You would have a category expenditure described as Capital Expenditure. You would have to buy cooking ranges, mixers, ovens, utensils, cutlery and the like. This is a starting expense, and you expect longevity. You do not expect to buy another cooking range for a decade, if not more.

Then you have operating expenses. Under this, you would have rent, salaries, ingredients, marketing and the like. In this, rent and salary will go out each month. The property owner does not care if 1 customer came or 10,000; the entire rent gets paid. These are called fixed costs.

By comparison, ingredients are truly a variable cost; you don’t use them up unless you make a dish. Hopefully, you do not make a dish unless a customer shows up.

Now, let us say your fixed cost is 50000. And your variable cost per customer is 50. You are charging the customer 100. You are making 50 over, and above your variable cost on every sale, hence you need to sell 1000 units to arrive at what is known as operational breakeven. At 1000 customers, you can pay off your fixed cost of 50000.

The moment you exceed 1000, you start recovering Capital Expenditure. At some point, you reach full breakeven.

Now, let us consider AI businesses, especially those claiming to build foundational models and raising billions as if they were millions.

Their Capital Expenditure is the data centres that they require. Data Centre capacity is measured in terms of Megawatts. The frontline companies, such as OpenAI and Anthropic, are building 100 MW data centres. This requires 50+ acres of land to begin with, and then they require an array of Nvidia GPUs. Each GPU costs about $5000. A data centre could have close to 100,000 to 200,000 of these. Then you have the building and all the other paraphernalia needed. Typically, a data centre costs about $3-5 billion to set up.

A 100 MW data centre is capable of catering to 10 - 50 million users based on the load it is receiving. For instance, if I send Claude a query asking it to convert a report into an interactive website, this can take 20 minutes of operation on several GPUs. Hence, the ability to cater to a large number of users can come under stress if the requests require a high degree of sophistication.

Then come the operating expenses. Depending on the cost of power, operating a 100 MW data centre can cost between $50 million and $150 million per year. Most of the foundational AI companies require Gigawatt scale to cater to the demand that they are receiving. Hence, the 100s of billions being spent.

But here is the fun part. The life of a GPU is between 5 and 10 years, depending on the amount of use it is subjected to.

Have you ever bought a laptop and, after a decade of use, it starts to sprout all kinds of issues? It starts to freeze up, crash, and be generally very slow. That is the result of, for the lack of a better expression, “wear and tear”.

The AI companies are throttling use because of the demand that they are seeing. Safe to say, very high wear and tear. So the 100s of billions being spent will have to be spent all over again in 5 years!!

Imagine you are building a hotel and you have to demolish and build all over again after 5 years! No hotel would ever break even.

Have you heard of an AI company that has managed to break even?

In no world is a $4 billion monthly revenue run rate small. If, at that scale, you are unable to breakeven, there is something fundamentally wrong financially. OpenAI boasts 50 million paying subscribers, and Anthropic claims 30 million paying subscribers.

At what scale should operating income overtake operating expenses? This is not an ad subsidised business such as Google or Facebook.

Also, it is one thing to get revenue, and it’s another how you get it. Across companies there are leaderboards set up for engineers who use the highest numbers of tokens used by the engineers on AI. Bad incentives lead to bad results.

Axios reports that an unnamed corporation managed to “accidentally” torch roughly $500 million in a single month. The cause? The company introduced Anthropic’s Claude to its workforce but forgot to put caps on usage. Employees ran long, unchecked queries and complex agentic workflows, racking up an enormous tab before the finance department even realized the meter was running.

Source: Inc

This is lunatic.

If you paid $25,000 a month to an engineer, you could have hired 20,000 engineers for a month with $500 million dollars. That would be the number of software engineers that a company like Apple or Facebook has on their rolls. I hope they get some output from all that AI use, although I highly doubt it.

This kind of wanton spending is propping up revenues. This is sure to disappear sooner or later.

Ultimately, the problem is that all of these companies are starting with the stated goal of putting a lot of humans out of work. They hope to achieve this using regurgitation machines.

There is a simple problem they are all struggling with. The human brain runs on 20W of energy; to replicate that capability, these companies require several megawatts of energy. That asymmetry is proving hard to overcome and manifesting itself in the form of economics.

There was a 500 KB limit on the document size, while also having a very specific 2130 X 1346 resolution required for the image.

Why?

Then something occurred to me. I have yet to come across a government site that is not a form. This is true of every country where I have had to engage with a government website.

Every single government site is always a form.

In India, we regularly use government websites for Income tax, GST, Aadhaar, Driver’s License, Passport, and Voter’s ID. But even beyond this, birth/death certificate, stamp papers, duty payment and the list goes on.

Every single site is a glorified Form.

Someone figured out how to make a form online in 1992. Why are we still sucking at it?

None of these forms has any elaborate back-end computation. No AI. Not like you are uploading a picture and AI figures out the rest.

It is just a form which often does not even validate for data type. You can type text in the phone number field.

In India, all of the government sites are either designed by TCS or Infosys. Why can’t our IT behemoths deliver decent forms??

Let us consider for a moment what that will involve. The only evidence of ownership that the hotel would have is an invoice for the purchase of the chair. That purchase, unless the hotel is new or recently refurbished, could have been made years, perhaps even decades ago.

So, are you free to carry away the chair if the invoice is not presented?

Not all properties are equal, and land, if it was bought even centuries ago, would have some paper trail.

In India, some forests have been tended to by tribes for centuries. It is such a vast country that parts of it have been left untouched. I was recently talking to a colleague, and she mentioned that there are places where you do not feel the presence of the government. How do you feel the presence of the government?

We do not think about it?

Every piece of public infrastructure, including roads and bus stops are a result of work done by the government. Our cities look a certain way, and our buildings are built in a certain way because there are rules laid down by the government. If you are robbed, there is an authority that you can report it to; the police are put there by the government.

If something is broken, there is someone to blame - the government.

But in the tribal recesses, none of these exists. There are no roads, no building rules, no authority, and you are the custodian of everything around you. You take care of what you consider valuable, and you ignore the rest.

So when the relentless march of capitalism suddenly requires land to build a factory, and the government realises that there is some land lying untouched in some corner, they show up to take charge of it.

When the tribal claims ownership, they are asked to present evidence. Some of them have been living in those forests since even before this country existed. Who was supposed to issue this evidence?

Going back to the hotelier, you would not assume that the chair did not belong to him because he did not have a paper to prove it, but the same standards do not apply to the tribals. They need either to present evidence or vacate! That pretty much explains all tribal militancy in India.

On May 21, 1967, as was his custom, Bigul Kisan went to harvest the crop on the fields of a landlord in the Naxalbari district in northeastern West Bengal. This particular day was different, however. He was to take over the land as his own, per the instructions of the Krishak Sabha (Farmer’s Union), a peasant organisation of the Communist Party of India (Marxist), commonly known as the CPM. As Bigul Kisan would describe the incident twenty years later, “The goons of the landlord attacked me. They nearly crushed my head. They took away my plough and oxen.” That spark set off a wave of peasant attacks on landlords.

On May 24, when the police tried to defend the landlords, two arrows from peasant bows killed Constable Sonam Wangdi. All the village men then went into hiding, leaving the women to form a barricade against further police force. The police stepped up its offensive, shooting and killing nine women and two children. Armed rebels continued to attack landlords. The CPM, which had unexpectedly vaulted into power as part of a United Front coalition government in the state assembly election that February, was initially sympathetic to the rebels. But by mid-June the CPM assumed its governmental role of maintaining law and order. The Chinese jumped in to counter the CPM’s anti-rebel position.

On June 28, Radio Peking lauded the rebels as the “front paw of a revolutionary armed struggle,” adding that the peasant movement was inspired by Mao Zedong, chairman of the Chinese Communist Party.

Indian authorities began repressive measures against peasant rebels on July 12. Backed by colonial-era laws, which permitted arrests without charges and trial, Indian police launched a major operation to round up the rebels and their leaders. By the end of July, about 1,500 armed policemen had established themselves in Naxalbari. Under the overwhelming force of the police, Bengal’s armed peasant movement folded.

Source: India is Broken, A people betrayed

Since the armed rebellion started in Naxalbari, the people engaged in the rebellion were referred to in Hindi as Naxalwadi. In the English media came to be referred to as Naxalites or Maoists.

About a dozen states in India have been known for Maoist activities. Characterised as terrorists, these were tribals who were forced to turn political because they wanted ownership over lands that they lived in. It is difficult for tribals to fight a government that derives revenues from crores of Individuals and has the freedom to employ it how it deems right.

In a defining moment in India’s battle against its decades-long Maoist insurgency, CPI (Maoist) top commander Tippiri Tirupati alias Devji also known as Devu Ji has laid down arms and surrendered to Telangana police.

Sources in the state’s Special Intelligence Wing confirmed to DH that Devji surrendered alongside senior Naxal leader Malla Raji Reddy alias Sangram, Narasimha Reddy alias Ganganna, and 16 other cadres.

Source: Deccan Herald

This surrender is a victory of the state machinery over the tribals. It is the loss of hope of a people who preserved their lands. They have given up on the possibility of preserving their lands and given in to the state machinery that will use it for more manufacturing plants, presumably. This is not a story of glorious victory; it is the story of attrition and the victory of capital.

While several American billionaires and politicians were named in the files, Pan Bondi said, if we were to go after all the people who are named in the files, the entire system would collapse.

The reaction of the American people was ‘Meh’. Late-night hosts made fun of it. The women who came forward felt stupid. Life moved on.

Incidentally, Donald Trump was named 5300 times in the files and as someone usefully pointed out on Twitter, Harry Potter is named fewer times in all of the 7 books put together.

Across the Atlantic, there was far more enthusiasm to do something about the people named in the file. At one point, it even seemed like the British government would collapse not because the current Prime Minister, Kier Starmer, was named in the files, but because he had appointed Peter Mandelson as the Ambassador to the US, and his name figured prominently.

The former Norwegian PM Thorbjørn Jagland was rushed to the hospital since he attempted suicide because his name surfaced in the files and triggered corruption charges.

Thomas Pritzker, the Chairman of Hyatt Hotels, stepped down from his role.

Kathy Reummler, the Chief Legal Officer of Goldman Sachs, who had previously served as the White House Counsel during the Obama Administration, stepped down.

Peter Attia, the Logetivity influencer, had his CBS show cancelled.

Richard Axel, a Nobel laureate and Professor at Columbia University, said he would be stepping down from his position.

Larry Summers, the former president of Harvard University who served as treasury secretary during the Clinton administration and head of the National Economic Council in the Obama administration, said he would step away from teaching at Harvard.

The list is long, but the highlight is that Former Prince Andrew, brother of King Charles, was arrested in England.

The Harry Potter of this saga was beginning to feel the heat. Hence, the bombing began in Iran.

Israel and the US had bombed Iran in 2025. Iranians had launched missiles towards Qatar and then informed the US that retaliation was on the way. Damages were avoided, but so was contagion. Iran did not want an escalation, unlike Israel. Turned out, Iran was the country with greater statesmanship and forbearance.

Iran had strategically ensured that it had put its country right next to all the US bases, expecting a situation like this.

Source: Al Jazeera

Israel and the US were aware, if they attacked, all the countries around Iran with US bases would be attacked. With absolutely no concern for their bases but also for the people living in all of those smaller nations, an attack was launched with no clarity of purpose or an end goal. This has to be perhaps the dumbest CIA operation.

Without a legitimate replacement in the country, they assassinated the Ayatollah. A regime change is either effected by sending troops to decapitate the state completely, as they did in the case of Iraq or by supporting an actor who has the respect of the masses and the military, like they did in Bangladesh when Sheikh Hasina was ousted.

So now what?

The Epstein story is buried, and all of the attention has shifted to Iran.

With no clarity of purpose or goal, the bombing of Iran continues. The economies of the Gulf nations will suffer as this bombing endures. For one thing, three of the largest airlines in the region - Emirates, Etihad and Qatar Airways are not able to fly.

They have given a huge jet and invested heavily in Trump’s businesses. The pressure is going to be on Trump to end this soon.

The real question is, what is the exit ramp? After having engaged in this escapade, when they stop bombing, what will be justified as victory?

Semitic people or Semites is a term for an ethnic, cultural or racial group associated with people of the Middle East and the Horn of Africa, including Akkadians (Assyrians and Babylonians), Arabs, Arameans, Canaanites (Ammonites, Edomites, Israelites, Moabites, Phoenicians, and Philistines) and Habesha peoples

Source: Wikipedia

By this definition, anti-semitism seems to be a euphemism for Islamophobia. It is a twisted interpretation that implies anti-semitism is hatred of the Jews.

Since we are on the subject of hatred of the jews, the people who hate them the most are British. The current Jewish population in Britain is estimated at 0.5%. When Britain occupied Palestine during the Second World War, there was a restriction on the number of Jews who could migrate to Palestine. The British did not want anything to do with the jews. Probably why they started by trying to appease Hitler until the war came home. They created the state of Israel because they wanted to play no part in accommodating any jews after the war was over.

The European “semites” went on to occupy Palestine after the Second World War, and their demand for a Jewish state was a cornerstone of Zionism.

Theodor Herzl was the founder of the modern Zionist movement. In his 1896 pamphlet Der Judenstaat, he envisioned the founding of a future independent Jewish state during the 20th century.

Source: Wikipedia

It speaks volumes that even the document that envisioned the founding of Israel was not even written in a Semitic language. It was written in a language of the Zionists, a European language.

The first government of Israel was full of Europeans.

David Ben-Gurion - Born in Poland

Dov Yosef - Born in Canada

Zalman Shazar - Born in Belarus

Moshe Sharett - Born in Ukraine

Eliezer Kaplan - Born in Belarus

Haim-Moshe Shapira - Born in Belarus

Pinchas Rosen - Born in Germany

Gloda Meir - Born in Ukraine

Bechor-Shalom Sheetrit - Born in Tiberias

Yehuda Leib Maimon - Born in Moldova

David Remez - Born in Belarus

Yitzhak-Meir Levin - Born in Poland

There was a grand total of one person on the cabinet who was born in semitic land.

Paternal Country of Origin - Present-day Israel

Source: Wikipedia

Even today, two-thirds of the jews in Israel trace their origins to non-Semitic lands. Israel is a collection of a bunch of Eastern European people.

Given the genocide in Gaza. If at all, any country that can be deemed anti-semitic, it is Israel. The G7 are active participants in the anti-semitism perpetrated by Israel.

Referring to the people of Israel as Semitic is like saying ‘Trump is pro-European’, because all the white people in the US descend from Europeans.

I was travelling to the US towards the end of January and did not want to post anything fresh that would result in my deportation. At the time, the hottest thing in the news was the ICE raids in Minneapolis.

The other piece I had written was about the most anti-semite nation in the world - Israel.

I chose to keep my mouth shut.

My choice was not irrational either. I knew an American citizen, who just happened to be brown, who was stopped at immigration and had their phones and laptops checked by the officials.

Censor is a word derived from the Latin word Cēnsēre, which means to appraise or to judge. But when that appraisal is done in the context of political acceptability, one is already on a slippery slope.

During the Nuremberg Trials, the reasoning offered by the Nazis for their actions was - it was the right thing to do at the time. Joseph Goebbels ran propaganda for Hitler the way he did because otherwise, he would have been fired.

It was the “politically acceptable” thing to do.

Pam Bondi is defending the people named in the Epstein files the way she does because otherwise, she might be fired.

Across the world, the state reserves the monopoly on violence. They inflict that violence through the police. The system that controls that violence determines behaviour. When that control slips into the hands of one all objectivity is lost. Polical acceptance becomes a proxy for self-preservation.

Not every German was a Nazi, every German did what was politically acceptable.

Doing the politically acceptable thing results in self-censorship. That kind of censorship enables the few at the top.

It got me thinking.

A few days later, my maid, whose earnings would put her comfortably in the top 15% of Indians, informed me that she was taking leave because she had to go to her village and buy land - 20 Acres of it!

In 1950, just after independence, 75% of the workforce in India was engaged in farming. Agriculture was the biggest contributor both to employment and to the GDP. Over the years, it has been declining, and today, only about 45% of the workforce is employed in farming.

Having said that, India’s population in 1950 was 358 million; today it stands at 1.45 billion. Therefore, if you look at it in absolute terms, the number of farmers in India has gone up from 268 million to 652 million. This, in turn, implies that the land that was being tilled by fewer people is now being tilled by a larger number of people, and proportionately, the land holding per farmer has charted a downward trajectory.

85% of the farmers in India are smallholder farmers with less than 2 hectares (5 Acres) of land.

With such small farms, it is not possible to eke out a living for an entire family. This has created a flood of migrants who are moving away from their villages to the cities in search of work. They live in urban slums and perform menial and daily wage labour in cities to earn a living. That this income is higher than what they would earn at their farm should cause us to pause and think.

The very same people were left stranded and walked back to their villages when the lockdowns were put in place during the pandemic. Some of them were forced to walk hundreds and thousands of kilometres to get back home. Nevertheless, they came back.

Farming is becoming more and more unviable at smaller scales. Not only that, but it is also no longer aspirational amongst those living in villages. The youth in the village want to leave for the cities. Part of the reason there is a fertility collapse in South India is that most rural households wish to educate their children so that they can find respectable jobs in the cities away from the villages. Most cannot afford to put together the resources to educate even one child.

It is getting to a point where there are no ”youth” to be found in villages.

Then, there is another India.

Techies who work 16-hour days in glass prisons to earn obscene sums of money. They burn themselves out following these insane work routines.

While the evidence is anecdotal, I have started to see many such people who are burnt out and investing their money to buy a plot of land to start farming. They can buy a fairly large plot (>30 acres) of land for a sum which would get them a small three-bedroom apartment in the city. They are educated, well-to-do, aware and have the patience to learn organic agriculture. Not only that, they have the ability to connect with the market and get themselves the right price for their produce. They want a simple life which can also provide them with a means for a steady and stable income while producing something real. Something that connects them to land and people.

These individuals buy large parcels of land. 10 hectares or upwards and cultivate several crops, and vegetables that they either package and sell on their own or, in certain cases, even process the produce to deliver an end product.

Is this how India moves or transitions to large holder farms? Will this be the way we make farming a more viable engagement?

Are all the rural folks going to end up in cities, and city folks going to end up in the rural belt?

With CDs and cassettes, you would listen to one album a time, or if you had mixed it yourself, you would have a few albums at most.

By the 2010s, internet speeds had improved wildly, and thus came streaming. Suddenly, the entire music library of the work was available at your fingertips. This broke music even further, and we are moved to consuming music piecemeal. One track at a time.

Last week, I was working from home, and I decided to just revisit an album by Michael Learns to Rock and play the entire album. One thing was that I seemed to remember was which track followed just because I had listened to the entire album so many times. But more importantly, there was a flow to the album. There was a story that they were trying to tell with the album, and all the songs were connected.

The other being that every album has good tracks and the not-so-good ones, and when you listen to the entire album, you learn to appreciate both. Songs do not have to be breakout hits for them to be worth listening to. There are so many that were not hits but are really enjoyable.

The ability to enjoy creativity in its entirety is being lost.

Everyone across the world has heard Gangnam Style by Psy, but cannot name another song by him.

It leads to the degradation of creativity as well. Everyone is chasing that one hit. They are not extending any thought into their work. They just want a quick hit. One big hit and they can call it a career.

I had written about this almost a decade ago in February 2016.

Songs have been a means through which emotions and stories were communicated. What makes a song beautiful is the poetry and the metaphors, which make several interpretations possible. It automatically implies that you must give it time to grow on you, for the interpretations to set in and for the songs to get internalised.

With the lack of time and attention, musicians are forced to produce music that is catchy and will hook you up soon. This also means that most songs do not have a soul. Most songs use a particular word repeatedly to get you to like the song.

Source

Since I started this blog with albums, I want to know do any of you know the album that Gangnam Style was from?

Thanks to Trump, the year felt like a decade.

The year began with the impending ascension of Trump to the presidency. He entered office and put Elon Musk in power of DOG-E, and American government Workers faced uncertainties, the likes of which they had not since the Second World War. He promised to reverse all of the policies that Biden had put in place and declared “Drill baby drill”.

The Climate

This year we had wildfires in Russia, Turkey, Syria, South Korea, Albania, Bulgaria, Croatia, Cyprus, France, Germany, Greece, Montenegro, the Netherlands, Portugal, Spain, United Kingdom. In addition to that, 25 US states recorded wildfires burning 5 million acres and taking a lot of celebrity homes in LA. Canada remained the unrivalled champion with 160 fires burning through 21 million acres.

In addition to that, there was flooding in Milwaukee and Texas in the US. Queensland and New South Wales in Australia. Bolivia, Indonesia, Vietnam, Philippines, Indian, Pakistan, Sri Lanka, Nepal, United Kingdom, South Africa, Botswana, and Bolivia.

Then a bunch of world leaders flew into the Amazon to agree on doing nothing, for now, about the climate. They all have, at most, a decade or a bit more to live and that time would be better used creating “shareholder value” rather than taking decisions which will cause inconvenience and disruption to their “economic models”.

Shifting Sands

The tariff wars that were unleashed in April drove the panic purchase of more goods across the world. Markets gyrated at first and then just disregarded any talk of tariffs. The Financial Express wrote, “like a dog to a bone, Donald Trump seems to run back to Tariffs”. He alienated friends while foes rejoiced. Europe and Japan were made to accept one-sided deals, which they happily did. At the same time, Indian and China defied the US and did not accept any deal. These economies that like to call themselves the global North could not stand up to bullying!

2025 was an important year to unveil the new world order. Whatever the motivation Trump may have had, he exposed the weakness of the G7. The erstwhile colonial powers do not have any.

Killing Fields

To add insult to injury, the US has put the burden of bankrolling the Ukraine war on the Europeans. Trump was making changes to the immigration policy that would destroy the pipeline of talent that flows into the US using the H1B visa. For a brief moment, the Europeans thought they could roll out plans to attract that talent. Now that money will be spent buying arms to keep the war in Ukraine alive.

India attacked Pakistan to avenge the civilians killed in Pahalgam. For a period of time in India, it seemed like war. Cell phone networks were being jammed in India. A lot of “information” was being provided from the areas of conflict. Then both sides declared victory, and the war was over.

Israel, in the meantime, bombed half a dozen Middle Eastern countries. Israel attacked Iran unnecessarily. Iran turned out to be far more restrained and calculated than it is given credit for. They launched a bunch of missiles towards the US base in Qatar to demonstrate what contagion in that region could look like. The US made Israel immediately back down.

AI

In 2022, ChatGPT was released, and it captivated the imagination of many. Riding on the back of the “imagination”, OpenAI pushed for incredible amounts of capital to be committed towards the build-out of infrastructure to support AI expansion. The year was dominated by trillion-dollar investment talk and trillion-dollar scale, while the reality stayed far away.

Oracle leveraged its Balance Sheet to build capacity for OpenAI. Nvidia started investing in OpenAI. Even the house of the mouse, Disney, committed capital to OpenAI. All the money is being spent on creating data centres and building power capacity to meet the needs of AI. On paper, yes, they are building out infrastructure.

The wrinkle in the story - chips do not last as long as other infrastructure does. Spending trillions on building roads, rail, ports, or airports is fine. It gives you returns over decades. Chips are finished in 5 - 7 years. Imagine building an entire road network all over again every 7 years!

Is AI economically viable?

As the year draws to a close, an uneasy question about a small number of companies making circular payments to each other is beginning to rear its head. In the meantime, China is releasing models of comparative capabilities working at a fraction of the cost. America’s decision to ban Nvidia from shipping chips to China forced them to adapt.

The rise of China during the early part of this century was powered by cheap steel that it dumped into global markets. It seems like the US ensured that it would do the equivalent by dumping cheap AI models into the market and destroying the US AI industry.

Looking forward to 2026 -

The AI thesis will really be put to the test, and there will be a crash or a correction depending on who writes the headlines.

European governmental budgets will buckle under the load of defence spending on behalf of Ukraine. Will they call Zelensky corrupt and cut him loose, handing over territory to Russia? Will they risk fomenting unrest in their countries to continue a war for purely ideological purposes?

The Insurance industry will come under pressure both from Climate and Trump policies in both health and trade. Already, an insurance industry CEO was shot. Rising premiums will result in hard questions.

Enter Sam Altman.

Perhaps because of his name, he wanted to create the alternate man, which he dubbed Artificial General Intelligence.

Artificial general intelligence (AGI)—sometimes called human‑level intelligence AI—is a type of artificial intelligence that would match or surpass human capabilities across virtually all cognitive tasks.

Source: Wikipedia

Most code is written to operate in a specific area. Photoshop can do everything relating to images, but it can’t even work on videos, let alone write prose. Software packages have been great at doing specific pieces of work, but there is no software package that can do it all. By contrast, a human can learn to code and manage HR. A human can work across contexts and understand the specific requirements of that area and excel at it.

The moment GPTs were able to regurgitate words across a variety of areas, Sam Altman became obsessed with the idea of creating AI that could work like a human. Buoyant from the hype he was receiving after the release of ChatGPT, he claimed AGI would be possible by 2025. Much like Elon Musk, that date has become a moving target.

Why AGI?

In his mind, that was the best way to usurp the incomes of many people and corner that money for himself.

His focus turned from pushing the boundaries of research to turning OpenAI into a for-profit entity. He estranged and got rid of all of the research minds that produced everything at the company. Ilya Sutskever had a falling out after trying to push Sam Altman out of the company and left thereafter to start his own venture. Mira Murati and John Shulman left to start Thinking Machines Lab. Andrej Karpathy, who at Stanford played a critical role alongside Fei Fei Li to usher in visual recognition, left to start his own venture.

The vision has broadened while the bench strength has narrowed.

When you set out to please everybody, you please nobody. The pursuit of AGI has delivered precisely that.

I used Gemini last week, and it smokes OpenAI. Google is really flexing it with their search and Docs integration. Anthropic has become the gold standard for coding. OpenAI, in the meantime, has turned into a bullet machine.

Ask anything, and it will deliver 35 bullet points, which has given birth to a new term, AI Slop.

Why the cancellation?

A new directive from the Directorate General of Civil Aviation (DGCA) came into effect regarding Flight Duty Time Limitations (FDTL), which directed airlines to provide rest to flight crew, specifically -

Further,

The minimum in-flight rest period shall be a consecutive 90 minutes period for each crew member and 2 consecutive hours for the flight crew members at control during landing.

Source: DGCA

There were clear limitations placed on how long the pilot and the crew could be asked to work, and further, the duration of rest that needs to be provided thereafter. The document is pretty coherent and well thought out; you can check the link in the source quoted above.

All the people working inside a plane have to be ready to jump into action in the event of any mishap. Their alertness is subject to their mind and bodies being well-rested. The revised document was published on the 8th January 2024.

The leadership did not care what the circular said and assumed that everything would be fine. Worst-case, they would not comply, pay a fine and move on.

In some ways, they have managed not to even pay the fine by manufacturing this chaos! *Slow-Clap*

Indigo the Efficiency Machine

Indigo has been one of the consistently profitable airlines in the Indian aviation industry. It has made money by running its operations like a machine.

It achieved this by doing a number of things. Till a few years ago, it ran a fleet that was exclusively Airbus 320s, which enabled standardisation as well as speedy solutions to any maintenance problems. The ground staff did not have to deal with different types of planes. They also flew the Airbus like a bus.

The plane lands, it is turned around in 45 minutes, passengers start boarding, and it takes off. The plane rarely spends more than 75 minutes on the ground. I would argue that most buses get more downtime. This was a model that was perfected into an art by Southwest Airlines in the US, which allowed them to rise incredibly fast.

On Flightradar, you can see how any aircraft is run around the country.

This equipment (VT-NCX), for instance, is expected to fly 7 flights on the 6th Dec.

Every company aspires to achieve 100% utilisation since a lower utilisation is seen as inefficiency.

Poor people do not have any slack in their finances; this is what keeps them poor. Their utilisation rate of money earned borders on 100%.

When your operations are aspiring for the highest levels of utilisation, a slight deviation from that can cause the entire operation to fall apart. We saw this happen during COVID when shipping lanes went into a tizzy.

But when that operation involves people, even those people have to run like machines. Here enters the FDTL.

In the case of Indigo, that sword was utilisation.

The point to ponder here is that Indigo’s consistent profits were a result of staff who worked 12-hour shifts without fail. The government asked the airline to provide them with slightly more rest, and the company is falling apart.

We rarely acknowledge that what every company calls profits is essentially the accumulation of all the negative externalities that they dish out to the world. In this case, the externalities were being borne by their flight staff.

It seems like the profits of every company are their net debt to society.

It was in this context that I was reminiscing about the winters in Delhi during my childhood. Not the toxic gas chambers of today, but Delhi in the 80s. Delhi would reach 2 degrees Celsius and would see thick fog, not smog. The fog used to be so dense that taking a walk outside could leave your clothes with a thin layer of moisture on them.

If you grew up in North India during the 80s, there were some stupid things that were a part of life. I sometimes imagine why we made it so hard on ourselves.

(Not) Woollen Gloves - Jobless grandmothers could be seen sitting in the sun in the afternoon, knitting all over town. They used to knit useful things like sweaters, but also really useless things like gloves. Those nearly porous things never ever kept my hands warm. Their only value addition being the loss of dexterity. But all kids would wear it, hoping against hope that it would keep their fingers from freezing over. The fact that not a single adult wore one should have been a tell.

The great travesty is that it was not even wool. It was made of acrylic pretending to be wool. There was this company called Vardhaman, which used to sell knitting yarn with not one whit of wool in it.

Razai - A Persian word that came to India and was adopted into Urdu. This is the Indian version of the quilt. It’s essentially 25 kgs of cotton inside a stitched-up cloth bag. The razai used to weigh a ton and was always deadly cold to get into. You had to expend a lot of body heat to warm it up, and it would feel like your feet were perennially cold. Not to mention, folding it was a 3 people job, and it would occupy half of the bed once folded and look like one shapeless mound.

Upside, it was great to jump on.

Layers - The solution to any clothing being ineffective at defying the cold was to wear more of the same. You would wear a full sweater on top of a half sweater and top that off with a jacket. It was stitched such that neither would you be able to raise your hands higher than your ears nor drop them fully. The same would repeat itself with socks. One sock over another over another, and then wear canvas shoes. Like why?

Bus stops - Everything in Delhi was made of Iron. I suppose that was to ensure longevity. Bus stops would have iron railings that you could sit on while waiting for a bus. Of course, it used to be freezing cold and if it was foggy, bound to be slightly wet as well. Sitting on it would mean your ass would be wet. Pre-warmed railings were an asset! If someone was getting off to board a bus, you made sure you got that spot.

I survived all that and sometimes wish I could have it back!

A GPU, which stands for graphical processing unit, renders images such as graphics on the computer screen. It needs to process all the pixels simultaneously so that it can render every pixel on the screen. They were an accessory 25 years ago for those who wished to play games on their computers.

The way AI algorithms work, they throw a ton of spaghetti on the wall and see what sticks. This implies the need for parallel processing, which GPUs excel at. CPUs which run your computers are better at sequential tasks.

Since Nvidia was the market leader in the space, it became the preferred supplier of GPUs to most companies that were building AI models. In some ways, Nvidia has become a proxy for the entire AI industry in the world. They have the fastest, most-preferred and well-tested GPUs in the world.

OpenAI announced ChatGPT, and the war for GPUs began. Companies are raising ever larger funding rounds to buy more and more GPUs to run larger and larger models all the pursuit of something called Artificial General Intelligence (AGI). AGI is about producing the sort of intelligence that can cut across subjects, much like humans do.

The agreed thesis is that if companies double down on the models and keep increasing the number of parameters, it would automatically deliver AGI.

For the past 2 years, companies have been doubling and doubling their orders to Nvidia. It reflects in their revenue.

There has been an even greater rise in their share prices. Their revenues have grown 5-fold, their share price 10-fold. That implies that investors are assuming that there is more growth to be secured.

Given all this, there is some strange dealmaking going on.

The profit machine, Nvidia, has been investing its profits into its customers!

NVIDIA and OpenAI today announced a letter of intent for a landmark strategic partnership to deploy at least 10 gigawatts of NVIDIA systems for OpenAI’s next-generation AI infrastructure to train and run its next generation of models on the path to deploying superintelligence. To support this deployment including datacenter and power capacity, NVIDIA intends to invest up to $100 billion in OpenAI as the new NVIDIA systems are deployed. The first phase is targeted to come online in the second half of 2026 using NVIDIA’s Vera Rubin platform.

Source: OpenAI

OpenAI will buy goods from Nvidia and give it paper in return.

Today, Microsoft, NVIDIA and Anthropic announced new strategic partnerships. Anthropic is scaling its rapidly growing Claude AI model on Microsoft Azure, powered by NVIDIA, which will broaden access to Claude and provide Azure enterprise customers with expanded model choice and new capabilities. Anthropic has committed to purchase $30 billion of Azure compute capacity and to contract additional compute capacity up to 1 gigawatt.

Source: Nvidia

NVIDIA and Nokia today announced a strategic partnership to add NVIDIA-powered, commercial-grade AI-RAN products to Nokia’s industry-leading RAN portfolio, enabling communication service providers to launch AI-native 5G-Advanced and 6G networks on NVIDIA platforms. NVIDIA will also invest $1 billion in Nokia at a subscription price of $6.01 per share. The investment is subject to customary closing conditions.

Source: Nvidia

Why would Nvidia invest its cash into its customers? Maybe its customers are not going to be able to afford its products any longer? Its customers are unable to make the kind of money that they should be able to and are about to fail? This is one way of keeping that from happening. By kicking the can down the road, Nvidia could be hoping that these companies would be able to find the kind of revenues that they hope to find.

When everyone is looking for gold, sell them spades. But the spades will continue to sell only if there is gold to be found.

You could loan them the money to buy the spades, hoping that at some point they will indeed find gold. What if they don’t?

People who would know more than we would,

A viral AI-driven analysis of US SEC filings has resurfaced the scale and frequency of Nvidia CEO Jensen Huang’s stock sales this year, showing the chipmaker’s founder executed near-daily liquidation of shares worth roughly USD 13 million per day for several months.

Source: Moneycontrol

This would translate to $286 million in shares sold per month, assuming 22 trading days. Why only $13 million? If he did sell more aggressively, it would get alarm bells ringing.

Thiel Macro LLC fund’s Form 13F filing indicated that the investor no longer had any shared in Nvidia as of September 30. As a result, his staggering departure fuelled the sale of more than 537,000 shares in the company from July to September. According to Investing.com’s analysis of the firm’s average stock price during this period, the shares’ sale price is estimated to be nearly $100 million.

Given the full-fledged Nvidia exit’s status, this shift has shrunk his fund’s equity book by nearly two-thirds. His holdings in the firm reportedly accounted for nearly 40% of his portfolio.

Source: Financial Express

Peter Thiel is selling all of his holdings in Nvidia.

The market was abuzz last week with the news that Michael Burry, who became famous for shorting the stock market during the 2008 mortgage crisis, was betting against Nvidia and Palantir Technologies. Both stocks fell recently, and the revelation of Burry’s position may have contributed to that.

Source: Motley Fool

There are several people making bets against Nvidia, and its own CEO is reducing his stake in the company. Why?

Unless the AI party is about to come to an end…

Quite frankly, Donald Trump has done everything in his power to crash the US stock market. Stupidly played with the tariffs globally. Antagonised allies. Played around with the largest trading partners, holding the relationship on a knife-edge. Kept flip-flopping on policies. The final straw in this stupid Olympics has been the government shutdown. It does not even seem like anyone in the US government is remotely bothered about keeping the federal government open.

Through all of this melee, the US stock exchange has persisted. It is up nearly 10% this year. But all of that evaporates very quickly when you take the AI trades out of it.

Source: Reuters

The tech stocks with an AI narrative are up 50% and that is the bulwark that is holding up the US Stock Exchange. Except that the AI story is what investors like to call unrealised potential.

AI has yet to translate into real value on the balance sheets of various companies. There is a rush to create infrastructure and solutions that would use that infrastructure, but who will be using these solutions at scale is not really something we know.

The Price to Equity Ratio, which is a measure of how much investors expect earnings to grow, reflects this. All of the tech major, Apple, Alphabet, Microsoft, have their P/E ratios in the mid-30s. The chip manufacturers such as Nvidia and AMD are in the 70s. Data centre suppliers such as Amazon are valued in the 40s.

Just for contrast, businesses that form the backbone of the American economy, such as ExxonMobil and FedEx, are valued at a P/E ratio of 14.

The entire US stock market is a big bet on AI at the moment. If you were to remove all the tech stocks, the market would be at best flat.

If the AI story comes apart, then what?

In India, there has been a great deal of consternation about the fact that we have not been building foundational AI models. We have not innovated much in this area and are net consumers of this technology. On the flip side…

The NIFTY100 is up 9% this year.

Source: Groww

Do you know the number of companies on the Indian exchanges that are up based on their AI narrative? Zip.

The Indian markets are free of the AI contagion, and if the AI train turns out to be a wreck, India is the anti-AI trade. It has survived Trump’s trade wars. It has been consistently growing, and it is mostly fuelled by domestic consumption.

So India offers the best hedge against the American markets!

Over the last decade, this law has opened up vast pools of capital for the organisations working in the development sector. HDFC Bank has the largest CSR Pool, which sits at close to Rs 1000 Crores (~$110 Million), which it gives away each year to various Non-Governmental Organisations (NGOs).

Source: CSR.GOV

The government publishes the stats every year, and these numbers can be analysed.

The interesting thing that I found in the stats is the steady growth of the outlay year on year. No company enjoys giving away money, and they would avoid it if they could. Despite the efforts of the best accountants, the CSR expenditure is rising at over 10% year on year.

To me, this is the strongest proxy for the India growth story. It also shows the strength of the underlying profit machines. Consistent and growing profits year after year.

If you are investing in an index fund in India, you need not be worried at all.

Any of you who have ever been in a negotiation knows that you need to try and guess what the other person is thinking. It takes intelligence to do that. The truly intelligent can do this not because they have a lot of data, but because they do it in the absence of it.

Most of the people engaged at the leadership levels of any organisation spend their time making decisions. Often their decisions are made in the absence of perfect information. The fact that they are not constrained by the lack of information and can function is what enables them to make a lot of money.

The current state of AI takes the opposite track and throws as much information as possible into a model and expects it to produce the right answer. Wherever knowledge gaps appear, the system hallucinates because it does not know how to reason in the absence of information.

Intelligence would be having the ability to reason and arrive at answers in the absence of information. Honestly, nobody seems to be building it.

In the name of religion, I can block entire roads, ambulances can find another route.

In the name of religion, I can create noise pollution and play stuff on infinite repeat; your brain turning to mush is not my problem.

In the name of religion, I can hang streamers all over the street even though it can be a physical hazard to someone.

In the name of religion, I can burst crackers in the middle of the road even when traffic is moving there.

In the name of religion, I can lynch someone for eating beef.

In the name of religion, I can pour milk on a stone and call it an offering while poor children struggle to get any.

In the name of religion, I can segregate people.

Indians will do everything in the name of religion except for the one thing that religion teaches you - be kind.

Credit rating for each nation is predicated on its banking system, amount of debt, etc. Western credit rating is top-notch, even if it precipitates a global banking collapse.

Credit rating agencies exist not to rate bonds, but to crash them when it suits the US. The US businesses then move in to buy the assets in those nations which have adopted free market principles. Once the plunder is complete, the ratings agencies change the rating.

Every developing nation needs to embark on austerity measures to set its economy on the right path, but the West sets it right by issuing more debt.

A 100% Debt to GDP ratio is only bad for a “developing nation” in the “global south”.

When a non-Western country does it, it is protectionism. When the West does it, it is national security.

To my knowledge, the burden of penetration is being borne by only one company, and it is called the State Bank of India.

The so-called fintech companies, in the meantime, have gone after the cream of India, penetrating the urban and peri-urban markets to saturation. Payments is an unforgiving business, and the margins are razor-thin. If you process about $1 billion in India, you would net about $10 -20 million in margins. Imagine the technology infrastructure one needs to process those kinds of volumes. That revenue would only suffice to pay for the technology one would be running.

The only means for these companies to turn profitable is through lending. A relatively high-margin business. For most lenders, the cost of recovery can be rather high and mitigating this cost can make a company valuable.

Indians do not normally borrow a lot.

It said, “At 42.9 per cent of GDP (at current market prices) in June 2024, India’s household debt is relatively low compared to other EMEs, however, it has increased over the past three years”.

According to the report, household debt stood at 42.9 per cent of GDP at current market prices in June 2024. The report underlined that the rise in household debt is primarily driven by an increase in the number of borrowers, rather than a rise in the average indebtedness per individual.

Borrowing by individuals in the household sector accounted for 91 per cent of the total household financial liabilities as of March 2024.

The data revealed that individuals primarily borrow for three main purposes, consumption which includes personal loans, credit card debt, and loans for consumer durables.

Source: Times of India

At 42.9%, the total household borrowing in India stands at about $1.7 trillion through formal channels. The per capita household debt in India is a little more than $1000. Contrast this with the US

Total household debt increased by $185 billion to hit $18.39 trillion in the second quarter, according to the latest Quarterly Report on Household Debt and Credit. Mortgage balances grew by $131 billion and totaled $12.94 trillion at the end of June. Auto loan balances also increased, rising by $13 billion to reach $1.66 trillion.

Source - New York Fed

The $18.39 trillion represents about 60% of GDP. But in absolute terms, it is 37 times that of India. This is with a population that is one-quarter that of India. The per capita household debt in the US is a little more than $38000.

Commissions are based on volumes. Of course, the fintech companies would love to turn over more money.

For lending to take place, you need to know the right candidate for the debt. Enter Unified Lending Interface (ULI).

The RBI announced the introduction of ULI last month. The platform is supposed to act as a data broker which provides account aggregators with all the information pertaining to the financial position of individuals so they can make decisions about processing loans. Ostensibly, the authorisation is consent-based, but given the number of dark designs tech companies use, they will either fool you or force you into consent.

ULI can provide access to all sorts of information, including land records.

To me, this entire platform seems like the product of intense lobbying.

Account Aggregators (AA) are non-banking financial companies, licensed by RBI, that act like a bridge, to deliver data from Financial Information Providers (FIP) that hold your personal or corporate financial data to Financial Information Users (FIU) that are providing financial services to you.

Source: Sahamati

Most of the Account Aggregators in India are tech companies that have been built within the last 2 decades. Almost all of them have been heavily reliant on affluent consumers in urban India for their growth. It does not seem to me that this technology layer, which will be used to penetrate deeper into the population that does not possess a credit score yet. Instead, it will be put to use to mine data about the affluent lot so they can be sold more lending products.

So get ready for unsolicited loan offers against your property, fixed deposits and mutual funds.

Every one of those states except for Yemen adopted a monarchy. Yemen was the only state to become a semi-presidential Democratic Republic. Their current economic predicament says a lot about which system served the region better.

On the other side of India, Singapore adopted what can only be described as a democratic dictatorship. Lee Kuan Yew was in office for 31 years.

When countries are poor, democracy is a bane. It forces leaders into short-termism and puts them in a position of constant squabbles where they must pull each other down. What is needed at that time is a benevolent dictator who will drive with a vision and deliver long-term outcomes.

In the book, ‘What Makes Democracies Endure?’ Przeworski et al. offer this fascinating insight – ‘a democracy can be expected to last an average of about 8.5 years in a country with a per capita income under US$1,000 per annum, 16 years in one with income between US$1,000 and US $2,000, 33 years between US$2,000 and US$4,000 and 100 years between US$4,000 and US$6,000 … Above US$6,000, democracies are impregnable … [they are] certain to survive, come hell or high water.’

One only has to look at the history of Asian economies (China, Indonesia, Korea, Malaysia, Singapore, Taiwan and Thailand) to see how this is borne out. And even beyond Asia, Pinochet’s Chile and Fujimori’s Peru are examples of economic success in lands bereft of democracy.

The reason for this ‘anomaly’ is that each of these dictators, whatever their faults (and there were many), was able to ensure some semblance of property rights, functioning institutions, growth-promoting economic policies (for example, in fiscal and monetary management) and an investment climate that buttressed growth – the things that democracy promises to do.

In India, there is a sharp Urban vs Rural divide in the per capita income.

The government of India does not supply statistics on Rural Per Capita Income Vs Urban Per Capita Income, so I am forced to use proxies and anecdotal evidence to determine this.

According to the Chief Economic Advisor, the per capita income of India is Rs 2.12 Lacs. (~ $2400).

As per a survey undertaken by the NSSO, the percentage distribution of agricultural households owning less than two hectares of land is 89.4%. 90% of the households in rural India have less than 5 acres of land. I would be willing to further make an argument that half of them would probably own about 2 acres of farm land.

An acre of paddy yields profits of about Rs 25,000 per season. Assuming one plants for a couple of seasons a year, this number would yield Rs 50,000 in annual income. Say, one were to plant a crop such as peanut or potato, the profits from an acre would be about Rs 50,000 per season. Again, two seasons of cropping would yield Rs. 1,00,000 in income. A farmer might not plant the same crop for both seasons, and let us split the difference and say they earn about Rs 75,000 per acre.

My assumption is that a good 50% of smallholder farmers have about 2 acres of land, which would give them Rs 1,50,000 (~ $1800) per year in income. This is not neatly spread like butter on your toast but comes in lump sums at the time of harvesting, which causes them to go into a debt trap and eventual poverty, but that is for another time.

Now coming to the anecdotes, my maid earns about Rs 30,000 per month working multiple houses. That translates to about Rs 3,60,000 (~ $3800) per year. Let me state, unlike the farmer, this is not household income but individual income. Her household income would at least be twice as much, if not more. Thus, the migration to cities.

But going back to the per capita income story, one can posit that the per capita income of Rural India is less than half of that of Urban India. While I did not plan this, if you take the average of Rs. 1,50,000 and Rs. 3,60,000, we arrive at Rs. 2,55,000, which is not too far from the per capita income touted by the Chief Economic Advisor.

If we go back to the assertion made by Przeworski mentioned above, it is plain to see that the Urban Indian is operating in a different system from rural India.

It is unequivocally true that democracy is not working for rural India, while it is just sort of, kind of working for Urban India. In some ways, Urban India is foisting democracy on Rural India. This has played out in the form of armed rebellions in certain places known as the Naxal movement; in the form of demands for statehood; and in the form of farmers’ protests in the capital.

Unfortunately, the rural Indian has no real recourse. Their protests have not yielded much. Why?

Revolutions rarely start on the street; they often start in the courtyards of palaces when an elite who could not make it to the top decides to tear apart the very system that would not allow him (it is usually a him) to rise.