Multi-Disciplinary Reading - Book Reviews

phreakv6 · August 20, 2025, 4:45am

Say Nothing, Patrick Radden Keefe, 2018 - I barely knew anything about Irish history, or their relationship with Britain, other than that there was/is some friction there. The book enlightened me on a topic (amusingly understated by the Brits as ‘the troubles’) I knew nothing about and it did so while not boring me to death. We have a natural tendency to pay attention to stories so when you tell history through character arcs, it tends to stick and that’s what the author mastered in this book.

The book opens with the kidnapping of a mother of 10 and while unraveling this mystery (which gets resolved only towards the end of the book), we understand the IRA, its ideology, actions, Britain’s involvement in Northern Ireland, the friction between Protestants and Catholics, the Irish Republic and the strife torn Belfast.

This sort of thing has played out in various forms where the concept of nation state is sometimes at odds with ethnic or religious identities like with Sri Lankan tamils, the kurds, chechens and palestinians. The general takeaways for me are how clannish human beings are to their land, language, religion and culture - even when the differences might be marginal, how easily misled youth might be to a cause to the point of following instructions to murder and maim and how indifferent life is to your beliefs and how glaringly stupid the deeds of yesterday stand to the beliefs of today. 10/10

Gaurav_Agarwal · August 21, 2025, 9:37am

Hi Phreak,

Does book points to or explain general human behaviour or is particularly linked the behaviour of Irish people?

Of course our cultures, traditions and DNA differs!

Anand_Sanyal · August 22, 2025, 7:56am

I didn’t read the book, but saw the series on Jio. Gripping. I suppose need to read the book also.

phreakv6 · August 22, 2025, 9:03am

The Great American Drug Deal, Peter Kolchinsky, 2019 - The book has a very interesting premise that high patented drug prices shouldn’t be seen as profiteering but as an investment that gives humanity generics after a period of time (patent expiry) for eternity - like a mortgage payment that lets you have the house at the end of payment term forever. The main argument being that people see these (high innovative drug prices) as rent when it should be seen as a mortgage payment towards the generic drug asset that benefits all society and hence cost should be borne by whole of society and not just vulnerable minority of patients

My notes -

If a bone building drug reduces hip replacement surgeries (which cost $40k now), the value of it should be seen through billions saved in avoiding these surgeries. Its value is passed through to generations when it goes generic
When valuable drugs like Lipitor (atorva) go generic, its like a house in NYC becoming affordable overnight.
Biotech social contract - Universal healthcare with low/no out-of-pocket costs while patented drugs go generic without undue delay
Lisinopril (generic) - 100 million prescriptions per year. Came to market in 1987 as Prinivil (Merck) and Zestril (AstraZeneca) - improved versions of generic Enalapril (all ACE inhibitors). Enapril itself was a better version of Captopril (first in class).
If insurance is paying for a drug, it requires the pharmacy to fill a zestril prescription with lisinopril generic to save money. The insurance company pays a fixed amount to pharmacy for generic, so pharmacy’s best interest is in sourcing the cheapest generic
Generics account for 90% of all prescriptions written in US by volume but only 28% by value (2018 data). Drug spends overall are only 9.4% of overall healthcare spend (so only 2.5% of overall healthcare spend in generic - there’s nothing to save here by tariffing)
Enrolling a patient for trials is extremely expensive - $200k per patient is common for cancer studies
Unlike doctor’s visit, MRI, hip replacement, nursing care etc. (services and procedures), drug cost come down over time. Its services that account for over 70% of healthcare spends. Services never go generic
If we make all branded drugs go generic with immediate effect, it could save just 6.7% of healthcare spending (1.2% of US GDP) - but it will mean no new drug ever and our mountain of generic will cease to grow
GoodRx and Drug Patent Watch compile list of drugs that could go generic in near future
Drug companies offer insurers rebates over list price. This rebate or discount are confidential and never disclosed. What public sees is list price and is often very misleading.
US makes the most of generic drugs. They make up 90% of prescriptions filled. In Switzerland its 17%, France 30%, Spain 47% and Germany 80% (Can see why US spends more on branded drugs. But some parts of Europe like Germany may have to pull their weight which is what Trump is pushing for)
US spends 2.4% of GDP on roads, water infra. At 1.3% of GDP, investing on new drugs is a worthwhile investment
We don’t charge copays for security of police dept or safety from fire dept. Why should healthcare be any different?
QALY (Quality adjusted life year) estimates reduction of life and quality of life for a patient with a particular condition in $s - Suggesting that its not worth developing drugs that can only be justified if priced higher than QALY. But it misses peace of mind in knowing eradication of a disease/condition for generations
From car to internet to statins - society is incapable of appreciating or anticipating innovation (Hence following models like QALY is deeply flawed)
Smokers save taxpayers money on medicare and social security by dying early - so should be encourage smoking?
A 20% drop in US branded drug prices (say $24k instead of $30k) would cut industry’s profits in half - but it would still be unaffordable to patients. At the same time it removes possibility of reinvestment of profits into future drugs.
Sick are subsidising the cost of the healthy through copays and deductibles today (the healthy may make use of generics in the future). Copays are justified by payers as a measure to prevent patients from seeking unnecessary treatments and services
Out of pocket costs mean that many people with insurance still can’t afford healthcare which defeats the purpose of insurance (91% pay < $500 but 2% pay over $1500 and are the most affected and drive the avoidable disproportionate outrage)
Emergency care doesn’t need insurance (uncompensated care) and large part of it is borne by taxpayer and rest is charged by increasing premiums. This is very expensive ($85b) and could have been avoided by spending upfront in prevention (by cheaper universal insurance without copay).
Medicaid (funded by state and federal subsidy) covers poor. Medicare covers older Americans (65+). Third one is for veterans and armed services - all funded by federal tax dollars - this leaves a large part who aren’t poor, old or veterans without coverage - some part of these are covered by pvt insurers by employers but a large part are uninsured (13%) - but these still wont be turned down at the ER (Emergency Response)
Bureaucracy in healthcare costs over $500b and is 2.4% of GDP (twice branded drug spend). Bureaucracy never goes generic
Insurance companies use Prior Auth ¶ and Step Edit (SC) to prevent over-utilisation. Docs have to do lot of paper work to get a drug to their patient if that drug needs PA (Paperwork discourages them). Docs have to use a generic or cheaper med before using the drug if it requires SC. Even if PA and SC hurdles are crossed, there’s still copay (even for generics!)
Cost sharing (Copay and Deductibles) allow lower premiums to everyone is the justification given by insurers - but whole point of insurance is for the fortunate to subsidise the unfortunate (Eliminating copay will increase premiums for everyone by just 2.3% and increase tax by 0.35%)
Pharmacy Benefit Managers (PBMs) extract profits whenever patient gets a drug, even if that spend is out of pocket. Sometimes its a rebate from the drug company and other times its from deductibles - its like mechanic paying insurer $50 for a $400 repair bill which you paid completely out of pocket since your deductible is $500.
Diabetics are likely to stop taking treatments earlier in the year when they are still paying out of their deductibles than later in the year after they have hit their out-of-pocket maximums. (Have noticed this in Esperion’s calls as well for their LDL-C drug)
The system only optimises for reduction in one drug vs other but not if a certain drug would reduce overall healthcare spends (esp services and ERs). This might only be possible in a single-payer system
Insurance companies spend 80-85% on patient care (MLR or Medical Loss Ratio). Of the 15-20% left, they must take care of operations and can keep rest as profit. This incentivises them to increase overall healthcare spends than decrease it
Payers (insurers) set budgets annually and price their premiums. If a blockbuster drug launches and they fail to anticipate demand, they stop many patients from being cured until the next year’s budget cycle (Again saw this repeatedly in Esperion calls). Customer mobility (users switching plans) also means not paying today might mean savings down the road (patient is another payer’s problem next year)
In-class monopoly - When a drug is the lone one in its class and is very good at it that no competitor can even get patients for trials. (Eg. Cerezyme for Gaucher’s disease, Roche’s Rituxan anti-CD20 antibody for leukemia, Norvatis Gleevex for CML, Merck’s DPP4 inhibitor Januvia for type-2 diabetes)
Long In-class monopolies are now rare - when a company launches a first-in-class drug, others are often not behind with their “fast-follower” drugs. While they are derided as non-innovative or me-too drugs - sometimes they are better on side effects or some people might respond better to them, so they serve a purpose.
Fast followers allows payers to play drugs off one another and reduce costs (for preferable formulary status - as is happening with wegovy vs zepbound). Eg. Two PCSK9 inhibitors (Repatha and Praluent) within weeks of each other, SGLT2 inhibitors (Empa, Dapa, Cana gliphozins) and CAR-T treatments (Escarta and Kymriah)
Fast-followers are rare in smaller markets so innovator may enjoy a longer monopoly
Payers (Pharmacy Benefit Managers or PBMs in specific) get rebates from innovator and so keep biosimilars out (there’s not as significant a price difference as in small molecules so free markets dont work). Biosimilars can’t match those rebates so the innovator doesn’t lose market share as quickly (Super interesting viewpoint)
CAR-T cell therapies and gene therapies are not genericizable by any approaches conceivable today
Sometimes me-too drugs are used in combination and dont compete - as in HIV where two polymerase inhibitors and one integrase inhibitor work together (LTD or TLD - which forms bulk of Laurus ARV generic business)
Sometimes multiple drugs are used in combination for a disease and not having capability in one might cripple a drug company - thats why every oncology company has a anti-PD(L)1 antibody. Not having one is a non-starter
Payers influence market share by tweaking copays even against a preferred drug (as in Wegovy vs Zepbound). This is esp. crucial when patients take the drug over longer periods. Sometimes good enough drugs are used as leverage against better drugs (Orforglipron might be used as a leverage against Wegovy, the way Wegovy is used as a leverage against Zepbound)
PBMs negotiate drug prices for payers (sometimes Payers own the PBMs, so they are just agents) - they retain a portion of the rebates. Express Scripts, CVS/Caremark, United’s OptumRx have 80% PBM market share. A drug company that offers lower list price thus threatens the PBM model (Principal-Agent problem). This is where bulk of US healthcare spends are sunk (what Trump should be going after, instead of prices of branded drugs or generics). Patient’s out-of-pocket costs are linked to list prices
Drug companies RoCE has halved since 90s but profits of middle mean - PBMs, insurers, wholesalers, pharmacies have clinbed from 20% to 40% of healthcare industry’s profits (and 2/3rd of healthcare industry’s profits)
Antibiotics need a guaranteed payment model to encourage innovation since market size is small and patents expire before they can recoup investment
Some drugs like Daraprim (malaria) cost $1/pill and have only 8000-12000 prescriptions in a year. GSK loses money just selling the drug. In cases like these, drug companies prefer to offload them rather than risk raising prices (bad publicity) - a smaller biotech that buys it though can raise prices (no reputation to worry about). This is what leads to price-jacking (Daraprim acquired by CorePharma and prices were raised to $13.50/pill. Pharma bro Martin Shkreli’s Turing acquired CorePharma and raised per pill price to $750 and became infamous)
Compounding pharmacies are holdovers from a pre-FDA era (thousands of them in US). They are allowed to purchase RM necessary and make and sell drugs. These homemade versions introduce dosage errors and contamination
Gilead developed the Hep-C drug Harvoni and priced it at $94500 in the US. However in 91 developing countries it was priced at $600-900 (Indian generics manufacturers were allowed to make and sell). This is a way in which companies maximise profits as higher prices will shut these markets out and lead to zero sales (some profit is better than no profit)
A drug that costs $20 to manufacture might sell at $1000 in the US and $50 in Mexico. Given the cost of making the drug - its profitable either way to the drug company. So they try to maximise profits by maximising volume by regulating price (I’ve been thinking about on similar lines for Orforglipron pricing in developing markets)
Affordability is a function of what customer can pay and not cost alone
If reference prices are enforced (Europe prices matching US), it will end up exporting higher prices from US to EU, rather than reduce US prices. In US treatment cant be denied whereas, in the EU it is. US is a price-insensitive market because of policy. This will shut out EU markets and drug makers will try to recoup profits in the US thereby rising prices.
Price controls won’t work because in a company’s portfolio, one or two drugs might sustain the pricing of rest of the portfolio, like in a restaurant with few great dishes priced higher
DTC (Direct to consumer) ads were banned by FDA in 1983 but was brought back with stricter regulations (cant be false or misleading, has to balance risks and benefits etc). It equips patients with better ability and understanding but can also lead to over-diagnosis and over-treatment
Epinephrine used to treat anaphylaxis was sold in vials and then EpiPen was approved in 1987 but it was priced too low and was barely profitable ($50/device). Mylan acquired it from Merck. Raising prices 3x and advertising to raise awareness, increased sales to $1b/year and also benefitted millions of patients (Benefits of DTC ads)
Chiral compounds are structural variants using same chemical formula (Enantiomers in chemistry - which are mirror images). Prilosec (Omeprazole) was a mix of two enantiomers - one effective and another dormant. AstraZeneca launched Nexium (Esomeprazole) which was purer with only the working enantiomer and extended patent. Esketamine (J&J) and Ketamine is another similar case. (BlueJet has been talking of Chiral chemistry, so found this interesting)
Pfizer’s Lyrica (pregabalin) went generic in 2019. Pfizer patents for once daily version wont expire for another 8 years - so it has Lyrica CR competing with Lyrica (generic) in the market. By offering generous copay assistance to paitents and rebates to PBMs, Pfizer can make sure Lyrica CR is preferred over generics (though generics are just as good and just have to dosed twice)

This book is by far the best I have come across to understand the US healthcare landscape when it comes to drugs (provider, payer, patient ecosystem). It also covers a lot of details on drugs themselves, so you get to learn a lot about pharma sector as well. Its backed by a lot of research, data and experience of the author being a biotech investor and also a scientist. Its also relevant in current times to understand feasibility of tariffs on branded drugs and generics. 11/10

pahaaadi · August 23, 2025, 4:27am

Very interesting. So in nutshell the real cost drivers in US are services, bureaucracy and middlemen, not drugs.

This makes sense on why UK is better. NHS is both negotiator and payer that helps them cut the middlemen. Unlike US where healthcare services operate in free market, these are centrally budgeted in UK.

It makes me anxious that we are slowly copying the free market practices from US in India. Fancy 5 star hospitals charge whatever insurers can’t push back on. I may be wrong (I am from village side :) but health insurance wasn’t in the first chapter of Personal finance 101, a few decades back.

ravilives · August 23, 2025, 5:46am

Throughly impressed by the wide range of books you read and have synthesize with clarity. Amazing.

I personally have a good collection on a range of books on various topics bought over the years, but it more or less stops there as not able to read or complete ( due to losing interest or lacking speed or reading multiple/ jumping between books) more than 1 in 10 books I buy😀

Kindly request if possible please if you could share your framework on how you approach reading- when you read/ the routines - how many you read at a time-any tips you could share in general. Would be really happy to hear your perspectives.

phreakv6 · August 24, 2025, 5:46am

Red Notice, Bill Browder, 2015 - The author having found some ridiculous bets in Poland and Eastern Europe after fall of communism (like 0.5 P/E from privatisation), finds similar bets in Russia (look up voucher privatisation in detail - incredible stuff) when it opened up to the world post the collapse of the Soviet Union. With his experience in consulting from BCG and banking at Salomon brothers, he sets up Hermitage Capital specifically to participate in similar bets in Russia. As an activist investor, he influences outcomes by exposing corruption in several privatisations. He absolutely thrives in bringing transparency when it helps his investments. Putin comes to power and goes after those oligarchs and consolidates power with his image of cracking down on oligarchy.

Things fall out however when he gets intertwined in higher echelons of oligarchy with his investments in Gazprom and Surgutneftegaz when he touches oligarchs closer to Putin (Abramovich, Deripaska, Potanin) by exposing how corrupt Gazprom officials were selling its assets for cheap. Browder thus becomes a liability after Putin’s power consolidation. This leads to a denial of visa and deportation. Things go down south very fast however when ownership of subsidiaries of Hermitage are stolen (using documents seized in a raid at Hermitage) and used to engineer a massive tax fraud by officials of the interior ministry (Kuznetsov, Karpov).

While Browder manages to get all employees of Hermitage out of Moscow, the person who exposed the fraud ($230m), who himself was just a tax lawyer (Magnitsky) is apprehended by officials and what ends up being massive human rights abuse, dies in custody due to torture. The last half of the book deals with how Browder uses his influence in the White House gets “Magnitsky Act” passed - something the US govt. can use to sanction officials involved in human rights abuses anywhere in the world.

I found it fascinating how driven and passionate some people can be, towards a cause (both Browder and Magnitsky). The book covers some incredible trades the author comes across - the sheer undervaluation that you can only dream of. It also covers a lot on how lobbying works in the White House. Overall the book was gripping page turner mixing finance, geopolitics and human rights. It is however possible we are only hearing one side of the story - Putin has made several allegations against Browder after the book got published. There’s perhaps truth and lies from both sides so read with a skeptical but open mind. 10/10

Shripad · August 26, 2025, 6:10am

Very well written review. Liked your comment “There’s perhaps truth and lies from both sides so read with a skeptical but open mind.”. It is very crucial when we read everything with open mind.

phreakv6 · August 27, 2025, 12:25pm

Apple in China, Patrick McGee, 2025 - Though the book is about Apple’s involvement in China, it covers a lot of Apple’s history as well to set the context of its evolution. It is a fairly balanced take based on facts. Whatever you may think of Foxconn, you can’t help but be impressed by the passion, drive and capability they have displayed (Terry Gou esp) that we can only dream of. Nobody pulled Apple into China - it just evolved step by step over a decade with Foxconn taking its manufacturing to mainland China and later on being displaced by the Red Supply Chain (mainland Chinese companies) due to politics

My Notes -

Shortly before Jobs returned back to the helm in 1997, Apple had factories in California, Colorado, Ireland and Singapore. Production moved to South Korea, Taiwan and then to Mexico, Wales, Czech Republic and China (Jobs always wanted manufacturing in the US)
Apple opened its first store in Beijing in 2008 (for the Olympics) and did sales of $1b, by 2012 it went up to $23b (massively underestimated demand). Apple equipment in the country was $7b by 2012 (its capital investments)
Since 2008, Apple has trained 28 million workers in China - more than entire labor force of California. China brilliantly played its long-term interests with Apple’s short-term needs. In ‘99 none of Apple’s products were made in mainland China. By 2009, almost all of them were
By 2015, Apple’s investments in China reached $55b per year! To put it in perspective, the CHIPS act envisions $52b investment in the US over 4 years.
China demand indirectly supports between 1-2.6 million jobs in the US across all industries. Apple alone supports 5 million jobs in China
Apple expected to do $414b in revenues in ‘25 (~$100b in profits). Since ‘07, iPhone alone has done $2T in sales. Google pays $20b/yr just to be default search engine on the iPhone
Total Apple devices in active use surpassed 2.35b (1.4b iPhones). These are richest quintile of people in the world that Apple has direct access to for promotions
Apple’s move to change how third parties track users killed business of advertisers like Google/Meta but Apple’s own privacy-first ad business went up from $1b in ‘20 to $30b in ‘26E
Foxconn’s revenues today are greater than META and NVDA combined (as of ‘24)
iPhone operations - 200 lines each averaging 3330 units/day (~$250m/yr)
Until 2019, iPhones were manufactured by Taiwanese groups working in China - but this has changed considerably since with mainland groups with political backing taking over
Only a dozen MNCs earn > $10b in China. Apple tops the list with $70b (almost 17% of revenue)
In 1981, Jobs was critical of IBM’s PC because it looked like crap but it was built in under a year with pre-existing parts with software by Microsoft (DOS). Jobs missed how brilliant the idea was (of a open platform). IBM had zero revenues in 1980 from PC but by 1982, had 16% market share outselling Apple two to one. By 1984, it was 3x Apple (open platform and giving control of software to Microsoft meant IBM’s brilliant strategy was still a dud for it as others competed it away on price)
IBM achieved market dominance through innovation in circuit board production (outsourced to SCI). SCI pioneered automated circuit board manufacturing and EMS (electronics manufacturing services). SCI sales was $45m in ‘81 and $500m in ‘85. Later came Solectron, Flextronics, Celestica and Jabil
In 1985 Jobs expected Mac to do sales of 80000/month. It did just 5000/mo. Demand was so poor that Jobs and Sculley didn’t give product-wise breakup to Wall st.
Mac was too expensive in ‘85 and needed a killer app to justify price. Apple designed a Laser printer (LaserWriter) and manufactured in Japan by Canon (Adobe PageMaker as the killer app). This was the first Apple product manufactured by third-party. Laser printers used to cost $30k then. Apple bundled Mac + LaserWriter at $10k as a bargain. It was however too late to save Jobs (quit in ‘85 due to underperformance)
Apple collaborated with Sony in early 90s for PowerBook. Sony crammed innards of a Mac into a portable form-factor of a 5lb laptop priced at $2300. Apple was astounded by Japanese precision and quality
Japan lost its edge as an affordable manufacturing base in ‘94 when dollar lost half its value against yen in 15 yrs. Apple Newton manufactured by Sharp was a disaster due to labor shortages and soaring currency. Sharp had a condescending view of Apple engineering as well (likely how Huawei sees Apple today) and came up with its own rival PDA.
US was bureaucratic (this is mid 90s) and falling behind Asia on cost and quality. Turnaround time of 1-2 days in Asia would be 2 weeks in US (the idiocy of Trump trying to make iPhones in the US)
Taiwan rose as a industrial base with Japanese investments of skill and capital to combat rising wages at home. Taiwan used to be a Japanese colony for 60 yrs until WW-II so culturally it was a good fit.
Apple made $11.1b in Sept ‘95 (TTM) up from $7.1b 3 yrs prior. But Windows ‘95 launch and subsequent panic and price cuts led to halving of revenue to $5.9b in ‘98. They didn’t have a hit product, was saddled with debt and company was sinking fast. (Amelio was CEO)
Jobs started NeXT after leaving Apple and in 8 yrs (by ‘93), having learnt nothing from Apple mistakes - had a expensive product (NeXTcube), expensive manufacturing and a even more closed ecosystem than Mac making it a complete flop
In ‘86 Jobs acquired Star Wars maker George Lucas’s graphics group for $5m and renamed it as Pixar. Pixar had a high-end image computer ($135k). By ‘93 Jobs sold off the hardware business and cut staff by 70% and narrowed efforts to story-telling. Toy Story was a smash hit (first full length computer animation). Jobs took Pixar public and became a billionaire
Jobs made a similar decision at NeXT computer of focusing on software - OS and WebObjects. Apple paid $400m for NeXT for the OS which was 5-7 yrs ahead of everyone else (only way to fight Win ‘95 onslaught). Jobs came back as well to lead Apple.
Apple chose LG to manufacture iMac (Jobs’ first product upon return) since it was already making its CRT monitors and iMac was a CRT monitor with a computer stuffed inside (Jobs wanted Japanese but they were expensive). Japan had invested significantly in South Korea since the 70s like it did in Taiwan (again, a former colony and similar culture). LG bid that it would pay for tooling and bear the cost to make the prototypes (something most EMS guys working with Apple will do in the future)
iMac was a big hit. It sold 278k units in 6 weeks (Aug-Sept ‘98) and 800k units by Christmas leading to 5th consecutive quarter of profit since Jobs’ return.
Apple cut 450 positions in Ireland and 350 positions in California (even Jobs came around eventually to give up his fancy for US manufacturing). LG setup factories in Singapore, Sacramento and Ireland and later in Wales and Mexico. iMac demand was overwhelming
Inventory would be on LG’s books until it crossed the yellow line in the factory and became a finished good (Another recurring theme in all Apple manufacturing is vendor-managed inventory - combined with just-in-time manufacturing)
Apple tried to get prices lowered as iMac volumes ramped up (another recurring theme) but LG stood firm. It thought itself indispensable having worked with Apple on the design and owning complete manufacturing. It made the cardinal error - In contract manufacturing client comes first. A little known Taiwanese company Hon Hai precision (Foxconn) was reverse engineering the iMac and Terry Gou managed to get the contract from Apple (by dialling Tim Cook)
Taiwan faced same issues Japan did (rising labor cost) and did same things Japan did - build factory in mainland China like how Japan built in Taiwan and Korea. Just like Japan, Taiwan had a cultural fit with mainland China. Taishang (as these Taiwanese were called) brought export-driven low-cost manufacturing
Michael Dell visited Terry Gou (Foxconn) in Longhua campus and was blown away by the austerity of Terry’s office (just a shed with a plastic table) - it was clear that every dollar Foxconn earned was going to the production line and not to marble floors in the reception - it had world class machinery in the factory (subsidised by local govt)
Gou’s ten levels of breaking down a product - Level 1 might be a knob on a TV. Level 2 might be a stamped part that goes into the knob (only mechanical interfaces, no complexity). Level 3 might be some wires and connectors going into the knob. Level 4 and 5 complex subassemblies. Level 6 - PCBAs. Level 7 and 8 - putting together a PC or laptop and Level 10 would be finished good ready for shipment (He would always aim to capture complete value chain)
In 1980 Shenzhen was a fishing village with population of 70k. By 1990 it was 1.7m. By early 2000 it was 7m (100x in 20 yrs). Terry Gou ranks second only to Deng Xiaoping in transforming China
Soon firms assembling electronics hired expensive designers, invested in R&D to design PCs as white label products that brands like Compaq or Dell could pick and rebadge it (ODMs like Asus, Acer, Inventec etc)
Gou shunned ODMs because he did not want to hire expensive talent. Instead he wanted to be vertically integrated and control BoM (bill of materials) as much as possible (OEM instead of ODM)
Gou wanted to take supply chain off the client’s hands and make money in sourcing components (more parts he controls, more opportunity to manufacture or trade). So he would give tooling away for free (bear all upfront cost for molds, dies, fixtures) and do final assembly for very cheap
Cook’s ruthless efficiency in materials management in the 12 yrs at IBM almost sent Apple to bankruptcy. Jobs hired Cook because he wanted someone good at all the things he was bad at.
When Cook negotiates with a supplier - he wouldn’t try to figure out “Whats a reasonable thing to ask them?” instead he would ask for everything until the vendor says no (and Terry wouldn’t say no).
Gou figured out earlier than anyone that the value of working for Apple wasn’t the profits, it was the learning. Gou offered to build iMacs for Apple at $40 less per unit than LG. Foxconn was merely a connector company at this point. Foxconn built iMac tooling in 25 days as per commitment impressing Apple. The standard was 12 weeks for tooling
Foxconn opened a Czech unit where workers put in fewer workers and were repsented by a trade union. They threatened with a strike before Christmas - Foxconn deposited bonus within a week and then closed division within half a year closing Europe for Apple. Within a decade in the same way, all production would move to China - with all final assembly done by one vendor - Foxconn
Apple’s engineers found precision machining capabilities at hard-drive makers in Thailand, turbine blade factory run by Singapore Airlines, watch makers in Switzerland and applied those technologies to their products (started with flat screen iMac G4)
The iMac stand was forged, machined, heat treated, polished and chrome plated with hardened 17-4 stainless steel - soon Apple cornered the entire world market for the material
The amount of time in trail-and-error work it took to design the iMac across China, Japan and Singapore forced Apple to urge partners to set up operations in China. This allowed Apple execs to go from injection molding, sheet metal stamping or machining facility to assembly site in hours/minutes instead of summoning someone from another country (Apple follows this to this day, now insisting all Indian iPhone vendors to setup shop in South India)
When iPod (2nd product under Jobs after iMac) was launched it sold 125k devices in first month but sales petered out to 20k a month within few months. The 2nd gen iPod launched within 9 months didn’t sell much either. Problem was that iPods required iTunes/Mac to function. Jobs was hoping iPods will sell more Macs but it backfired :-) iTunes for windows catapulted the 3rd gen iPod into a blockbuster (manufactured by Inventec in Taiwan). The iPod 3rd gen demand boom caused Inventec to open second factory in China (Apple was squeezing Inventec on margins with inc. in volumes - another recurring theme)
Foxconn reverse engineered iPod without specs and unveiled it to Apple to show capability (you have to be floored by this level of ingenuity and drive). Apple was pissed with Inventec by then (on slow ramp up and margins) and Terry was willing to work for pennies on the dollar. Gou was again going for capability and not profits - he would rotate workers on Apple projects to maximise the learning
Foxconn had a stronghold in metals in 2002. It was the go-to source for all things Aluminum starting with the iMac cinema display. Jony Ive was impressed by the stainless steel back of the original iPod but it would attract fingerprints (it forced the user to polish the unit, developing an unconscious nurturing connection) so when Jony wanted to shift to anodized aluminum for iPod Mini, Foxconn was the natural choice in 2004 (Apple’s bestselling product at that point)
Gou put up greenfield capacity for iPod in just 6-9 months - installed new machinery, hired tens of thousands of people and got production line up and running (9 months, even permit for site cant be acquired in America). iPod sold 937k units in 2003 to 4.4m in 2004 and 22.5m in 2005 (market share from 33% to 82% between ‘02 to ‘04)
Jobs attributed his cancer to the volume of work he took up in ‘97 running both Apple and Pixar
Apple abandoned sampling (testing random samples) in QA process and instead tested every iPod through costly tests - two-thirds of line was dedicated to testing and validation. Nobody would do that because of the cost (A habit they will carry on to iPhone)
Foxconn intended to squeeze Apple on profits (by controlling supply chain) but Apple instead squeezed Foxconn. In 2000, Foxconn PAT margin was 10.6% and dropped to 4.6% in 2007 and then to 2.4% in 2011. Between 2007 and 2011, revenues went up from $53b to $107b but prfits only went up from $2.41b to $2.53b (and Foxconn was fully responsible for warranty on iPhone!). Same time Apple margins jumped from 1.1% in 2003 to 26.7% in 2012
I would prefer to talk inventory in hours and not days - Tim Cook
For the unibody Macbook pro machines out of a single block of Aluminum - Apple bought 10000 CNC machines in a single year - costing between $100k to $500k each (deal with FANUC to buy entire pipeline for years, shutting down competitors)
When Apple co-created a process with a supplier, Apple owned the IP for it
It doesn’t make sense to hire smart people and tell them what to do; We hire smart people so they can tell us what to do - Jobs
“Fool me once, shame on you. Fool me twice shame on me” - American version. “If I cheat you and you don’t catch me, its your fault” - Chinese version
Since products made in China were for export only, Apple had to route them to Singapore and import it back to sell in China
Throughout the 2000s, manufacturing labor costs rose by 15.6% a year. Attrition was 300-400%, sometimes 25% a month and avg tenure only 68 days (competition for labor from Xiaomi, Vivo, Oppo and Samsung was very high). Churn around Chinese New Year was the highest at 50% (workers won’t return back from home to the city)
Samsung was making chips, flash memory, displays batteries for Apple until they copied the iPhone which pissed off Apple. This led Apple to work with TSMC. TSMC couldn’t commit to whole order and took only 50% (rest still with Samsung) and borrowed $7b to invest $9b to bring up a new chips fab in 11 months. Later shifted whole chip supply to TSMC (and risking itself with Taiwan situation)
Plenty of vendors went bankrupt when the technology they supplied was obsolete in the next generation of product (later they would insist vendors not make more than 50% from Apple to prevent negative headlines)
ToT (Transfer of Tech) as explicit condition for market access was illegal as per WTO. But when Siemens, Bombardier and Kawasaki entered China for high-speed rail, thats precisely what they did for local market access and soon Chinese companies caught up fast and were competing with them in the US on copied tech
China is a regionally decentralised authoritarian regime. While Beijing set the goals, how it was achieved was left to the provinces, municipalities and countries. This led to massive cauldron of independent experimentation. What worked in Guangdong would then be replicated in Shanghai and so on.
China might be mimicking the West but it wasn’t actually becoming Western. The assumption they would gain Western values leading to collapse of CCP was intellectual laziness (back in 2000s)
In 2016 China created a limit for companies temporary workers to just 10% - almost all of Apple’s vendor were in violation. Beijing creates policies like this when it wants to arm twist companies towards a wink-wink nod-nod arrangement and compel companies towards favors. Starting 2013, China started tightening the screws on Apple
Qualcomm which was collecting license fee on every smartphone sold using their chips was arm-twisted into paying a $1b fine, reducing license cost and losing IP.
iPhone accounts for < 20% of smartphones sold around the world but accounts for 80% of the industry’s profits
Apple would constantly send and train Lens Technology’s (company which cut corning gorilla glass) competition to keep them on their toes
Apple Squeeze - rigorously train local partners and give away manufacturing knowledge and help them scale and maintain quality but squeeze them for soul crushingly low or nil margins. iPhone margins was 33% for Apple while Oppo/Vivo/Xiaomi were earning 7%/6%/2% (with tech they learnt from Apple)
Apple demanded to access every detail of its vendor’s operating costs, from wages, cost of dorms, bill of materials and expenses of its machinery. Apple most times had better sense of its suppliers’ op. costs than the suppliers themselves. It would disintermediate vendors by procuring supplies for them and obscuring costs (like they did with Foxconn)
Because Apple didn’t want its vendors to be solely dependent on them (50% sales from others), they would strangely end up driving up Android sales - even encouraging their vendors to do so (because of negative news headlines when vendors went kaput losing Apple business). This is how Huawei, Xiaomi, Vivo and Oppo thrived. This ended up being even better than a JV based tech transfer
To CCP, Apple pledged to invest $279b in China over next 5 years in 2016 (they anyway invested $55b in 2016 - so they were just marketing) - to campaign for Chinese govt. to go easy on them (To put it in perspective the Marshall plan to spur development of 16 European countries post WW-II was $13.3b or $131b in 2016 dollars)
Catfish effect - put a catfish among sardines and they survive better and taste good (Norwegian fishermen would do this while transporting them). China saw Tesla as the catfish for its fledgling EV industry in 2016 (and we know how this went). We should probably let Chinese companies into our country and nudge our high RoE sardines to survive
Luxshare Precision won the order to assemble Airpods in 2017 - at cost with no margin at all on one condition - that Tim Cook visit her factory and be photographed on the assembly line. Two reasons 1. Owners stood to make more from stock market - predictably stock was valued at $38b even surpassing Foxconn, though Foxconn was 13x more on revenues 2. Luxshare wanted govt attention - to get factory land and other incentives for cheap.
Luxshare went on to make Apple Watches in 2019 and 2021 and even iPhone (first mainland business to do so). Between 2016 to 2023, its revenues soared 1455% to more than $32b (3/4th revenue from Apple). The faucet was turned off for Taiwanese as the mainland Chinese (the Red Supply Chain) would work for even cheaper and politics was forcing Apple’s hand
In top 200 vendors for Apple, there were 16 Chinese companies in 2012. 41 in 2019 (surpassing US) and 51 in 2021 (surpassing Taiwan!). Even this is misleading as the Chinese companies were many times bigger than the others. Apple’s whole labor demand in Vietnam was 45k (across 14 vendors) vs 72k at a single Chinese supplier Biel Crystal (glass). Even the non Chinese companies had presence in China (151/200 top had)
Other mainland Chinese suppliers than Luxshare - BYD Electronic (hardware enclosures), Goertek (Airpods and Airpods pro) and Wingtech (Mac mini and Macbooks) - collectively $6b revenue in 2015. $25b in 2020 and expected to do $52b in 2025 (Apple shifted from Taiwanese leaders like Foxconn, Wistron, Pegatron and Quanta and the workers shifted as well from these companies to the red supply chain)
Huawei Mate started nibbling away at iPhone market share in 2018 leading to a dud launch of the iPhone XR (first revenue warning in 16 years). Mate was awfully good, outshining Apple in features rather than just price. Within a year Huawei was outselling Apple not just in China but globally (Even in premium market of $600-800 phones, Huawei market share soared from 10% in 2018 to 48% a year later)
Trump sanctions on Huawei preventing Google services and cellular chips from being sold to Huawei collapsed its business and gave a huge lifeline to Apple (Huawei share of Chinese market collapsed to 7%)
Mac Pro assembly in Texas by Flextronics was a disaster. If screw is too short and you needed a longer one, you call someone and you will have 1000 at the factory tomorrow. In Texas it would take 2 months. Apple had to ship Chinese engineers to Texas to complete the project
Trump tariffs of 25% against Chinese products would have hammered Apple’s margins but it survived by gaining exemptions (this is in 2019)
By 2022, more than 70% of Airpods production, nearly a third of Apple Watch assembly and about a quarter of iPad manufacturing had moved out of China - mostly to Vietnam and Thailand. Apple has nudged the Taiwanese manufacturers to drive this shift who have driven the shift to India, Vietnam etc.
India ramped up iPhone production from zero to 15 million units between 2016 to 2023 (5.9 million in first half of ‘25). China between 2006 and 2013 went from zero to 153 million units! India mostly does FATP (Final Assembly Test and Pack out) mostly by Taiwanese companies Wistron and Foxconn.
Lower labor costs in India are offset by logistics cost of sending freight from China. Japan, Taiwan and China started out by making components, creating foundational technical expertise. By contrast India has been doing FATP for 7 years only now building up competency in parts
Industrial robots deployed in China - 290k. In India its 5400. Culturally too China suppliers and govt officials had a “Whatever it takes” approach to win iPhone orders, backed by docile hardworking labourers. In India its officials aren’t incentivised to boost growth and its labour has more of a voice. Tata Electronics however has shown intent and drive buying Pegatron and Wistron and scaling operations
Foxconn has limited interest in establishing operations in India given its investments in China and its knowhow of political and cultural landscape. Only thing driving it is Apple’s mandate. China has blocked equipment and workers from traveling to India but Foxconn has used its Taiwanese workers to compensate. China wants tech transfer to be a one-way gate
Today Apple Silicon in every iPhone, iPad, Macbook, desktop Mac, Airpods, Apple Watch are made on one small island by TSMC (Apple has doubled down on China risk moving away from Intel chips)
If Taiwan were prevented from exporting chips, global losses would be $600b-$1t on an annual basis for first several years
Samsung is likely to be the biggest winner if something were to happen to TSMC since it runs its own chip foundries in Korea and Texas
TSMC’s investment in Phoenix, Arizona would still be depended on Taiwan for advanced packaging so this fab would be a paperweight if something were to happen to Taiwan
Just 4 years after Trump sanctions, Huawei began shipping phones with its own chipsets and its own OS (HarmonyOS) and in early 2024, overtook iOS in Chinese market share. Huawei and Xiaomi outsold Apple in 2024
Apple market value has grown by $700m/day since Cook took over - but it might very well end being like Jack Welch at GE since the risks Apple faces of a blowup are now substantial

If it isn’t clear, China isn’t doing low value add work as is perceived by most of us. It appears as low value add only because they are foregoing margins. Most of the contributions aren’t inside the iPhone (as parts) but in the machinery and processes that made it. China doesn’t see itself as owing Apple anything - Apple created jobs but it also made a lot of money. It is also clear that Apple so far has been nimble and switches suppliers on a dime - be it Samsung or Inventec or Quanta when its not happy. However it might have sunk itself irretrievably in the Chinese Red supply chain and also traded short-term gains for long-term moats and stability (relentless margin expansion by squeezing suppliers in exchange for supplying know-how) both of which are at threat in the future. The book is very well researched and written well and is a very good read. 10/10

Meetkatrodiya · August 27, 2025, 7:36pm

@phreakv6 very good summuary. Trying to write something- credits to Suyash Bhave.

Suzuki and Apple – Does History Rhyme?
At first glance, drawing parallels between an auto company of the 1980s and a modern-day technology giant might seem odd.

Suzuki’s entry into India:

In the early 1980s, the Indian auto market was a closed, monopolistic space with just Ambassador and Premier Padmini dominating. Financial and operational efficiency metrics — working capital cycles, margins, return ratios — were hardly in focus.
After Sanjay Gandhi’s failed attempt to create a viable car company, the Indira Gandhi government facilitated a JV between Maruti Udyog Ltd and Suzuki of Japan in 1982. Interestingly, Suzuki itself was still a relatively minor player in Japan back then.
What Suzuki brought in was not just cars, but processes: Total Quality Management (TQM), Just-in-Time (JIT), and Kaizen. These methods revolutionized efficiency, discipline, and quality — permanently reshaping the Indian auto industry.

Completely Knocked Down (CKD) / Semi Knocked Down (SKD) to 95% Indigenization:
When Maruti Suzuki began operations in 1983, local value addition was barely ~3%. By 1991 it had risen to ~65%, and today ~95% of the car is indigenized. This gradual transformation shows how a nascent industry, initially dependent on imports, can build deep local capabilities over time. A similar trajectory could play out for India’s EMS/contract manufacturing sector, where value addition is still limited today.

Government Push – Then and Now
In the 1980s, the government rolled out the Phased Manufacturing Programme (PMP). While it didn’t offer subsidies like today’s PLI scheme, PMP created a framework for rapid indigenization:

Imports were allowed in the early phase, but with strict, time-bound local sourcing targets.
Penalties existed for missing indigenization milestones.
Forex regulations were eased, and access to the domestic market was provided as an incentive.

Suzuki’s Handholding
Suzuki didn’t just sell cars; it built an ecosystem:

Long-term purchase agreements gave revenue visibility to local vendors.
Japanese suppliers were encouraged to enter India through JVs with local players.
Suzuki engineers trained Indian engineers, raising process and quality standards.
Japanese banks were nudged to provide financing, while Indian DFIs like ICICI and IDBI, along with IFC, were mobilized to fund ancillary growth.

The Keiretsu Model in Action:
Maruti Suzuki’s success wasn’t just about assembling cars — it was about building an entire supply chain ecosystem. The Keiretsu model — a network of interlinked companies — was pivotal to this. Leading Japanese auto suppliers followed Suzuki’s lead and entered the Indian market:

Denso (alternators and starter motors)
Sumitomo and Motherson Sumi (wiring harnesses)
Asahi Glass, Krishna Maruti (seats)
Sono Koyo Steering, Kansai Nerolac
Machino Plastics, Jay Bharat Maruti
Nippon Thermostat (India), Lumax

This ecosystem-building approach led to a transformational shift in India’s automotive landscape. It opened the door to Indo-Japanese collaboration, which later expanded to broader Indo-Asian partnerships. Over time, other major collaborations followed suit, including:

Hero Honda, Kinetic Honda, Ind Suzuki (TVS)
Bajaj Kawasaki, Mahindra Renault
Jindal MG, and many more.

This model, initiated by Maruti Suzuki, redefined India’s auto industry, fostering deep collaboration between Indian and global players, which continues to shape the market today.

Basic read

Parallels Between Suzuki’s Auto Journey and India’s EMS Opportunity:

Lack of Existing Domestic Ecosystem:
In the 1980s, Indian auto component manufacturers, like those supplying parts for the Premier Padmini, were operating with outdated systems and technologies. Today, India’s smartphone manufacturing and assembly capabilities are still in their nascent stages — a sector that has only begun to scale up post-2020. The lack of a deep domestic ecosystem mirrors the auto industry’s early days.
PMP-PLI Similarities:
Low Value Addition Initially:
Both industries began with minimal local value addition. Maruti’s early days saw very little indigenization, and similarly, India’s smartphone sector started with very limited local value added. However, both industries have the potential to scale this over time.
Efforts to Develop the Ecosystem:
Just as Maruti Suzuki worked to develop an entire supply chain ecosystem — with Japanese suppliers setting up operations in India — the EMS sector in India has a long road ahead in terms of building out a deep supplier base (Imo most important thing). The government’s incentives will also play a huge part.

Divergence – Suzuki vs Apple:

Company Size & Market Impact:
In the 1980s, Suzuki was a small player, with only 1–1.5% of the global passenger vehicle market share, primarily known for two-wheelers. Apple, on the other hand, is a global giant today, commanding 20% of global volume and 80% of global profit.
Vendor Hesitation & Established Demand:
Initially, there was hesitation from auto vendors to partner with Suzuki. Many were unsure about the number of orders Maruti would receive, with an expected 40,000 orders — but Suzuki exceeded expectations with 1.25 lakh orders. In contrast, Apple’s demand in India is already well-established, with no such vendor reluctance.
Local vs Global Story:
The Maruti Suzuki JV was primarily an India-focused story, with Suzuki trying to build a presence in India, which was a nascent market. On the other hand, Apple’s presence in India is a global story.
Political Support – Then vs Now:
The Maruti Suzuki JV had strong domestic support, with no major international opposition, while Apple’s India operations face active resistance from both China (due to supply chain shifts) and the US (due to strategic, geopolitical concerns, make in US move). Unlike in Suzuki’s time, Apple faces a much more complex and contested environment.
Focus of Government Schemes:
The PMP in the 1980s was mainly inward-looking, aimed at domestic consumption, whereas the current PLI scheme is outward-looking, encouraging exports and helping India become a global manufacturing hub.
Strategic Vision – Now vs Then:
As against the PMO family’s pet project (Maruti Udyog Ltd – Mr. Sanjay Gandhi), today, it’s a well thought out plan (Make-in-India, China+1, Aatmanirbhar Bharat, etc).
Competitive vs Monopolistic Landscape:
Today’s landscape is more competitive, with room for multiple players, not just one.
Critical Stakes for Both Parties:
For Apple, reducing dependence on China and securing supply chains is critical. For India, fostering indigenization, creating jobs, and scaling manufacturing are vital. In contrast, back in the day, Suzuki’s success didn’t hinge on such urgent, multifaceted factors.

The Success of PLI in Mobile Manufacturing: A Transformation from FY15 to FY25

FY15:

Domestic Mobile Phone Production: Approximately 6 crores units.
Import Dependence: Around 74% of phones sold in India were imported.
Exports: Total export value stood at ~₹1,500 Crores.

Fast Forward to FY25:

Domestic Production Dominance: Over 99% of the phones sold in India are now manufactured locally.
Exports: massive leap to ₹2 Lakh Crores in total exports. Key Exporters Companies like Foxconn, Tata Electronics, and Pegatron have significantly contributed to this growth.
Value Addition: The industry now adds 18-20% value to the mobile manufacturing process.

Approved PLI Companies for mobile phones & specified electronic components:

Value Addition in Mobile Manufacturing:

China’s Early Struggles & Evolution:
Initially, China’s value addition to mobile phones was under 4%, a point that was even mocked in a Wall Street Journal article. However, as Apple consolidated its manufacturing base in China to avoid managing multiple jurisdictions, the value added grew significantly, particularly in areas like displays, camera modules, PCBs, and enclosures. By now, Chinese value addition is estimated at around 50-55%, with 45% of the Bill of Materials (BOM) for an iPhone coming from semiconductor chips sourced from Taiwan.
India’s Progress & Government Support:
In March 2025, the Government of India launched the Electronic Component Manufacturing Scheme (ECMS), with a substantial budget of ₹23,000 Crores, focusing on the non-semiconductor BOM (including components like displays, batteries, etc.). The scheme comes with flexible incentives (not just linked to turnover/production) to incentivize deeper value creation. Additionally, a separate policy for semiconductors is being developed to boost India’s role in the semiconductor value chain.
India’s Design Capability:
A key advantage India holds is its design prowess. For instance, the Google Pixel’s Tensor chip design team is based in Bangalore, underlining India’s capability in chip design. While China still leads in the manufacturing side, India is closing the gap on design, with India’s share around 20%, compared to China’s 25-30%. This brings India much closer to China in terms of design capability, an area where India has significant growth potential.

China Can Only Slow India Down – Not Stop It:
India has historically overcome strong foreign interference in its bid to keep
acquiring / indigenizing critical technology. E.g., despite setbacks in core national
security and strategic programs like nuclear power & weapons (Homi Bhabha
mysterious place crash), and space technology (multiple ISRO scientists dying under mysterious circumstances), India has still managed to develop a robust nuclear arsenal and is today a global cost leader in satellite launch technology.

Geopolitical Pressures Accelerating India’s Growth:
It could be argued that President Xi Jinping’s aggressive actions (against Apple, India, or the US by restricting exports of critical materials and equipment in response to tariffs) (China targets India's manufacturing sector, delays machinery delivery, pulls iPhone engineers - Times of India) and President Trump’s erratic tariff policies (which threatened global supply chain stability) have actually pushed Indian policymakers to improve the ease of doing business and commercial environment in India. These external pressures have accelerated Apple’s shift towards India. Apple resists Trump’s warning, says no pause on India's iPhone expansion - The Economic Times

Timing of Diwali – An Unusual Challenge:
An unusual challenge that might arise is the timing of Diwali. In China, the major production season typically starts in H2 (second half) of the year and runs until January to support the Western festival shopping seasons like Black Friday and Christmas. This is followed by the Chinese New Year holiday from January to March.

In contrast, India’s festive season, which includes several holidays, primarily falls in the peak H2 production period.

Meetkatrodiya · August 27, 2025, 8:13pm

Beijing has allowed Apple to exploit its workers, so that China can in turn exploit Apple. – Patrick McGee
Apple wouldn’t be Apple today without China. China wouldn’t be China today without Apple. – Patrick McGee

Rise of Foxconn:
Apple’s journey with contract manufacturers (CMS) began with SCI and LG. However, after LG tried to squeeze Apple with tough negotiations, Apple shifted to Foxconn. A key reason for Foxconn’s victory in securing the CMS contract was its extreme frugality (e.g., no ACs in offices but setting up ACs in production areas), hunger for business, and the founders’ risk-taking mentality.

At that time, Apple was much smaller but very focused on complex technology and had a zero tolerance for imperfections. Foxconn realized that if it could meet Apple’s demanding standards, it could serve anyone. Despite Apple being a low-margin customer (due to its automation-unfriendly designs and its ability to squeeze margins and working capital), the real value lay in the learning. Foxconn gained world-class manufacturing capabilities through this relationship.

Foxconn’s Performance:

1999: Revenue of $1.8 billion (smaller than competitors like Solectron, SCI, and Flextronics).
2010: Revenue surged to $98 billion (~50x in 11 years), making it larger than the top 5 competitors combined.

However, despite this massive revenue growth, Foxconn’s margins were under pressure:

2000: 10.6% margin
2007: 4.6% margin
2011: 2.4% margin

Apple outsourcing is not really outsourcing. They will have a control in almost every stage of the manufacturing process. They have extremely strict quality control as well. They will have full details of the entire cost structure of the contract manufacturer. All critical resources such as machinery, raw materials (RM), inventory, and employee payroll will be under the contract manufacturer’s (CM’s) control, with Apple having full, unrestricted access to these assets.

Conclusion: With Apple’s entry into India, many companies could stand to benefit. While the rise of Foxconn is exciting, one key lesson I’ve taken away is the importance of margins. This may not be applicable to all players, as Apple is also pursuing a strategy of vertical integration. As Apple seeks greater control over its supply chain, the beneficiaries of this shift will be those players who align with Apple’s objectives - particularly those providing critical resources, machinery, raw materials, inventory, and skilled labor, all of which Apple will have unrestricted access to.

Rishabh_Nair · August 28, 2025, 2:23pm

Just had to be a major U2 fan to know all about this :)
or for that matter the Beatles (they all have Irish heritage)

draveshia · September 17, 2025, 6:53am

Hi P
Awesome reading. I hav been trying to write notes , for the books i read, after reading your posts. Find it very difficult while reading. Do you dictate into an app or write. Also how do you find books and topics to read

abhishek_sinha · October 9, 2025, 8:22am

Thinking in Bets: Why Life (and Investing) is More Poker, Less Chess

Book reading streak has improved drastically in the recent past - putting out summaries / sharing has not :)

I finished reading Annie Duke’s “Thinking in Bets” and it’s one of those rare books that changes how you think about decisions, especially in investing.

The core concept: We’re terrible at separating decision quality from outcomes. e.g. Stock went up 5x? We were geniuses. Stock tanked 40%? Bad luck or unforeseen events.

Annie Duke (professional poker player) challenges this line of thought with a comprehensive approach. Life is poker, not chess. You can make all the right moves and still lose the hand. You can make terrible decisions and get lucky.

The “wanna bet?” framework is powerful. When you force yourself to think “what’s the probability I’m right here?”, you move from certainty to humility. From “I’m right” to “I’m 70% confident.”

The book also talks about building a truth-seeking pod: people who challenge your thinking, not confirm it. Echo chambers kill returns.

Not a long book, very readable. If you invest seriously, worth your time.

Main takeaway: Focus on improving your decision-making process, not judging yourself by outcomes. Good process + probabilistic thinking = better odds over time.

The Book reduced to equations
This is purely for easy retention of key concepts - please don’t take these as rules, do read the book :)

Outcome ≠ Decision Quality
Good decisions can have bad outcomes; bad decisions can have good outcomes.
Decision Quality = (Expected Value × Probability) - (Ego × Certainty)
Your best thinking happens when you maximize EV while minimizing ego-driven overconfidence.
Life = Poker ≠ Chess
Hidden information + luck makes life probabilistic, not deterministic.
Truth-Seeking Group > Echo Chamber
Surround yourself with people who challenge your thinking, not confirm it.
Regret = |Actual Outcome - Imagined Outcome| × Hindsight Bias
We torture ourselves by comparing reality to what didn’t happen but could have happened.
Hindsight Bias × Motivated Reasoning = Self Deception
We remember our past predictions as more accurate than they were.
Bet Size = Edge × Confidence × (1 - Ruin Risk)
Even with an edge, sizing matters: over-bet and you risk catastrophic loss.

rupaniamit · October 9, 2025, 7:40pm

Gelain Antoine, a renowned industry expert, brings decades of insight to his comprehensive book, a culmination of articles spanning from the early 2000s to the pre-COVID era. This authoritative guide offers a nuanced understanding of the industry’s complex landscape, including:

Industry Insights: Supply chain dynamics and expert analysis
Global Shift: China’s emergence as a key player in the traditional duopoly market
Innovation Strategies: Models and approaches driving industry growth

Through his extensive experience, Antoine provides valuable perspectives on navigating the industry’s intricacies, making this book an essential resource for professionals and enthusiasts alike.

Notes:

• I believe the aerospace industry is entering a new age, an age “defined not so much by technology prowess, corporate wealth accumulation or relentless growth, as has been the case for the last 30 years, but rather by restraint, humility and resilience”

• The US A&D (Aerospace & Defense) industry has been going through reconfiguration process from late 1990s to early 2000s where the prime suppliers spun off more than $15 billion worth of business portfolio (mainly upstream activities such as components and minor modules and sub-systems). PE players came in and played the role of catalyst for the emergence of a new breed of independent suppliers, positioned between primes and traditional subcontractors.

• Defense and commercial aerospace are counter-cyclical and very unlikely that both will be in a down cycle at the same time.

• Commerical aircraft’s innovation discussion circles around engines, power systems, connectivity, and big data. While these are important, stakes are much bigger in the aerostructures, materials and detailed parts, which concern much bigger pool of players and are the largest contributor to an aircraft’s cost.

• Excluding powerplants , the airframe and all its constituent parts account for two-thirds of a commercial aircraft’s cost. Aero structures are not typically thought of the next aerospace battleground because innovations have so far been slow and relatively few and far between. The most significant has been the introduction of composite materials, which have taken an increasing share of the flyaway weight, reaching 50% in the Boeing B787 and Airbus A350. Yet it took 30 years for the use of CFRP (carbon fiber reinformed polymer) to move from minor subassemblies, such as fins and rudders, to the full fuselage section. And composites still account for only 5% of the total aircraft raw material demand which has so far limited the destructive impact.

• But it looks increasingly likely that aero structures will determine The winners and losers in the next development phase of the commercial aerospace industry. Technologically, ongoing advances in both metal alloys and thermoplastics combined with rapidly maturing manufacturing techniques such as 3D printing, robotics, out-of-autoclave processing and fusion welding are generating a step change in possibilities for weight and cost savings, as well as lead time on airframe structure and parts.

• It is likely that many incumbent suppliers probably will end up “stuck in the middle”, with neither the depth nor the breadth of expertise required to remain competitive, due to the requirement of making big investment in the new technologies and capabilities.

• Airbus and Boeing are investing huge amounts in optimizing and transforming their aerostructure capabilities and supply chain.

• Aerostructures are the only area where OEMs can realistically integrate suppliers from strategic countries such as China, India and Russia into their supplyvhain. These countries have significant domestic capabilities in materials and aerostructures as well as huge resources and ambitions to beef them up, making their integration in OEM 's global supply chain highly probable.

• Both Boeing and Airbus reached their peak in 2018, collectively and evenly delivering sixteen hundred aircrafts and generating $120bn revenue. At that point they were both anticipating glorious days ahead, predicting a twenty-year demand of 40, 000 new aircrafts worth $6 trillion, that they would happily share between them.

• There is now little doubt that China is well on its way to succeeding where several other countries have failed in becoming a full -fledged player in the large commercial aircraft manufacturing sector. After getting their hands on both ends of the value chain, aircraft design and final assembly, the Chinese now understand that what will make or break their industry over the long term is what happens in the middle of the value chain.

• While British Aerospace was historically the “head”, Rolls-Royce has always been the “heart” of the country’s aerospace sector.

• Like the country as a whole The British aerospace industry is at a crossroads. In search of its true identity torn between the pragmatic recognition that its future is in extricable tied to the future of Europe as a whole and the deep -rooted belief that is destined for greater achievements on the world stage, for which its European membership may be perceived as a drawback.

• Belonging to a cluster should be an essential part of any aerospace player’s strategy be it around an OEM or a large tier 1 or smaller supplier. Not only will a cluster allow companies to aim higher and further, but It will also help them strengthen their home bases and reinforce their uniqueness, their competitiveness. Wings and roots are the two things one needs to prosper. It is true for a child; it is also true for an aerospace company.

• When it comes to major international defense contracts, politics often Trump economics, which is why, for example, countries like South Korea and Japan have always bought American when it came to strategic assets such as fighter jets and anti-missile defense, often in spite of the economics of the deal.

• The takeaway for companies involved in exporting sensitive products is they should consider politics as not just an externality over which they have no control but as important input into the contract negotiation as well as their business strategy as a whole.

• As with any close system though, true rigidity and bureaucracy have crept in and the whole system has become largely inefficient. That is how you end up with a price tag of $200 million for a fighter aircraft, satellite or launcher.

• A&D industry players should therefore not tie their future to the ability of institutional customers to fund their R&D. Instead, they should take the initiative and approach their innovation agenda more like startups do in the Silicon Valley, by being long on vision and talent and short on time and budget. If they manage to do that, then commercially driven innovation will become an opportunity rather than a threat.

• In the space sector, downstream services are by far where the stakes at the highest. They account for 85 % of the overall $ 100 billion plus annual space market value, vs 10 % for satellites and only 5 % for launchers.

Its a must read book to understand how industry has evolved over so many decades. Recommended.

abhishek_sinha · October 10, 2025, 1:44pm

Thanks for sharing this @rupaniamit . Looks like an interesting book. Those interested in learning more about Rolls Royce may find this interesting - https://joincolossus.com/episode/forster-rolls-royce-turbines-and-tribulations/

karanthharsha999 · October 26, 2025, 5:47pm

I have been fascinated by the way technology has changed around us over the years. The recent advancements of AI and the way it has been presented to us is more compelling and confusing for people like me. I have always thought that the best prompt comes from people like Shashi Tharoor. This became challenging to me and I wanted to know more about AI and prompts. Hence the reading of the above book.
Much of the topics are known to many or new to a few. Here is the summary of the book:-
CHAPTER 1
The rapid advancements in AI and ML have created new skills and knowledge. Prompt engineering is a critical skill for anyone working with AI models, as it directly impacts the quality and the reliability of the outputs generated.
Whether you are new or a seasoned AI practitioner, understanding prompt engineering will enhance your ability to harness the full potential of AI models.

HISTORY OF AI

Using machines to solve problems defines technological advancements in civilization. The better our machines, the better we use them thereby the better our problem solving. Computers are the electronic version of mechanical computing devices, which have existed a very long time, the middle ages saw clocks track astronomical events, what They did with motor and rotors was actually create ways to mathematically predict the motion of celestial bodies.

TURING AND THE CONCEPT OF THINKING MACHINE.

The notion of machines solving problems is not new. The roots of what we know of AI today goes back to World War II. A secret group of code breakers was formed at the government code and cypher school on the grounds of the grounds of the British country house Bletchley park in Buckinghamshire. One of the best in the team was ALAN TURING. Turing was a brilliant mathematician , helped to use electromechanical devices to break the code of German electromechanical enciphering device called enigma machines. To decode Enigma machines a group of primitive computer was made and named it colossus. It was the world’s first computer in terms of it being digital programmable platform.
To decode the encoded message, the recipient would need an enigma machine set up like the sender’s. They would type in the encoded message, and the machine would reverse the process to reveal the original text.
After world war II, Alan tuiring wrote about computer machinery and intelligence in an academic paper in 1950, where he proposed the concept of machine intelligence, which is now known as Turing test.
His concept of machine learning hinges on the ability of a machine to exhibit behavior indistinguishable from that of a human, particularly in the form of linguistic and conversational tasks.
this focus on behavior and performance marks the foundational shift in thinking about AI.
Turing is considered the father of computer science.

It was not just computer scientists involved in taking the concept of AI forward. AI had to be interdisciplinary field of study. Professionals, from mathematics, psychology, engineering, and computer science needed to come together. The notion of ML, natural language processing (NLP), and symbolic reasoning were evolving together.

FUNDAMENTAL CONCEPT OF MACHIINE LEARNING AND ARTIFICIAL INTELLIGENCE

AI is a broad field concerned with creating systems capable of performing tasks that typically require human intelligence. These tasks include problem solving, understanding, natural language, recognizing patterns and making decisions. ML, is a subset of AI, focuses on specifically on the development of algorithms that enable computers to learn from and make predictions based on data.
Data is the life blood of these technologies, the quality and quantity remains the backbone of the performance and accuracy of the models. Data collection, cleaning and pre-processing are critical steps in any AI project. By understanding these fundamental concepts, users can better experience and appreciate the potential and limitations of AI and ML. they can make more informed decisions about their implementation and guide their organizations toward strategic, ethical and effective use of these powerful technologies

OVERVIEW OF MACHINE LEARNING AND STATISTICAL ANALYSIS
Data is what they use, statistical analysis on very large data sets has been a crucial use for computing for years. BIG DATA or BUSINESS INTELLIGENCE(BI) are business / technical methods to draw insights from the signals in the data. Computers can make mistakes, so the focus is on having good data and refined and tested calculations. This is where ML comes into play.

ML is a field of computer science that uses statistical techniques to allow computers to learn from data without explicitly programmed for specific tasks. It is about designing and implementing algorithms that enable machines to utilize data for improvement and make decisions with limited human intervention.
ML enables AI systems to adapt to new scenarios and perform tasks recognizing patterns and inferring rules from data. Ml and AI are closely related.Ml is effectively behind the scenes, and AI is where the statistical analysis occurs and inferences are made.ML enables computers to build models from sample data to make predictions or decisions rather than following strictly static progrp1redictions. This adaptive nature of ML brings life to AI. NLP (Natural Language Processing) helps in understanding and generating human language and in computer vision. These capabilities are not just theoretical; they are used in real world applications like virtual assistance, self driving cars, and personalized medicine, highlighting the trans formative power of ML in AI.

TYPES OF ARTIFICIAL INTELLIGENCE

NARROW AI
It is designed to perform specific tasks within a limited domain. These systems excel at their designated task / functions but lack the ability to transfer knowledge or adapt to new situations outside their programming.
Examples :- chess engines, voice assistance and recommendation systems.
GENERAL AI :- Also known as strong AI or Artificial General Intelligence (AGI). Refers to hypothetical AI systems that would match or exceed human level intelligence across a wide range of cognitive tasks. Such systems would be capable of reasoning, problem solving, learning and adapting to new situations much like humans do.

DEEP LEARNING:- It is a subset of ML that uses artificial neural networks inspired by the human brain.Deep learning has led to significant breakthroughs in ares like image and speech recognition, NLP, and game playing AI.

NLP :-It is a field of AI focused on enabling computers to understand, interpret, and generate human language. Applications of NLP include machine translation, sentiment analysis, chatbots and voice assistants.

EXPERT SYSTEMS:- They are a type of AI designed to mimic human decision making abilities in a specific domain. They are composed of two main domains, knowledge base and an inference engine. The knowledge base contains a collection of facts and rules about the domain, while the inference engine applies logical rules to the knowledge base to deduce new information or make decision. Medical diagnosis, financial forecasting, and technical support, where they make users informed decisions by providing expertise recommendations based on the data and the rules set into the system.
Despite their effectiveness in narrow domains, expert systems are typically limited by the quality of the knowledge they contain and the complexity of the rules they can process .

COGNITIVE COMPUTING :- Refers to advanced AI systems designed to simulate human though processes in a computerized model. These systems use technologies such as NLP,ML and reasoning to understand, interpret and respond to complex data in a way that mimics human cognition. The goal of cognitive computing is to enhance human decision making by providing insights and recommendations based on the systems ability to analyze vast amounts data, recognize patterns, and understand context. Examples healthcare, finance,and customer service.

ROBOTICS:- Branch of AI that involves the design , manufacturing and operation of robots- machines that can perform tasks autonomously or semi automatically. Robotics create systems capable of interacting with systems of the physical world. They are equipped with sensors to perceive the environment, processes to analyze data and make decisions.and actuators to carry pout physical actions. The goal of robotics is to develop machines that can perform complex tasks safely and effectively and efficiently often in environment that are challenging or hazardous for humans.

DATA TRAINING AND THIE IMPORTANCE IN ML.

The importance of data cannot be overstated, as it serves as the foundation upon which ML models are built and trained. Quality data enables models to recognize patterns, male informed decisions, and ultimately deliver valuable insights. Bad data means bad results.
The quality and the quantity of the data directly impacts the performance of AI and ML models.
In ML and AI, it is critical to understand the distinction between supervised and unsupervised learning types.

Supervised learning :- supervised learning is a type of ML where the model is trained on a labeled dataset. This means each input set is paired with correct output. Examples like classification, where the goal is to assign input data to predefined categories, along with regression, where the objective is to to predict continuous values.
Unsupervised learning deals with unlabeled data. Here the model is tasked with identifying patterns, structures or relationships within the data without prior knowledge of the correct output. Common applications of unsupervised learning include clustering, where the model identifies data points that significantly differ from the norm.
Semi-supervised learning offers the advantage of supervised learning where the model benefits the guidance of labeled examples- combined with the efficiency of unsupervised learning, which utilizes the vast amounts of available unlabeled data. This approach can lead to improved model performance and broader applicability of AI solutions in various business contexts.
Addictive to the choice of supervised versus unsupervised AI training is an advanced concept that deals with AI using ML to potentially make predictions. These learning techniques are called few shot learning and zero shot learning.

FEW SHOT LEARNING it refers to the ability of a model to learn and make accurate predictions after being exposed to only a few examples of each class. Traditional ML models typically require large amounts of labeled data to perform well. Few shot learning leverages advanced techniques to generalize from a limited number of examples, making it highly valuable in applications where data is scarce or rapidly changing. Example:- Medicine is broadly a field where statistical analysis is central to the successful treatment of patients.
ZERO SHOT LEARNING:- It takes a few shot learning further by enabling models to make accurate predictions for classes they have never seen before.
Example:- A zero shot learning model trained to recognize animals might use descriptive attributes like stripes and is large enough to correctly identify a new, unseen animal, such as zebra, based on those characteristics.

Modern AI broadly encompasses a wide range of technologies and methodologies aimed at creating systems capable of performing tasks that typically require human intelligence.
On the other hand General AI or Strong AI, is the concept of a machine with the ability to apply intelligence to any problem rather than just one specific problem, essentially mimicking the cognitive abilities of human. General AI can learn, understand and apply knowledge in different domains, making it a more flexible and adaptable form of intelligence. General AI remains mostly a theoretical concept and has not been achieved in practise.

The current wave of chat services like CHATGPT evolved out of of the work being done on LLM systems. LARGE LANGUAGE MODEL(LLM) is an advanced type of AI designed to understand, generate and manipulate human language. These models are typically built using deep learning techniques and are trained on vast amounts of text data to develop a broad understanding of language, including grammar, context and various nuances. Examples of LLM include OpenAI’s GPT4, and Google’s BERT. GPT models are based on transformer architecture, which uses mechanisms called attention and self attention to weigh the influence of different words in a sentence. The training models are specifically trained through inputting prompt ( question) and answer pairs. The more you enter a wider range of expertise or depth the better is your LLM model. There are a few input points, so you can only use ML, although anything complicated that relies on interpolation would need training.
The interesting thing is that you do not just have to train an LLM on facts; there are trading methods where you do not give it an answer or give it bounds to work within.

Next chapters when time permits.

Vivek_Shetty · December 8, 2025, 7:01am

A Sense of Self - Memory, The Brain And Who We Are
By Professor Veronica O’Keane

The way we make memory is a mystery science is still unraveling. A Sense of Self is a book about how the brain synthesizes a self through memory, essentially creating the artificial duality between oneself and the universe. Dr Veronica O’Keane, a Professor of Psychiatry and a Neuroscientist at the Trinity College in Dublin, handles this complex subject with the clarity of a scientist and empathy befitting a mental health physician.The unfortunate reality of medicine is we learn a lot about physiology through the study of pathology, and this is especially true in Neuroscience. Suffering patients pay a heavy price for the knowledge we gain, and Dr O’Keane treats their stories with sensitivity and respect.

Memory is not just passive storage of sensory information acquired from the environment, but also it’s processing, pruning and contextualising.
Stimuli from the environment are sensed by transducers in the body, called sense organs. These convert various forms of energy into electrical energy, which then fire particular areas of the brain. It is the firing of this circuit that essentially is the representation of that particular stimulus. This circuit includes cells of an area of the brain called hippocampus, which essentially ‘stores’ working memory.
This sensory information is also relayed to the amygdala, which is the way the body attaches an emotion to it. The amygdala is connected to the hypothalamus, and hence the autonomic nervous system (ANS). Amygdala orchestrates the feeling of an emotion by stimulating various parts of the ANS (includes things like a change in heart rate, rate of breathing, sweating, other secretions etc).
Another area of the brain, the insula, then ‘interprets’ this feeling (interoception) through feedback from the ANS. This is how an emotion is attached to the memory.
Cells in the hippocampus also assign a place and a time to the memory. This is what adds external context to the sensory information.
Longer term memory ‘moves’ to the cortex, thus freeing up the hippocampus to effectively hold short term memory.
In this way, inputs from the surroundings, given internal context through emotion, and an external context through spacetime, stitch together a biographical narrative that gives an individual a sense of self. This isn’t information set in stone either, it modifies and evolves with accumulation of experiences and with age.

Professor O’Keane draws from extensive medical literature as well as stories from her own clinical practice to simplify this complex yet fascinating subject. The book is a thoroughly enjoyable read. I would highly recommend it to anyone interested in the brain.

rupaniamit · December 15, 2025, 9:51pm

Introduction

Conservative Republicans perceive this debt as an immediate threat to our economy, a danger that is in our faces right now. Their solution is to espouse austerity, to cut federal expenditures to arrive at a balanced budget and avoid becoming like Greece. Liberal Democrats on the other hand, see the debt as a problem on a distant horizon. One to be addressed by increasing taxes, especially on the wealthiest elements in American society, to achieve a balanced budget. These differences in perception and proposed solutions aim at the same target: a balanced federal budget.
The fate of the Bowels- Simpson Commission’s report on fiscal responsibility and reform illustrates this point. It recommends both tax hikes and program cuts. Apparently for this reason, both conservative Republicans and liberal Democrats turned their backs on it. Although the report should have served as a springboard for negotiation and compromise, it was dead in the water upon publication. This book constitutes an attempt to get beyond the tax increase & program cuts alternatives. It is an effort to confront the debt issue from outside of the proverbial box.
The more the government borrows, the less is available for private sector investment. This is the notorious “squeeeze” effect. Increasing government debt squeezes out funds for corporate research and development, employee retraining and other purposes. Put briefly, the “squeeze” inhibits economic growth.
Every dollar spent on interest is a dollar less for other purposes. In other words, those elements of the federal budget deemed “discretionary” suffer. The mandatory elements of the budget Social Security, Medicare, Medicaid, and the interest on the debt must be provided for, but defense and national security, education, energy, infrastructure repair and development, climate change and other needs wind up with less.
$19 trillion is the gross national debt. Yet the gross debt that conceals a fact that it is composed of two elements: the public debt and intragovernmental holdings The public debt consists of bonds, bills, and notes that investors purchase at the current market interest rates. Those investments are secured not by congressional legislation but signed and sealed by contract with the U.S. Treasury and backed by the full faith and credit of the United States. In any case, according to the Bureau of the Fiscal Services on June 6, in 2016, the public debt amounted to more than $13.8 trillion.
The other element of the gross national debt “intragovernment holdings”, is what the government borrows from the various trust funds that it has created over the years. While there are a number of such funds from which Treasury borrows annually, the principal in-house creditors are the Social Security and Medicare trust funds. Especially the former. In other words, the term “intragovernment holdings” refers to the money that the U.S. Treasury borrows from other parts of the government at interest rates defined by law. These securities are not Marketable to the public and the borrowing and the interest payments are intra-government transactions only. Briefly put, the government borrows from itself and pays interest to itself on these loans. On June 6, 2016, the Bureau of the Fiscal Service reported that the intragovernment holdings portion of the gross national debt totaled more than $5.3 trillion. That amount combined with the public debt adds up to gross debt of more than $19 trillion.
It separates the two recognizing that the real debt challenge facing the United States is $13.8 trillion, not $19.2 trillion. About 74% of GDP not more than 100% of GDP.
Consensus between the two parties is nowhere in sight. For this reason it is useful and perhaps even comforting to recall that since gaining independence in 1776 the United States has been debt-free for only two years and ten months (from January 1835 to October 1837). We have lived with public debt for nearly all our national history.

Notes from Chapters

Balancing the budget, if not generating surplus revenue, should be one of the main targets of public policy. Whether short term or long term.
Economic growth is the key to keeping us on track to ongoing debt reduction. Without growth in GDP, tackling the debt problem will resemble Don Quixote Tilting with windmills.
In 2000 and 2001, the Treasury reported surpluses for the first time in a very long time. Suddenly many in government, business, and academia were predicting debt freedom by the second decade of the 21st century.
When Bush Left office in early 2009. The gross national debt had soared to $10 trillion or 68% of GDP. Under Obama administration because of bailouts and stimulus, the national debt grew to about 100 percent of GDP to get the US out of 2008 financial crisis.
The 2008 Great Recession caused Federal Revenues to fall. The millions who lost their jobs ceased paying income taxes. Instead of contributing to the national treasury, they filed for and collected unemployment insurance, which of course was their right. But depending on the unemployment insurance hardly constitutes affluence. It promotes anxiety about the future. All personal discretionary expenses were halted and delayed until they found a new job. Those who had jobs stopped/reduced their discretionary spend thinking that they would get pink slips in next round of job cuts.
Fear does not promote economic growth. Robust consumption does. Thus for this reason increasing the disposable income of the greatest number of families should be one of the great objects of public policy.
Vigorous economic growth, it will be recalled, is the key to deficit and ultimately debt reduction. Vigorous economic growth depends in large measure upon maximizing the disposable income of the greatest number of families.
CBO projects The US population will increase from 325 million at the beginning of 2015 to 394 million in 2040. Currently, those over the age of 65 constitute about 25% of the population. By 2025, they will account for approximately 33% and by 2040, about 39%. As they become an ever larger segment of the society, seniors must remain active participants in our consumer-driven economy in order to secure and sustain vibrant economic growth. This fact alone underscores the importance of Social Security to the well-being not only for senior citizens, but also for the entire nation. Put another way, maintaining and better yet increasing the disposable income of this rapidly growing element in American society is essential to national prosperity.
Almost all of the social security programs funding; about 96 percent in 2015 was derived from the payroll tax, which amounts to 12.5 percent on the earnings. The payroll taxes are credited to the Social Security Trust Fund, which is made up of two legal separate entities, Old Age Survivor’s Insurance (OASI) and Disability Insurance (DI). In 2014, combined OASDI revenue from the payroll tax amounted to $756 billion. At the same time, total outlay for benefits amounted to $848.5 billion. In other words, benefit payments exceeded tax revenues by roughly $58.8 billion. Still Social Security Trust Fund had $2.7895 billion in reserve. How could this be?
Social Security Fund has Sources of income other than tax it collects from the workers. Under legislation enacted in the 1970s, surpluses in government-sponsored trust funds must be loaned to the Treasury. These loans are “intra-government” holdings. And Treasury is supposed to pay interest on these loans provided by the government funds. If and whenever these Trust funds need cash to meet their current operating expenses, the law authorizes Treasury to redeem “at par value” the amount that will allow the trust funds to meet their obligation. This process covered the 2014 gap between social security revenues and benefit payments. In 2014 these holdings were well over $2 trillion and the interest paid to OASDI that year amounted to more than $98 billion or about 11% of total OASDI revenue.
Any perceived threat to Social Security solvency will engender anxiety and fear among retirees and those who are approaching retirement. The consequence is certain and therefore predictable. Those dependent upon and soon to be dependent upon OASDI benefits will cut back on spending. They’ll become increasingly marginal consumers, buying only what they consider utterly necessary. Economic growth will slow down. Layoffs will follow and the newly unemployed will quickly cut back on their spending. Recession will deepen.
Defeating fascism in Europe and the Japanese in Asia transformed the United States into a superpower for the first time in its history. Indeed, when World War II ended in 1945, US military power was unchallengeable.
Preserving and extending NATO, a Cold War relic has cost and continues to cost the United States vast amounts of money. Given our annual budgetary imbalances, this is a matter that merits serious review.
There are about 174 bases in NATO countries. Moreover the larger installations – the ”Little Americas” as Professor Wynne calls them – house the families of military personnel. Family residence on these bases has been standard policy for many years. Supporting wives, husbands and children on this basis up the cost of American taxpayer. These little Americas attempt to create suburban middle-class American communities. They provide schools, restaurants, movie theaters, athletic facilities, country clubs, and all elements of stereotypical life in the United States. Construction and maintenance of course elevates the cost.
By Professor Mahin’s count, the United States currently maintains 113 military installations in Japan. More than 50,000 American military personnel serve there. More than 15,000 Marines are on duty in Okinawa alone at the cost of between $150 and $225 million more than if they were stationed in the United States. The total annual cost of our presences in Japan is about a “billion dollars a year.” Sixty-four years have passed since the first treaty between The US and Japan. In the interim, neither North Korea China, nor the Soviet Union Before its demise in 1991, ever attacked Japan.
US has 83 military bases in South Korea today.
In fiscal year 2012, the detention center in Guantanamo Bay operation cost taxpayers $448 million. Estimated cost for Financial year 2013 and Financial year 2014, were respectively, $454 million and $443 million. In fact, from 2002 through 2011, operating the facility cost a staggering $3.8 billion dollars. This amount was on top of whatever The bill was to support the 5,800 personnel on duty there. their families and their Little America environment.

Conclusion:

This brief book has aimed at identifying possible methods of narrowing our annual federal budgetary deficits without increasing taxes or slashing safety net programs. Some of the ideas for reducing national debt are vigorous economic growth, avoidance of intra-government debt and interest on it, federal investment in the private economy, and paring back foreign policy obligations to accommodate more closely our revenue realities.

This book provides a decent introduction to the concept of U.S. national debt. It’s a good starting point for readers with little or no prior knowledge of the topic. However, it tends to overlook several inefficiencies within the system that would have provided a more comprehensive picture. I would rate 8/10.

vb_1lavya · December 18, 2025, 9:53pm

Amit Bhai, thanks for the notes. I haven’t read the book yet, but couple of questions.

Why is intergovernmental debt may not be considered real debt?. It sounds like a real liability since it is drawing from the future even though of another federal entity.
Which other points did you mean could have made the narrative more comprehensive?.

Thank you,
Varun