We’ve constantly been sharing a list of our recent reads in our weekly emails for The Good Investors.

Do subscribe for our weekly updates through the orange box in the blog (it’s on the side if you’re using a computer, and all the way at the bottom if you’re using mobile) – it’s free!

But since our readership-audience for The Good Investors is wider than our subscriber base, we think sharing the reading list regularly on the blog itself can benefit even more people. The articles we share touch on a wide range of topics, including investing, business, and the world in general. 

Here are the articles for the week ending 01 February 2026:

1. Anthropic Lowers Gross Margin Projection as Revenue Skyrockets – Sri Muppidi

Anthropic last month projected it would generate a 40% gross profit margin from selling AI to businesses and application developers in 2025, according to two people with knowledge of its financials. That margin was 10 percentage points lower than its earlier optimistic expectations, though it’s still a big improvement from the year before…

…If Anthropic also counted inference costs for Claude chatbot users that don’t pay for a subscription, its gross margin would be about 38%, or a few percentage points lower than for paid users, based on The Information’s analysis…

…Anthropic has previously projected gross margins above 70% by 2027, and OpenAI has projected gross margins of at least 70% by 2029, which would put them closer to the gross margins of publicly traded software and cloud firms. But both AI developers also spend a tremendous amount on renting servers to develop new models—training costs, which don’t factor into gross margins—making it more difficult to turn a net profit than it is for traditional software firms.

The inference costs are in addition to costs from training the models. Anthropic last month expected its costs for training its AI models for 2025 to be roughly $4.1 billion, up roughly 5% from its summer projections. OpenAI, meanwhile, expected to spend $9.4 billion on compute for training its AI models last year.

2. A business that scales with the value of intelligence – Sarah Friar

We launched ChatGPT as a research preview to understand what would happen if we put frontier intelligence directly in people’s hands…

…As ChatGPT became a tool people rely on every day to get real work done, we followed a simple and enduring principle: our business model should scale with the value intelligence delivers…

…Looking back on the past three years, our ability to serve customers—as measured by revenue—directly tracks available compute: Compute grew 3X year over year or 9.5X from 2023 to 2025: 0.2 GW in 2023, 0.6 GW in 2024, and ~1.9 GW in 2025. While revenue followed the same curve growing 3X year over year, or 10X from 2023 to 2025: $2B ARR in 2023, $6B in 2024, and $20B+ in 2025. This is never-before-seen growth at such scale. And we firmly believe that more compute in these periods would have led to faster customer adoption and monetization.

3. 50x in 5 Years – Joe Raymond

I discovered Cable Information Systems (CIS) in a page of one-liner descriptions of companies in the OTC edition of Moody’s Manual.

It was trading for a dollar per share…

…Believe it or not, Cable Information Systems had 50,000 subscribers in 1980 which placed them in the top 10 U.S. cable companies.

The company had about 1 million shares outstanding which were inactively trading at a dollar a share in the pink sheets.

At the time, it was said that cable subscribers were going to be worth $1,000 each to an operator of cable services. Thus, it became apparent that Cable Information with 1 million shares outstanding was worth $50 million although it was selling at a market value of only $1 million.

A second way of valuing a cable company was to apply the then-going multiple to cash flow, deduct debt, and divide by outstanding shares. Doing that I also came up with $50 a share.

So, using the two ways of valuing a cable company at the time, I found a $1 stock worth $50.

I asked Peter if he knew of anyone that cared to sell shares, and he told me that some of the employees were shareholders and, from time to time, some of them were interested in selling. I asked him if he would give them my name and number and he said gladly, they’d be happy to know of me.

Over a period of months, some of these employees called and asked if I would buy their shares. I said yes, I am glad to pay the current market price of approximately $1 per share.

Before buying, I told any caller offering shares to me, “Look, I want to make clear to you that I’m buying because I think the shares are worth a heck of a lot more than a dollar and if I were you, I would not be eager to sell.”

As employees, I wanted them to know I felt strongly it was not a good idea to sell. After questioning them and explaining why they should not sell, some people still sold me their shares…

…Late in the year, 1981, Peter telephoned me to tell me that he was selling out at $48 in cash to John Malone, who was the biggest cable operator in the United States.

My first reaction was, “Wow, two dollars short of what we had calculated it was worth.”

But Peter told me that there were two dollars being put into escrow and they will probably be paid to shareholders as well, bringing the total consideration to $50…

…Here’s what Larry was looking at in Moody’s Manual back in 1977:

Sales were growing double digits and accelerating. Margins were expanding.

The stock traded between $0.38 and $1.00 in 1977. The normalized P/E ratio was 1x on the low end and 3x on the high end.

There was some debt, as was common with fast growing cable companies at the time. The EV/EBITDA at $1 per share was 5x.

4. My Interview With Andy Jassy: OpenAI, Trump, Power and the Future of AWS – Jessica E. Lessin and Andy Jassy

Andy Jassy: I think that we’re excited about agentic commerce. I think that it has the chance to make it easier for customers to find what they want. If you know what you want, it’s pretty hard to find a better experience than popping onto Amazon and searching and finding it.

But the one place still where physical retail has some advantages, in my opinion, is the ability to go in, not know what you want, ask questions, refine those questions, have somebody point you to different things. And I think agents are going to help customers with that type of discovery. And it’s part of why we’ve invested so much in Rufus, which is our shopping assistant, which has really gotten quite good.

And I think that over time that we will work with other third-party agents as well. I think today the experience hasn’t been great yet. You know, I think that a lot of these third-party agents, they don’t have your buying history, they don’t have what you like, a lot of the information about pricing and the product is off.

But over time, I do believe that will get better. I also think there needs to be the right value exchange between the agents and between the retailers themselves, but I am optimistic that those will work out. We’re having conversations with lots of people and I’m very bullish on agentic commerce…

…Jassy: As you know, the chips are such an important part of the performance and the cost structure for people running technology infrastructure. We learned in the CPU side of the business, we had this deep relationship with Intel, which we still do. But when you have a significant leader, it’s not always their priority to take price performance down for customers.

And one thing we learn about customers over and over and over again is they want better price performance. And so we built Graviton, our own custom CPU silicon, which is about 40% more price performance than the leading other x86 processors. And that has been really great for our customers and business.

And about 90% of our top 1,000 customers now use Graviton in a very significant way. And we just saw this same movie happening in the AI space. And we have a very deep partnership with Nvidia, and we will for as long as I can foresee. But customers badly want better price performance. And so that’s why we built Trainium.

Our Trainium2 chip has been fully subscribed. Anthropic runs hundreds of thousands of Trainium2 chips as they’re training their next model of Claude on top of it. It’s a multi-billion dollar business. And we just released Trainium3 which is our next version of chip, which is 40% more price performant than Trainium2.

And Trainium2 was about 30% to 40% more price performant than the other leading GPUs out there. If you want to allow customers to be able to use AI as expansively as they want, you must take the cost of inference down. And the chip is a big piece of it…

…[Jassy:] I think we’re just in this stage right now where there is so much demand. And, you know, we’re not at this point, we’re not just trying to guess whether there’s demand. We have so much demand. I think the industry would tell you as a whole, there is still not enough capacity, even though it’s gotten better than it was 18 months ago, we could still be growing faster if we had more capacity…

…[Jassy:] We’re in this really interesting stage of AI adoption, in my opinion. It’s very bar-belled.

You have a lot of use by the AI labs who are consuming gobs and gobs of compute right now, and maybe a runaway app or two like ChatGPT. Then the other side of the barbell are enterprises who are really using AI for cost avoidance or productivity. Customer service, business process automation, things like that.

But the middle of that barbell are all the enterprise workloads in production that are not using inference yet. That will. We’re still at this relatively early stage. I believe that the middle part of the barbell is going to be the largest absolute segment. And I think when enterprises get to deploying their production apps using inference and AI, they’re going to want those applications to run close to the rest of their other applications and where their data is.

And just the largest amount by a fair bit, resides in AWS. And so we’re making it easier and easier for customers to be able to run their core workloads with their AI workloads.

5. An Early Buffett Partnership Investment – Joe Raymond

The first investment Buffett disclosed in his partnership letters was Commonwealth Trust in 1958…

…Buffett started buying Commonwealth at $50 and thought it was worth $125…

…Warren was paying 5x earnings and 80% of book value. Seems like a good deal for a bank earning 20% on equity.

The second is the nature of the bank.

Commonwealth Trust had $50 million of deposits and only $20 million of loans, most of which were residential mortgages. It also had $21 million of government securities.

The asset mix appeared highly conservative, at least from a credit perspective.

While the assets looked solid, there was little equity in the business ($2 million of equity on $53 million of assets). You don’t see this sort of leverage today, but it was common practice amongst small thrifts in the ’50s…

…A sharp increase in reserves, coinciding with rising interest rates, caused a big hit in 1954. This was magnified by the fact that Commonwealth’s equity was only 4% of assets going into the year. Book value per share fell 34%.

By the time Buffett was buying in 1957, interest rates were moderating, reserves were healthy, and earnings and equity were about to resume their growth…

…Warren didn’t hold long.

He sold his shares for $80 apiece about a year after buying them.

This was a 25% premium over the prevailing market price at the time and represented a profit of 57% for the partnerships…

…Buffett said the buyer at $80 could expect to do well over time and that he was selling to recycle the proceeds into a better opportunity (Sanborn Map)…

…About a year after Buffett sold, Commonwealth Trust merged with Hudson County National Bank (HCNB). It was a share-for-share deal, and the combined bank kept the Hudson County name…

…Over the next eight years, HCNB grew its book value from $135 to $183 per share (4% CAGR) and paid $57 per share of dividends. The average stock price in 1968 was $228 (1.25x book value).

So, the buyer from Buffett at $80 in 1958 had $228 by 1968 plus $58 of dividends.

Including dividends, the total annual return was in the mid-teens…

…This is a good example of successful value investing.

Corporate performance was mediocre, but big follies were avoided. Equity grew slowly and dividends were paid.

A cheap entry price and average exit price produced a mid-teens IRR over more than a decade.

---

Disclaimer: The Good Investors is the personal investing blog of two simple guys who are passionate about educating Singaporeans about stock market investing. By using this Site, you specifically agree that none of the information provided constitutes financial, investment, or other professional advice. It is only intended to provide education. Speak with a professional before making important decisions about your money, your professional life, or even your personal life. We currently have a vested interest in Amazon (parent of AWS). Holdings are subject to change at any time.

We’ve constantly been sharing a list of our recent reads in our weekly emails for The Good Investors.

Do subscribe for our weekly updates through the orange box in the blog (it’s on the side if you’re using a computer, and all the way at the bottom if you’re using mobile) – it’s free!

But since our readership-audience for The Good Investors is wider than our subscriber base, we think sharing the reading list regularly on the blog itself can benefit even more people. The articles we share touch on a wide range of topics, including investing, business, and the world in general. 

Here are the articles for the week ending 18 January 2026:

1. “The Compute Theory of Everything” – Abdullah Al-Rezwan

Albanie referred two seminal essays by Hans Moravec: “The Role of Raw Power in Intelligence” (1976), and “When will computer hardware match the human brain?” (1998)

I glanced through the first essay, but read the second one. I was moved just by reading the abstract of the paper:

“This paper describes how the performance of AI machines tends to improve at the same pace that AI researchers get access to faster hardware. The processing power and memory capacity necessary to match general intellectual performance of the human brain are estimated. Based on extrapolation of past trends and on examination of technologies under development, it is predicted that the required hardware will be available in cheap machines in the 2020s.”…

…Despite acknowledging valid reasons to harbor skepticism, Moravec relied on his simple observations on computing:

“Computers doubled in capacity every two years after the war, a pace that became an industry given: companies that wished to grow sought to exceed it, companies that failed to keep up lost business. In the 1980s the doubling time contracted to 18 months, and computer performance in the late 1990s seems to be doubling every 12 months…

…At the present rate, computers suitable for humanlike robots will appear in the 2020s. Can the pace be sustained for another three decades? The graph shows no sign of abatement. If anything, it hints that further contractions in time scale are in store. But, one often encounters thoughtful articles by knowledgeable people in the semiconductor industry giving detailed reasons why the decades of phenomenal growth must soon come to an end.”

2. Venezuelan Historical Primer: Friend, Foe, Vassal – Collapse Intelligence Agency

Before the US Shale Revolution (Fracking) in ~2010, the consensus view among energy majors was that US domestic light sweet oil was dying.

It was thought the world had burned all the easy, high-quality oil. Future reserves were geographically concentrated in the Middle East or were “Trash Grade” (Canadian Bitumen, Venezuelan Extra-Heavy, Mexican Maya).

US Refiners (Valero, Chevron, LyondellBasell) decided that to stay profitable, they had to spend billions upgrading their facilities to process the “Trash Grade” oil that nobody else wanted. They built massive Delayed Cokers and Hydrocrackers.

By building machines that could eat $10/barrel sludge and turn it into $50/barrel gasoline, they guaranteed massive margins that simple refineries in Europe couldn’t touch…

…Meanwhile, Gulf Coast refiners weren’t building just for Venezuela; they were building for the neighborhood.

In the 90s, Mexico’s massive Cantarell Field was pumping huge volumes of “Maya” crude (heavy/sour)

Venezuela had Orinoco (extra heavy/sour).

The logic was that the Gulf of Mexico basin was destined to be the global hub for processing heavy oil. Refiners poured tens of billions of dollars into capital expenditures (CapEx) to optimize specifically for this metallurgic sludge…

…When fracking exploded in 2010, the US flooded the market with Light Sweet Crude (LTO).

The US refiners looked at all this light oil and realized, “We can’t use it efficiently.”

If you put Light Oil into a refinery built for Heavy Sludge, you run the equipment inefficiently. You under-utilize the coker units (billions in wasted sunk costs).

The US exports its own high-quality light oil to Asia/Europe (who have simple refineries) and must import heavy oil to satisfy the diet of the Gulf Coast processing complex.

The capacity exists because Venezuela effectively paid to build it (via Citgo) and US executives in the 90s bet the house that heavy oil was the only game in town. Formerly permissive national economic policies supercharged the technological development.

The recent US military operation isn’t just about seizing new resources; it’s about feeding a starving industrial monster that was specifically designed to eat only what Venezuela produces. And that industrial monster must feed the US economy because now the shale party is about to end. The US administration knows this. They have made a 100% rational decision to force a bloody showdown with Venezuela to fund US energy needs.

3. The AI revolution is here. Will the economy survive the transition? – Michael Burry, Dwarkesh Patel, Patrick McKenzie, and Jack Clark

Jack: Yes, something we say often to policymakers at Anthropic is “This is the worst it will ever be!” and it’s really hard to convey to them just how important that ends up being. The other thing which is unintuitive is how quickly capabilities improve—one current example is how many people are currently playing with Opus 4.5 in Claude Code and saying some variation of “Wow, this stuff is so much better than it was before.” If you last played with LLMs in November, you’re now wildly miscalibrated about the frontier…

…Dwarkesh: The million-dollar question is whether the METR productivity study (which shows that developers working in codebases they understood well had a roughly 20% decrease on merging pull requests from coding tools) or human equivalent time horizons of self-contained coding tasks (which are already in the many-hours range and doubling every four to seven months) is a better measure of how much speedup researchers and engineers at labs are actually getting. I don’t have direct experience here, but I’d guess it’s closer to the former, given that there isn’t a great feedback verification loop and the criteria are open-ended (maintainability, taste, etc.).

Jack: Agreed, this is a crucial question—and the data is conflicting and sparse. For example, we did a survey of developers at Anthropic and saw a self-reported 50% productivity boost from the 60% of those surveyed who used Claude in their work. But then things like the METR study would seem to contradict that. We need better data and, specifically, instrumentation for developers inside and outside the AI labs to see what is going on. To zoom out a bit, the massive and unprecedented uptake of coding tools does suggest people are seeing some major subjective benefit from using them—it would be very unintuitive if an increasing percentage of developers were enthusiastically making themselves less productive…

…Michael: Do you think the podium will keep rotating? From what I’m hearing, Google is winning among developers from both AWS and Microsoft. And it seems the “search inertia” has been purged at the company.

Dwarkesh: Interesting. Seems more competitive than ever to me. The Twitter vibes are great for both Opus 4.5 and Gemini 3.5 Pro. No opinion on which company will win, but it definitely doesn’t seem settled.

Jack: Seems more competitive than ever to me, also!…

…Jack: Coding has a nice property of being relatively “closed loop”—you use an LLM to generate or tweak code, which you then validate and push into production. It really took the arrival of a broader set of tools for LLMs to take on this “closed loop” property in domains outside of coding—for instance, the creation of web search capabilities and the arrival of stuff like Model Context Protocol (MCP) connectivity has allowed LLMs to massively expand their “closed loop” utility beyond coding.

As an example, I’ve been doing research on the cost curves of various things recently (e.g. dollars of mass to orbit, or dollars per watt from solar), and it’s the kind of thing you could research with LLMs prior to these tools, but it had immense amounts of friction and forced you to go back and forth between the LLM and everything else. Now that friction has been taken away, you’re seeing greater uptake. Therefore, I expect we’re about to see what happened to coders happen to knowledge workers more broadly—and this feels like it should show up in a diffuse but broad way across areas like science research, the law, academia, consultancy, and other domains.

Michael: At the end of the day, AI has to be purchased by someone. Someone out there pays for a good or service. That is GDP. And that spending grows at GDP rates, 2% to 4%—with perhaps some uplift for companies with pricing power, which doesn’t seem likely in the future of AI.

Economies don’t have magically expanding pies. They have arithmetically constrained pies. Nothing fancy. The entire software pie—SaaS software running all kinds of corporate and creative functions—is less than $1 trillion. This is why I keep coming back to the infrastructure-to-application ratio—Nvidia selling $400 billion of chips for less than $100 billion in end-user AI product revenue.

AI has to grow productivity and create new categories of spending that don’t cannibalize other categories. This is all very hard to do. Will AI grow productivity enough? That is debatable. The capital expenditure spending cycle is faith-based and FOMO-based. No one is pointing to numbers that work. Yet.

There is a much simpler narrative out there that AI will make everything so much better that spending will explode. It is more likely to take spending in. If AI replaces a $500 seat license with a $50 one, that is great for productivity but is deflationary for productivity spend. And that productivity gained is likely to be shared by all competitors…

…Michael: At some point, this spending on the AI buildout has to have a return on investment higher than the cost of that investment, or there is just no economic value added. If a company is bigger because it borrowed a lot more or spent all its cash flow on something low-return, that is not an attractive quality to an investor, and the multiple will fall. There are many non-tech companies printing cash with no real prospects for growth beyond buying it, and they trade at about 8x earnings…

…Michael: Well, value accrues, historically, in all industries, to those with a durable competitive advantage manifesting as either pricing power or an untouchable cost or distribution advantage.

It is not clear that the spending here will lead to that.

Warren Buffett owned a department store in the late 1960s. When the department store across the street put an escalator in, he had to, too. In the end, neither benefited from that expensive project. No durable margin improvement or cost improvement, and both were in the same exact spot. That is how most AI implementation will play out.

This is why trillions of dollars of spending with no clear path to utilization by the real economy is so concerning. Most will not benefit, because their competitors will benefit to the same extent, and neither will have a competitive advantage because of it.

I think the market is most wrong about the two poster children for AI: Nvidia and Palantir. These are two of the luckiest companies. They adapted well, but they are lucky because when this all started, neither had designed a product for AI. But they are getting used as such.

Nvidia’s advantage is not durable. SLMs and ASICs are the future for most use cases in AI. They will be backward-compatible with CUDA [Nvidia’s parallel computing platform and programming model] if at all necessary. Nvidia is the power-hungry, dirty solution holding the fort until the competition comes in with a completely different approach…

…Jack: The main thing I worry about is whether people succeed at “building AI that builds AI”—fully closing the loop on AI R&D (sometimes called recursively self-improving AI). To be clear, I assign essentially zero likelihood to there being recursively self-improving AI systems on the planet in January 2026, but we do see extremely early signs of AI getting better at doing components of AI research, ranging from kernel development to autonomously fine-tuning open-weight models…

…Michael: If I had the ear of senior policymakers, I would ask them to take a trillion dollars (since trillions just get thrown around like millions now) and bypass all the protests and regulations and dot the whole country with small nuclear reactors, while also building a brand-new, state-of-the-art grid for everyone. Do this as soon as possible and secure it all from attack with the latest physical and cybersecurity; maybe even create a special Nuclear Defense Force that protects each facility, funded federally.

This is the only hope of getting enough power to keep up with China, and it is the only hope we have as a country to grow enough to ultimately pay off our debt and guarantee long-term security, by not letting power be a limiting factor on our innovation.

4. Is Venezuela’s Oil Worth the Hassle? – Tomas Pueyo

This depends on how much oil can be extracted from Venezuela. Today, it’s ~1.1M barrels per day.

A barrel of oil is currently worth about $60:

But Venezuela’s oil is worse quality than most, so it sells for cheaper, ~$8 less as of today, or $52…

…But how much does it cost to extract a barrel of Orinoco oil and transport it and treat it to be sellable?

So of these $52, about $23 are hard costs, and each barrel yields around $29 in profit…

…The oil [in the Orinoco Valley] is extremely dense (heavier than water), extremely viscous (like pitch or molasses) and extremely dirty (over 5% sulfur and masses of metals like vanadium). The only deposit like this elsewhere in the world is Canada’s Athabasca oil sands.

To extract the oil, you have to first pump large amounts of steam into the formation, to melt the hydrocarbons, then use electrical pumps at the surface or in the bottom of the well, up to a kilometer deep, to lift it to the surface. Once there, the “oil” is far too viscous to transport by pipeline or ship, and far too heavy and dirty for most refineries to tackle. So it is diluted by mixing with a much lighter crude oil, or the “condensate” liquids from a gas field, or refined naphtha (a solvent which you can buy as “white spirit” in UK DIY stores). The resulting diluted crude oil (DCO) is exported as Merey blend. This is still one of the heaviest, dirtiest crude oils in the world (16 API, 3.5% sulfur, high acidity and metals content), but it flows just well enough to be transported if kept warm, and some of the world’s more complex refineries can handle it, and make transport fuels from it, although usually alongside other lighter crudes…

…The two best estimates suggest it would take tens of billions to maintain the existing infrastructure, and tens of billions more to go beyond that.

5. A Few Things I’m Pretty Sure About – Morgan Housel

I think the majority of society problems are all downstream of housing affordability. The median age of first-time homebuyers went from 29 in 1981 to 40 today. But the shock this causes is so much deeper than housing. When young people are shut out of the life-defining step of having their own place, they’re less likely to get married, less likely to have kids, have worse mental health, and – my theory – more likely to have extreme political views, because when you don’t feel financially invested in your community you’re less likely to care about the consequences of bad policy…

…There’s a long history of Americans cycling through how they feel about government and how politicians treat each other.

The 1930s were unbelievably vicious. There was a well organized plot to overthrow Franklin Roosevelt and replace him with a Marine general named Smedley Butler, who would effectively become dictator. The Great Depression made Americans lose so much faith in government that the prevailing view was, “hey, might as well give this a shot.”

It would have sounded preposterous if someone told you in the 1930s that by the 1950s more than 70% of Americans said they trusted the government to do the right thing almost all the time. But that’s what happened.

And it would have sounded preposterous in the 1950s if you told Americans within 20 years trust would collapse amid the Vietnam War and Watergate.

It would have sounded preposterous if you told Americans in the 1970s that within 20 years trust and faith in government would have surged amid 1990s prosperity and balanced budgets.

And equally absurd if you told Americans in the 1990s that we’d be where we are today.

---

Disclaimer: The Good Investors is the personal investing blog of two simple guys who are passionate about educating Singaporeans about stock market investing. By using this Site, you specifically agree that none of the information provided constitutes financial, investment, or other professional advice. It is only intended to provide education. Speak with a professional before making important decisions about your money, your professional life, or even your personal life. We currently have a vested interest in Alphabet (parent of Google), Amazon (parent of AWS), and Microsoft. Holdings are subject to change at any time.

We’ve constantly been sharing a list of our recent reads in our weekly emails for The Good Investors.

Do subscribe for our weekly updates through the orange box in the blog (it’s on the side if you’re using a computer, and all the way at the bottom if you’re using mobile) – it’s free!

But since our readership-audience for The Good Investors is wider than our subscriber base, we think sharing the reading list regularly on the blog itself can benefit even more people. The articles we share touch on a wide range of topics, including investing, business, and the world in general. 

Here are the articles for the week ending 11 January 2026:

1. Ten things about Venezuela: on oil, geopolitics and drugs – Michael Cembalest

 Venezuela is not a large part of the global oil production picture, at least not right now.  The impact on global oil markets from the US invasion/arrest of Maduro should be minor…

…The US is still highly reliant on petroleum for 90% of transport energy consumption with the remainder mostly made up of natural gas and biomass, and for ~33% of industrial production (mostly high temperature heat and industrial feedstocks).   The amounts of oil used for residential and commercial heating is pretty negligible, less than 10% of the respective totals…

…The oil intensity of GDP is gradually declining in most of the world.  At some point, this ratio may drop low enough that disruptions in oil supplies will be less of an issue for growth and consumer spending…

…While US oil production is tilted towards light oil, US refining capacity is more even split among light, medium and heavier grades.  Note how heavy and medium normalized oil production in Venezuela aligns better with US refining gaps…

…Venezuela also possesses largely untapped reserves of critical minerals like coltan (niobium-tantalum), rare earth elements (REEs), nickel, gold, bauxite and iron ore.  The Orinoco Mining Arc, which spans 111,843 sq km, contains documented deposits of coltan (tantalum ore), cassiterite (tin ore), rare earth elements, bauxite, gold, and lithium reserves.  Coltan is used for manufacturing tantalum capacitors used in advanced electronic systems, including military communications equipment, missile guidance computers and radar systems. Rare earth elements enable permanent magnets required for precision-guided munitions, aircraft actuators and electromagnetic systems. Cassiterite provides tin for solder in electronics assembly, including defense systems while bauxite feeds aluminum production for aerospace applications…

…Iran and Venezuela have exchanged oil, gold and infrastructure assistance using Iran’s Islamic Revolutionary Guard Corps and Hezbollah-linked front companies for money laundering and sanctions evasion…

…Over 120 Russian troops reportedly operate in Venezuela and lead the “Equator Task Force.”  Russian advisers provide training across multiple domains including infantry, drone operations, special forces, military intelligence, signals intelligence, armor, aircraft, artillery and domestic surveillance

China has extensive ties with Venezuela; note the disproportionate amount of Chinese loans to Venezuela vs other Latin American countries (most of these loans were originated over a decade ago).  China’s military connections with Venezuela involve arms sales (missiles, jets, naval vessels), defense cooperation and strategic support; it’s not clear what the benefit has been for Venezuela, at least based on last week.

2. Steam, Steel, and Infinite Minds – Ivan Zhao

My co-founder Simon was what we call a 10× programmer, but he rarely writes code these days. Walk by his desk and you’ll see him orchestrating three or four AI coding agents at once, and they don’t just type faster, they think, which together makes him a 30-40× engineer. He queues tasks before lunch or bed, letting them work while he’s away. He’s become a manager of infinite minds…

…With AI agents, someone like Simon has graduated from riding a bicycle to driving a car.

When will other types of knowledge workers get cars? Two problems must be solved.

First, context fragmentation. For coding, tools and context tend to live in one place: the IDE, the repo, the terminal. But general knowledge work is scattered across dozens of tools. Imagine an AI agent trying to draft a product brief: it needs to pull from Slack threads, a strategy doc, last quarter’s metrics in a dashboard, and institutional memory that lives only in someone’s head. Today, humans are the glue, stitching all that together with copy-paste and switching between browser tabs. Until that context is consolidated, agents will stay stuck in narrow use-cases.

The second missing ingredient is verifiability. Code has a magical property: you can verify it with tests and errors. Model makers use this to train AI to get better at coding (e.g. reinforcement learning). But how do you verify if a project is managed well, or if a strategy memo is any good? We haven’t yet found ways to improve models for general knowledge work. So humans still need to be in the loop to supervise, guide, and show what good looks like…

…Before steel, buildings in the 19th century had a limit of six or seven floors. Iron was strong but brittle and heavy; add more floors, and the structure collapsed under its own weight. Steel changed everything. It’s strong yet malleable. Frames could be lighter, walls thinner, and suddenly buildings could rise dozens of stories. New kinds of buildings became possible.

AI is steel for organizations. It has the potential to maintain context across workflows and surface decisions when needed without the noise. Human communication no longer has to be the load-bearing wall. The weekly two-hour alignment meeting becomes a five-minute async review. The executive decision that required three levels of approval might soon happen in minutes. Companies can scale, truly scale, without the degradation we’ve accepted as inevitable…

… At the beginning of the Industrial Revolution, early textile factories sat next to rivers and streams and were powered by waterwheels. When the steam engine arrived, factory owners initially swapped waterwheels for steam engines and kept everything else the same. Productivity gains were modest.

The real breakthrough came when factory owners realized they could decouple from water entirely. They built larger mills closer to workers, ports, and raw materials. And they redesigned their factories around steam engines (Later, when electricity came online, owners further decentralized away from a central power shaft and placed smaller engines around the factory for different machines.) Productivity exploded, and the Second Industrial Revolution really took off.

We’re still in the “swap out the waterwheel” phase. AI chatbots bolted onto existing tools. We haven’t reimagined what organizations look like when the old constraints dissolve and your company can run on infinite minds that work while you sleep.

3. Our Approach to the Future – Hirotaka Shimizu

Venture companies seeking to go public typically expand by increasing sales through their hard-earned business models. Once sales exceed the break-even point, they begin to generate profits. During this process, they develop the organizational structures, governance frameworks, and compliance systems required of listed companies, steadily advancing toward an IPO. Only a limited number of these companies, under favorable conditions, ultimately succeed in going public.

Yet many of those that do achieve an IPO, often after significant struggles and setbacks, find their growth peak around the time of listing. According to the Ministry of Economy, Trade and Industry’s March 2024 report, “Research on How Startups Can Continue to Grow after Listing,” market capitalization growth typically peaks in the first year after listing and then declines uniformly from the second year onward. In fact, although the Tokyo Stock Exchange (TSE) Growth Market is intended to function as a gateway to higher-tier markets, only about one quarter of listed companies successfully make such a transition. Most are unable to achieve their anticipated growth trajectory and remain on the Growth Market. This is why IPOs are sometimes called jokingly by the public as “the final goal” of venture companies. To address this issue, the TSE reportedly plans to revise its continued listing criteria for the Growth Market by requiring companies listed for five years or more to have a market capitalization of at least ¥10 billion, thereby encouraging stronger post-listing growth.

Now then, why does growth come to a halt? There must be a reason. In my view, many venture owners concentrate too much of their attention and energy on their hard-earned business models. Yet all business models, even highly unique ones, have a shelf life. Every business inevitably moves from a growth phase to a maturity phase, and eventually to a decline phase. Companies that push aggressively during the growth phase and succeed in going public often discover that the differentiation they once created has diminished by the time they reach maturity. Their once-unique business models are imitated by competitors, or they unavoidably face intensified competition from companies with adjacent business models. As a result, they find themselves in a red ocean. In addition, the company growth cycle itself is shortening as information and technology continue to advance. This trend is particularly evident among venture companies in B2B marketing and technology domains. Once they face such situations, developing a new business model becomes increasingly difficult. Furthermore, listed companies are required to disclose financial information on a quarterly basis. In my view, this requirement can also discourage new investment, given the potential impact on share prices. I suspect that the current framework functions as a kind of “trap” into which many companies that manage to go public eventually fall.

To avoid this outcome, companies must continually conceive and pursue new business models while their existing models are still in a growth phase. However, most business managers fail to direct their attention to this imperative. In my view, this is because they lack long-term, ambitious goals. If managers were to set long-term goals, they would recognize that such goals cannot be achieved through a single business model and would therefore feel a natural imperative to develop the next one. Companies should, in my opinion, pursue growth driven by long-term goals, such as missions, visions, principles, aspirations, and ambitions, rather than relying on business models. I believe that the continued pursuit of these goals ultimately enables sustained corporate growth.

4. Peace and prosperity in Venezuela will come from democracy, not oil – Ricardo Hausmann

But then, concern: just hours after the raid President Donald Trump declared that he would now “run” Venezuela. He talked much about oil but not at all about democracy other than to dismiss María Corina Machado, Nobel peace laureate and leader of the democratic opposition…

…Instead, Mr Trump made clear, America will work with the dictator’s own vice-president. He spoke as if he owned the country and its assets. Venezuelans will be recipients of his benevolence, not agents of their destiny.

Removing a dictator—especially if leaving his henchmen and -women in charge—is not the same as rebuilding a country. And there is much to rebuild. When Mr Maduro came to power in 2013, Venezuelans were four times richer than they are today. A disaster followed: the largest economic contraction ever recorded in peacetime, triggering the departure of 8m Venezuelans. Brutality, repression and corruption accompanied the catastrophe.

At its heart was a systematic dismantling of rights: property rights, independent courts and free elections. Speaking out became a crime. As rights vanished, so did security, investment, trust and the power to imagine. People stopped planning for the future because the future no longer belonged to them.

The lesson is simple: prosperity does not come from oil, decrees or even benevolent rulers, but from rights. Rights create private property and security. They allow people to invest, innovate and dream. Restore rights, and society can recover.

Venezuelans now need neither revenge nor Trumpian improvisation, but a return to freedom and peace. The technology for that has already been invented: democracy, which is not just about voting but is a system for aggregating preferences while protecting liberties. Democracy aligns political authority with social consent and is the formula for sustained prosperity. Venezuela enjoyed it for much of the latter part of the 20th century. 

5. Trump’s Enormous C-Length Win over China – Collapse Intelligence Agency

When we talk about “Oil,” we are using a lazy bucket term. In reality, a barrel of oil is a soup of thousands of different molecules. Each geographic barrel is a unique fingerprint.

“C-Length” refers to the number of Carbon atoms chained together in a single molecule.

This is the fundamental biophysics of the economy. The length of the carbon chain determines State of Matter (Gas vs. Liquid vs. Solid) and Energy Density (how much work it can do).

Short Chains (C1–C4): Gases. They float away.

Medium Chains (C5–C12): Thin Liquids (Gasoline). They evaporate quickly.

Long Chains (C13–C20): Oily Liquids (Diesel/Jet). The “Goldilocks” zone for heavy work.

Very Long Chains (C50+): Solids (Asphalt).

The US/Venezuela/China trade war is essentially a fight over C20+ chains…

…To run a modern economy, you need a specific ratio of products: roughly 40% Gasoline, 30% Diesel, 10% Jet, 20% Industrial/Asphalt. This matches the general demand pattern of the economy.

But nature never gives you that exact ratio in the ground.

Scenario A: Refining Light Oil (US Shale – Mostly C5-C10)

You have too much Gasoline/Naphtha.

To make Diesel (C16), you have to mathematically glue molecules together.

Biophysics: It is energetically difficult and expensive to “Oligomerize” (fuse) small chains into big ones. You cannot efficiently run an industrial economy on shale oil alone because you can’t make enough Diesel/Jet fuel without massive waste.

Scenario B: Refining Heavy Oil (Venezuelan Orinoco – Rich in C20-C100)

You have huge long chains.

The Coker: You heat them up and “Chop” them. A C50 chain can be snapped into three C16 chains (Diesel).

Biophysics: It is thermodynamically efficient to “Crack” (break) a long chain into specific smaller pieces. This is why US Coking Refineries are the “Golden Key.” They take the cheapest feedstock (C50+ sludge) and turn it into the most valuable product (C16 Diesel)…

…The US possesses the “Holy Grail” of refining: Single-Site Deep Conversion.

US Advantage: A barrel of Orinoco sludge enters a Texas refinery and leaves as 80% High-Value Diesel/Jet and 20% solid Petcoke. It is processed in one location, efficiently.

Russian Flaw – The Mazut Glut: Russia cannot fully refine its own heavy barrels. Its refineries lack the depth of US “Coking” capacity.

Russia is forced to export massive volumes of Mazut (M-100)—a cheap, low-value heavy fuel oil—because they can’t crack it into diesel domestically. They have to ship this half-refined trash to buyers who can finish the job.

China: “Teapot” refineries in Shandong have effectively become the “Trash Cans” for the Eastern Bloc. They import Russian Mazut and Venezuelan Bitumen blend to crack it into diesel and asphalt.

The Eastern Bloc relies on shipping half-refined residue between countries to achieve what Texas does inside a single fence line. That creates a massive Thermodynamic Friction (shipping fuel oil is heavy and dirty) that the US avoids…

…Iranian Oil (Soroosh/Nowruz) and Russian Mazut: Heavy, but optimized for fuels (Energy).

Venezuelan Oil (Merey 16) and Canadian Tar Sands: The global gold standard for high-yield Bitumen (Asphalt).

China consumes massive amounts of asphalt for its ceaseless road/infrastructure construction. Losing Venezuelan supply implies a structural shortage of road-paving material.

With Venezuela (Orinoco) gone to the US, and Canada (Tar Sands) logically aligned with the US (despite mercantile friction), China has only one source left for heavy, complex oil: Iran.

The Bottleneck: This forces China into a single-point dependency. If the US/Israel acts against Iranian export terminals (Kharg Island), the Eastern Bloc has minimal access to the heavy oil required for their specific refinery configurations.

Russia can’t help: Russia produces “Urals” (Medium Sour), it’s true heavy oils are limited in production and export.

Canada via the TMX pipeline supplies 200 000 bpd. This is the bpd spoken for CHINA crude. TMX total is 800 – 900 thousand bpd. And this pipeline is MAXED out. China can’t get any more. TMX schedules are spoken for. Other consumers have contractual claim.

You can’t pave a road with Iranian Soroosh/Russian Heavy efficiently; you get less asphalt and more waste…

…By seizing Venezuelan Orinoco heavy oil, the US also effectively secures the highest-value feedstock for its specialized machine, forcing China to run its “Teapot” refineries on inferior or politically volatile alternatives. This heavy oil sludge can be more easily cracked into lower forms as needed for desired usage.

Heavy oils give US optionality in refining. It is more efficient to “chop” that it is to “glue.”

The US will very likely install governance and corporate structure that is supplicating to its national needs. It can begin to squeeze the Eastern Bloc slowly by reducing exports of Merey 16. Or it can simply increase prices. China was able to buy this sanctioned oil at discount.

Now the US controls this oil supply. It’s categorization is “Clean.” So China pays fair market prices for continuing their infrastructure construction.

The same way that China uses REE controls.

We can make an estimation that China currently relies upon Venezuelan bitumen for roughly 50% of its asphalt production needs.

Depending on the mood of the US administration, this is about to get very expensive or outright disappear from China’s procurement.

Whether by design or coincidence, the US now has a very real wartime advantage against China.

It’s likely the US does not recognize this fully. They just wanted China OUT.

---

Disclaimer: The Good Investors is the personal investing blog of two simple guys who are passionate about educating Singaporeans about stock market investing. By using this Site, you specifically agree that none of the information provided constitutes financial, investment, or other professional advice. It is only intended to provide education. Speak with a professional before making important decisions about your money, your professional life, or even your personal life. We currently have no vested interest in any companies mentioned. Holdings are subject to change at any time.

We’ve constantly been sharing a list of our recent reads in our weekly emails for The Good Investors.

Do subscribe for our weekly updates through the orange box in the blog (it’s on the side if you’re using a computer, and all the way at the bottom if you’re using mobile) – it’s free!

But since our readership-audience for The Good Investors is wider than our subscriber base, we think sharing the reading list regularly on the blog itself can benefit even more people. The articles we share touch on a wide range of topics, including investing, business, and the world in general. 

Here are the articles for the week ending 04 January 2026:

1. Authenticity after abundance – Adam Mosseri 

Everything that made creators matter—the ability to be real, to connect, to have a voice that couldn’t be faked—is now suddenly accessible to anyone with the right tools. Deepfakes are getting better and better. AI is generating photographs and videos indistinguishable from captured media. The feeds are starting to fill up with synthetic everything…

…We are now seeing an abundance of AI generated content, and there will be much more content created by AI than captured by traditional means in a few years time. We like to talk about “AI slop,” but there is a lot of amazing AI content that thankfully lacks the disturbing properties of twisted limbs and absent physics. Even the quality AI content has a look though: it tends to feel fabricated somehow. The imagery today is too slick, people’s skin is too smooth. That will change; we are going to start to see more and more realistic AI content.

Authenticity is fast becoming a scarce resource, which will in turn drive more demand for creator content, not less. The creators who succeed will be those who figure out how to maintain their authenticity whether or not they adopt new technologies. That’s harder now—not easier—because everyone can simulate authenticity. The bar is going to shift from “can you create?” to “can you make something that only you could create?” That’s the new gate…

…But flattering imagery is cheap to produce and boring to consume. People want content that feels real. We are going to see a significant acceleration of a more raw aesthetic over the next few years. Savvy creators are going to lean into explicitly unproduced and unflattering images of themselves…

…Social media platforms are going to come under increasing pressure to identify and label AI-generated content as such. All the major platforms will do good work identifying AI content, but they will get worse at it over time as AI gets better at imitating reality. There is already a growing number of people who believe, as I do, that it will be more practical to fingerprint real media than fake media. Camera manufacturers could cryptographically sign images at capture, creating a chain of custody…

…In a world of infinite abundance and infinite doubt, the creators who can maintain trust and signal authenticity—by being real, transparent, and consistent—will stand out.

As for Instagram, we’re going to have to evolve in a number of ways, and fast. We need to build the best creative tools, AI-driven and traditional, for creators so that they can compete with content fully created by AI. We need to label AI-generated content clearly, and work with manufacturers to verify authenticity at capture—fingerprinting real media, not just chasing fake. We need to surface credibility signals about who’s posting so people can decide who to trust. And we’re going to need to continue to improve ranking for originality.

2. 2025’s biggest investing lesson – slow down – Chin Hui Leong

HERE Is the uncomfortable truth about 2025: The year’s biggest wealth destroyer was not tariffs, AI disruption, or interest rate uncertainty.

It was speed…

…At the start of 2025, traders reacted swiftly to every hint about interest rate movements.

A strong jobs report? Sell immediately – fewer rate cuts ahead.

A weak inflation print? Buy before everyone else does.

This behaviour assumed that being first to interpret the data would translate into superior returns.

Let us test that theory with 2024’s track record. Goldman Sachs predicted five rate cuts. We got three.

Traders priced in a 73 per cent chance of a March 2025 cut. The first cut came in September, six months later. The market expected 1.5 percentage points of cuts. We got one.

In other words, the number of cuts was wrong, the timing was off, and the size of the cuts were lower than expected.

Yet despite these spectacular misses, the S&P 500 rose more than 23 per cent in 2024.

The lesson: You can be completely wrong about interest rates and still do well in the market – if you stay invested.

The investors who traded every data point, trying to front-run the US Federal Reserve, generated fees and anxiety.

The investors who ignored the noise generated returns…

… In his book Your Money and Your Brain, author Jason Zweig explains that our minds recognise patterns even when none exist.

But this is the kicker: We cannot switch this mechanism off at will…

…Consider this: Every major decline in 2025 was accompanied by an avalanche of negative headlines – detailed articles on what went wrong, podcasts dissecting the damage, and social media hot takes piling on.

Amid that onslaught, any good news was buried.

The investors who reacted to the noise sold at lows. The investors who waited for the noise to clear bought those shares from them.

Speed did not protect portfolios. Patience did.

3. Running Out of Runway – Poe Zhao

Last week’s dual IPO filings from Zhipu AI and MiniMax reveal a paradox at the heart of China’s AI model market. Both companies have proven they can build competitive technology. Both have validated their business models at the unit economics level. Both are running out of time…

…Zhipu grew from ¥57 million in 2022 to ¥312 million in 2024, a 130% compound annual growth rate. MiniMax achieved even more dramatic expansion, with revenue surging 782% to $30.5 million in 2024. In the first nine months of 2025, MiniMax generated $53.4 million, already exceeding its full-year 2024 results.

But losses grew faster. Zhipu’s adjusted net loss exploded from ¥97 million in 2022 to ¥2.47 billion in 2024. That’s 20x growth. MiniMax went from $7.37 million in losses in 2022 to $465 million in 2024.

The cash burn is brutal. Zhipu: ¥300 million monthly. MiniMax: ¥2 billion monthly. Zhipu’s mid-2025 reserves stood at ¥2.55 billion. Do the math. Six months later, both companies rushed to file IPOs. The December timing was necessity, not choice…

…Research and development consumed ¥2.2 billion of Zhipu’s budget in 2024. That’s a 26x increase from the ¥84 million spent in 2022. Within that R&D figure, ¥1.55 billion went directly to compute services. Computing infrastructure alone ate 70% of the entire R&D budget.

MiniMax shows better cost discipline but faces the same fundamental pressure. Training-related cloud computing costs reached $142 million in the first nine months of 2025. The company has managed to improve efficiency. The ratio of training costs to revenue dropped from 1,365% in 2023 to 266% in the first three quarters of 2025. But even at 266%, you’re spending nearly $3 on training for every $1 of revenue.

This creates the first paradox. At the transaction level, these businesses are profitable. Sell an API call or a subscription, you make money. Scale that up, you should make more money. But scaling requires maintaining competitive model quality. Competitive model quality requires constant compute investment. The compute investment grows faster than revenue. The more you sell, the more you lose…

…China’s entire large language model market totaled ¥5.3 billion in 2024, according to Zhipu’s prospectus. Enterprise customers contributed ¥4.7 billion of that. Individual consumers accounted for just ¥600 million.

Do the math. Zhipu burns ¥300 million monthly. MiniMax burns ¥2 billion monthly. Combined, that’s ¥2.3 billion per month. Annualize it and you get ¥27.6 billion. The two companies alone are burning through more than five times the entire current market size annually. And they’re not alone. Multiple other companies compete in the same space…

…Zhipu bet on scale. The company invested heavily in frontier model development. R&D spending jumped from ¥529 million in 2023 to ¥2.2 billion in 2024. Compute infrastructure dominated that budget. The strategy assumes that leading-edge capabilities justify the burn rate. Stay at the frontier, win the highest-value customers, eventually reach economies of scale.

MiniMax took the efficiency route. The company’s prospectus explicitly positions itself as capital-efficient. Cumulative spending from founding through September 2025 totaled approximately $500 million. The prospectus contrasts this with OpenAI’s estimated $40–55 billion in cumulative investment. That’s a 100x cost difference for comparable multimodal capabilities…

…This reveals what makes the situation structural rather than cyclical. Your strategy becomes irrelevant when competitive dynamics dictate behavior. Zhipu chose scale. MiniMax chose efficiency. DeepSeek’s emergence forced both to spend more regardless of their chosen path. In a true market, companies can differentiate on cost, quality, or features. In this market, everyone must match the pace of iteration or become obsolete. The pace keeps accelerating. The costs keep compounding.

4.Why We Worry – Part I – Fawkes Capital

This year alone, Google will spend roughly $60 billion more in annualised capex than it did before ChatGPT launched. Since late 2022, the company has deployed an additional $85 billion in cumulative capex on AI-related development. With similar spending levels expected next year, Google’s capex now exceeds its net profit – a sharp departure from pre-AI years, when capex represented only about 25% of profit…

…What is Google receiving in return for this extraordinary level of investment? At present, Google processes roughly 1.4 quadrillion AI tokens per month. If we make a simplifying assumption and apply Google’s API input pricing across all of those tokens, the result is an additional $21 billion of annualised revenue.

For context, this is not an especially compelling trade-off: $85 billion of incremental capex for $21 billion of low-margin revenue. In effect, Google is deploying vast sums of capital for what amounts to a modest 5% uplift in annual revenue, and materially lower returns than its core search and advertising franchise generates. A major outlay for just a 5% uplift in annual revenues doesn’t sound like a great use of capital to us.

And if this is the underlying economic reality for the industry leader, it is difficult to see how outcomes will be more favourable for its competitors. Over time, we doubt that the return on capital employed (ROCE) from datacentres will meaningfully improve from today’s levels…

…If Big Tech and data centre operators collectively spend around $400 billion on AI infrastructure in 2026, then, by our estimate, at least $80 billion in annual net income would need to be generated to justify that investment. The hurdle is high because processors, which make up the bulk of capex, have a useful life of only about five years. Back-solving this requirement implies that something like 333 million paying users of ChatGPT – roughly the entire US population – would be needed to support such economics.

Today, the numbers fall drastically short. Only around 5% of users (about 20 million people) pay for ChatGPT, and both paid and non-paid user growth has begun to stall in recent months. OpenAI’s attempt to introduce advertising as a revenue stream has met fierce consumer resistance. And unlike Google, directing users to websites does not generate economic value for OpenAI. This raises the critical question: how will OpenAI, or any non-advertising-based AI provider, monetise its service at the scale needed?…

…. A similar pattern emerged during the late-1990s dot-com bubble. Telecom operators, despite enormous capital outlays, found their services rapidly commoditised. Usage growth slowed, pricing power collapsed, and the industry could not extract the household revenues required to justify the capex binge. High returns initially attracted more competition, which eventually eroded margins for even the leading “pick-and-shovel” equipment suppliers. Investor belief in unassailable competitive advantages proved illusory. Once reality set in, the bubble burst and triggered a shallow recession…

…SemiAnalysis notes that Google’s TPU infrastructure now rivals NVIDIA’s latest commercially available GPUs – at significantly lower cost. Sensing the threat, NVIDIA has resorted to taking equity stakes in companies that depend on its support (OpenAI among them), effectively subsidising its customer base to stave off competition and preserve margins. This is not a sustainable strategy. Amazon’s upcoming Trainium 3 chip has also narrowed the performance gap and is likely to be cost-competitive upon release.

With credible alternatives emerging, NVIDIA’s 75% gross margins – the foundation of its current valuation – will not hold indefinitely. When investors fully appreciate this, and when that realisation intersects with the economic unsustainability of OpenAI’s model, the conditions for a sharp correction may be in place.

5. AI will kill all the lawyers – Sean Thomas

‘Last week we did an experiment, a kind of simulation. We took a real, recent and important case – a complex civil court appeal which I wrote, and it took me a day and a half. We redacted all identifying details, for anonymity and confidentiality, and we fed the same case to Grok Heavy AI. And then we asked it to do what I did. After some prompting, the end result was…’ He shakes his head. ‘Spectacular. Actually staggering. It did it in 30 seconds, and it was much better than mine. And remember, I am very good at this.’

He sits back, wry yet resigned. ‘It was at the level of a truly great KC. The best possible legal document. And all done in seconds for pennies. How can any of us compete? We can’t.’…

…James believes AI will work its way up the legal hierarchy. First the gruntwork, then the drafting, the citation, the argumentation. Eventually the majority of legal jobs will be replaced. ‘Process lawyers are obviously doomed. AI will handle the most complex probate and conveyancing cases in seconds. The most complicated human skill will be,’ he chuckles, sadly, ‘to scan and digitise paper documents. Barristers will make arguments in courtrooms that are drafted by AI, and then people will wonder why they are paying human barristers £200,000, and they too will disappear.’…

…I ask what he thinks this will do to his colleagues – psychologically, economically, emotionally. ‘At first, they will fight, like radicals. A losing battle. There will be attempts to outlaw the use of AI in various legal areas. But it won’t work, the economics will see to that. So lots of people who make a lot of money will, suddenly, not make that money. God knows what that might do to property prices, to politics, to all of us. Because it won’t just be the law.’

---

Disclaimer: The Good Investors is the personal investing blog of two simple guys who are passionate about educating Singaporeans about stock market investing. By using this Site, you specifically agree that none of the information provided constitutes financial, investment, or other professional advice. It is only intended to provide education. Speak with a professional before making important decisions about your money, your professional life, or even your personal life. We currently have a vested interest in Alphabet (parent of Google), Mastercard, Meta Platforms (parent of Instagram), and Visa. Holdings are subject to change at any time.