china media tried to cover, china focused on world class virology capacities and public health investment—2015 completed with french help construction of first highest security biosafety level 4 biology lab The Wuhan Institute of Virology (operational 2018), now 69 of these level 4 labs in construction/exist in 27 countries—more risk of accident and sub-standard practices, also now easier and easier for biohackers at home/school to DIY virology testing, name of this testing to make more infectious is “gain of function research” which has stigma now despite centrality to vaccine development and gene therapy, gene drive = engineer genomes of sexual reproducing species so engineered partner passes along its editing tool kit to partner and made to be dominant trait so passes to future gens with certainty quickly (can backfire, used in malaria field for example)
