What We’re Reading (Week Ending 17 August 2025) - 17 Aug 2025
Reading helps us learn about the world and it is a really important aspect of investing. The late Charlie Munger even went so far as to say that “I don’t think you can get to be a really good investor over a broad range without doing a massive amount of reading.” We (the co-founders of Compounder Fund) read widely across a range of topics, including investing, business, technology, and the world in general. We want to regularly share the best articles we’ve come across recently. Here they are (for the week ending 17 August 2025):
1. Beyond the “Search” Box – Abdullah Al-Rezwan
Semrush tracked 260 billion rows of clickstream data on U.S. desktop users who began using ChatGPT in Q1 2025, comparing their Google Search sessions in the 90 days before and after adoption to a control group that never used ChatGPT. This setup allowed them to isolate whether ChatGPT adoption caused changes in traditional search behavior compared to natural trends over time.
The overall result of the study shows after adopting ChatGPT, users increased their Google Search sessions from 10.5 to 12.6 per week while also adding about 5 ChatGPT sessions weekly, suggesting ChatGPT use complemented rather than replaced Google searches.
Semrush shared some cohort level data by month which all show that despite sustained ChatGPT usage after adoption, Google Search usage remained resilient.
One may wonder if you keep using ChatGPT for longer than a year, perhaps it eventually changes your Google usage. That also doesn’t quite seem to be the case yet since a 500-day study by Semrush of users who began using ChatGPT in January 2024 found that Google search activity remained steady while ChatGPT usage stayed consistent after adoption.
2. Podcast: Amazon’s advertising strategy (with Adam Epstein) (Transcript here) – Eric Benjamin Seufert and Adam Epstein
Adam Epstein: I’ve been working in ad tech for seven plus years, and people have been decrying the end of the agency for as long as I can remember through the use of automated and simple software. But AI adds a new layer of complexity to everything, and complexity is good for agencies particularly. I’m not sure who coined the phrase, but they basically said agencies are cockroaches. And I believe that to probably be the case.
At least for the next three to five years, I actually don’t even think agentic AI will be a headwind for agencies—I think it will be a tailwind on two dimensions. First, the most scaled agencies in the world have been able to scale themselves not through data and technology, but through scaled processes, standard operating procedures, training collateral and docs to create expertise and uniform level of service across all their clients and team members. Well, guess what’s really good for training an LLM? Literally all of those documents.
Every agency I’ve talked to for seven years comes to me and says, “How is your off-the-shelf ad tech different than the off-the-shelf ad tech that you’re going to sell to the next agency tomorrow?” And the answer has always been it hasn’t been any different—it’s been exactly the same. But with agentic AI, you now no longer buy software—you hire software. You hire software, and you train software, and you develop a new teammate that you train and mold exactly as you would a new team member.
Agencies want this level of customization. They’re actually in a perfect position to do so because they’ve invested in collateral that allows them to train an LLM in a very efficient manner. We’re just catalyzing them and giving them the tools to do exactly that.
The other interesting thing with services businesses is that you typically need to linearly scale headcount as you scale customer and revenue growth. But I believe agentic AI will bring in a world for media agencies in particular where they’ll be able to exponentially increase customers and revenue while maintaining a flat headcount. Agentic AI will take all the operational work that teams are currently running and allow these agencies to scale in ways they’ve never been able to scale before. It’ll be a massive tailwind from an operating margin perspective, and I think people will actually start to value agencies on a different multiple than what they have in the past, given the fundamentally different margin profile.
3. Robotaxis & AI | Uncharted Territories Magazine | Tech Update Summer 2025 – Tomas Pueyo
Waymo is destroying the competition. It has surpassed Lyft in rides in SF, and is on track to surpass Uber within 8 months or so.
And this is with Waymo taking 2x longer and costing 70% more than Lyft!!!1 That’s how much better the Waymo experience is: People really care about not having a driver!…
…Uber said ride-hailing could grow by 25x if its price dropped under $1/mile…
…Uber couldn’t make it happen. But in Austin, now Tesla costs $1 per mile.
As a comparison, ride hail customers are currently paying nearly $3/mile.
If Tesla maintains this type of pricing, it won’t make sense for drivers to continue their job, and Uber and Lyft will crash.
8% of US workers are professional drivers…
…I didn’t realize how important this is until I read this article:
Something like 40,000 people die in traffic accidents in the US every year. The number is over one million per year globally.
There are over 5 million non-fatal injuries from car crashes each year that require medical attention in the US.
In 2010, the total costs from these events was $836 billion, or ~$2700 per American per year.
But these costs are just the tip of the iceberg because most of the cost of transportation, at >$2 trillion per year, comes from adjusting to human inadequacies.
Wait, what? Car accidents are costing trillions to the world economy? How?
- A big share of the materials in cars are due to safety. Without accidents, you can strip them out, saving all their money. Austin Vernon calculates we could make car weights 10x lower.
- Automobile shapes today trade off safety and aerodynamicity. Without safety, they can become more aerodynamic, and move faster at a cheaper cost.
- Cheaper transportation costs massively improve the economy.
- Lower weights on roads means less road wear, and hence less maintenance cost.
4. What If Money Expired? – Jacob Baynham
More than a century ago, a wild-eyed, vegetarian, free love-promoting German entrepreneur and self-taught economist named Silvio Gesell proposed a radical reformation of the monetary system as we know it. He wanted to make money that decays over time. Our present money, he explained, is an insufficient means of exchange. A man with a pocketful of money does not possess equivalent wealth as a man with a sack of produce, even if the market agrees the produce is worth the money.
“Only money that goes out of date like a newspaper, rots like potatoes, rusts like iron, evaporates like ether,” Gesell wrote in his seminal work, “The Natural Economic Order,” published in 1915, “is capable of standing the test as an instrument for the exchange of potatoes, newspapers, iron and ether.”…
…Gesell believed that the most-rewarded impulse in our present economy is to give as little as possible and to receive as much as possible, in every transaction. In doing so, he thought, we grow materially, morally and socially poorer. “The exploitation of our neighbor’s need, mutual plundering conducted with all the wiles of salesmanship, is the foundation of our economic life,” he lamented.
To correct these economic and social ills, Gesell recommended we change the nature of money so it better reflects the goods for which it is exchanged. “We must make money worse as a commodity if we wish to make it better as a medium of exchange,” he wrote.
To achieve this, he invented a form of expiring money called Freigeld, or Free Money. (Free because it would be freed from hoarding and interest.) The theory worked like this: A $100 bill of Freigeld would have 52 dated boxes on the back, where the holder must affix a 10-cent stamp every week for the bill to still be worth $100. If you kept the bill for an entire year, you would have to affix 52 stamps to the back of it — at a cost of $5.20 — for the bill to still be worth $100. Thus, the bill would depreciate 5.2% annually at the expense of its holder(s). (The value of and rate at which to apply the stamps could be fine-tuned if necessary.)
This system would work the opposite way ours does today, where money held over time increases in value as it gathers interest. In Gesell’s system, the stamps would be an individual cost and the revenue they created would be a public gain, reducing the amount of additional taxes a government would need to collect and enabling it to support those unable to work.
Money could be deposited in a bank, whereby it would retain its value because the bank would be responsible for the stamps. To avoid paying for the stamps, the bank would be incentivized to loan the money, passing on the holding expense to others. In Gesell’s vision, banks would loan so freely that their interest rates would eventually fall to zero, and they would collect only a small risk premium and an administration fee.
With the use of this stamp scrip currency, the full productive power of the economy would be unleashed. Capital would be accessible to everyone. A Currency Office, meanwhile, would maintain price stability by monitoring the amount of money in circulation. If prices go up, the office would destroy money. When prices fall, it would print more.
In this economy, money would circulate with all the velocity of a game of hot potato. There would be no more “unearned income” of money lenders getting rich on interest. Instead, an individual’s economic success would be tied directly to the quality of their work and the strength of their ideas. Gesell imagined this would create a Darwinian natural selection in the economy: “Free competition would favor the efficient and lead to their increased propagation.”…
…Although many dismissed Gesell as an anarchistic heretic, his ideas were embraced by major economists of the day. In his book “The General Theory of Employment, Interest and Money,” John Maynard Keynes devoted five pages to Gesell, calling him a “strange and unduly neglected prophet.” He argued the idea behind a stamp scrip was sound. “I believe that the future will learn more from the spirit of Gesell than from that of Marx,” Keynes wrote…
…That very year, the owner of a dormant coal mine near the Bavarian town of Schwanenkirchen tried in vain to get a loan from a bank to begin mining again. Stymied by the representatives of traditional finance, he went to the Wära Exchange Association, a group that was created to put Gesell’s ideas into practice. The group agreed to give the mine owner 50,000 Wära, a depreciating currency equivalent to 50,000 Reichsmarks.
The mine owner then gathered the unemployed miners and asked if they would go back to work, not for legal tender, but for this new currency. They agreed that any money was better than no money. The mine owner purchased food, clothing and household goods from warehouses that were already using the Wära currency. The miners, now back digging coal, used their wages to buy these goods from the mine owner. Soon, other businesses in town wanted to use the currency to benefit from the sudden influx of cash. Because the currency depreciated at 1% per month, everyone was eager to part with it and it circulated rapidly throughout the economy. Soon, in whole districts, the Wära currency replaced the Reichsmark, which alarmed the bigger banks and the government. Finally, the Reichsbank ended the experiment by banning the currency.
Two years later, in the Austrian town of Wörgl, Gesell’s ideas came to life again. In 1932, Wörgl’s mayor, a socialist locomotive engineer, desperately wanted to get his constituents back to work. A supporter of Gesell’s ideas, he devised a plan where Austrian schillings would be replaced with Work Certificates that depreciated at 1% per month.
The mayor hired townspeople, paid in Work Certificates, to improve roads, install streetlights and build a concrete bridge. Work Certificates circulated rapidly from merchants to tenants, to landlords, to saving accounts. People paid their taxes early to avoid paying for stamps. In one year, the Work Certificates traded hands 463 times, creating goods and services worth almost 15 million schillings. By contrast, the ordinary schilling was exchanged only 21 times.
The experiment was called the Miracle of Wörgl. Vienna newspapers took notice. The government of France expressed interest. Two hundred mayors in Austria devised similar programs in their communities. Again, however, the financial authorities grew uneasy, arguing that these local stamp scrips undermined the currency-issuing power of the national bank. By the fall of 1933, the Austrian Supreme Court had prohibited their circulation.
Gesellian experiments happened in the U.S. and Canada too, inspired by the Great Depression. In 1932, in Hawarden, Iowa, a limited amount of stamp scrip was put into circulation to pay for public works. The same year, a similar program was deployed in Anaheim, California. In 1933, Oregon attempted to print $80 million in stamp scrip, but the U.S. Treasury stopped it. The government of Premier William “Bible Bill” Aberhart in Alberta, Canada, introduced depreciating “prosperity certificates” (which people quickly renamed “velocity dollars”) in 1936.
That decade in the U.S., 37 cities, eight counties and some business groups attempted to issue almost 100 different types of stamp scrip. All these experiments were local, small in scope and short-lived. In 1933, the economist Irving Fisher, who called himself “a humble student of Silvio Gesell,” tried to persuade President Franklin Delano Roosevelt to adopt a national stamp scrip, and even convinced an Alabama senator to introduce a bill that would have issued up to $1 billion in depreciating currency. It never came to a vote. Roosevelt, who was preparing to take the country off the gold standard, worried that any further economic innovations would be too destabilizing…
…Gesell’s idea for depreciating money “runs counter to anything we’ve ever learned about the desirable properties of money,” David Andolfatto, a former senior vice president of the Federal Reserve Bank of St. Louis and the chair of the economics department at the University of Miami, told me recently. “Why on Earth would you ever want money to have that property?”
But during the economic downturn that followed the Covid pandemic, Andolfatto recognized the potential value of an expiring money in times of crisis. The relief checks that the government sent out to U.S. households didn’t immediately have their desired effect of stimulating the economy because many people saved the money rather than spend it. This is the paradox of thrift, Andolfatto explained. What’s good for the individual is bad for the whole.
“Well, what if we gave them the money with a time fuse?” Andolfatto remembers wondering. “You’re giving them the money and saying look, if you don’t spend it in a period of time, it’s going to evaporate.”
In a paper he wrote for the Fed in 2020, Andolfatto called this concept “hot money credits.” He pointed out that when the economy goes into a funk, there is a “coordination failure” where people stop spending and others stop earning. Withholding money in times of fear creates a self-fulfilling prophecy by further stifling the economy. So, could Gesell’s idea of expiring money be the cure?
“The desirability depends on the diagnosis,” Andolfatto told me. “It’s like a doctor administering a drug to a healthy person and a sick person. You administer the drug, and it has some side effects. If the person is healthy, you’re not going to make them any better. You might make them even worse. If they’re sick, it might make them better.”
The problem, Andolfatto said, is that issuing pandemic checks with an expiration date would hurt those with little savings. People with money in the bank would use their expiring money just like normal money. People with no savings, on the other hand, might find that expiring money forced them to spend and did little to stabilize their financial situations…
…Keynes believed Gesell’s expiring money amounted to “half a theory” — it failed, Keynes argued, to account for people’s preference for liquid assets, of which money is just one example. “Money as a medium of exchange has to also be a store of value,” Willem Buiter, a former global chief economist at Citigroup, told me. In a Gesellian economy, he continued, the affluent would simply store their wealth in another form — gold bars, perhaps, or boats — which could be converted into money when they wanted to transact.
Buiter doesn’t believe Gesellian money can really address serious social inequality, but he did note times when it was advantageous for a central bank to drop interest rates below zero, like when inflation and market interest rates are low and should go lower to maintain full employment and utilization of resources. Positive or negative interest rates could easily be applied to digital money in a cashless economy, for which Buiter and others have advocated. But it’s hard to imagine how a government today could practically implement a Gesellian tax on hard currency. “You’d have to be able to go out and confiscate money if it’s not stamped,” Buiter said. “It would be rather brutal.”
5. Intel’s One True Stakeholder is Here – Doug O’Laughlin
There is a rumor that the Trump administration could be taking a stake in Intel…
…And it’s no surprise that the future of American semiconductors has Intel written all over it. But there’s no other way than forward, and I think it’s time to consider what needs to happen realistically, and that’s the death of the Intel we once knew to make room for what’s next. The key is that while CPUs don’t matter, the only American leading-edge foundry left making them is critical.
The problem is that the company that funds it might run out of money, and that’s why they need to publicly threaten to stop financing the future of the foundry, because it’s a problem they can’t do alone. That is why I believe they so publicly announced the ending of future nodes past 14A…
…The calculus for America is pretty simple. In my view, there is very little strategic importance to the Intel CPU business. The x86 ecosystem was once the most incredible compute ecosystem, but AMD designs better chips than Intel could; Intel has the one thing that AMD does not, a Fab. The fabless business at Intel has a real issue in that making a CPU is becoming a relatively commoditized business. ARM has made it possible for almost any hyperscaler to have its ARM-based CPU, while AMD continues to outdesign Intel at its core job, and that’s not even discussing the longer-term RISC-V ecosystem.
Adding up the CPU side, I see a business with massive competition and Intel not at the top of the stack. Intel has to deal with increasing competition in’s core profit center while at the same time covering the increasingly heavy burden of a leading-edge fab. There is only one leading-edge foundry (TSMC), and a second American option is the single highest value-added project of all time…
…We cannot rely on Taiwan for the future of semiconductors. The more capacity we get from TSMC, the more we remain reliant on R&D in Taiwan rather than the US. Intel must be standalone and must have the capabilities to do the two things the US critically needs. High-end logic and military capabilities. I’d argue the second is met chiefly, but the first Intel is hopelessly behind.
What’s worse is that Intel has a bad customer, itself. Intel needs a good customer to be the anchor, and sadly, the core customer is a CPU company that is struggling to find its way in an accelerated compute world…
…Trump can bully Broadcom, Nvidia, Qualcomm, Apple, and AMD to put orders towards Intel, while possibly forcing Amazon, Microsoft, Google, and others to make a large investment in the fab itself (or push orders). Additionally, forcing semicap companies like KLAC, Applied Materials, and Lam Research to invest and give resources in exchange for approved licenses is another example of a carrot and a stick. I think Trump could forge the giant partnership to happen, but then execution is all up to Intel. And LBT is still once again qualified for the job.
Disclaimer: None of the information or analysis presented is intended to form the basis for any offer or recommendation. We currently have a vested interest in Alphabet (parent of Google and Waymo), Amazon, Apple, Microsoft, Tesla, and TSMC. Holdings are subject to change at any time.
