What I learned this Week: July 13, 2025
When something is important enough, you do it even if the odds are not in your favor
Lessons of the Week
Bristol, our 7 month old puppy German Shepherd, is eager to play, so I have little to say as an intro in my lessons of the week. There is a lot in this issue. The related articles are insightful in the context of those trying to understand the massive wave of disruption building across AI, investing, and chronic disease. I blame all typos on Bristol’s priorities to go play with a tennis ball.
What I learned this week….
In 1774, firewood output accounted for 28 percent of US GDP. (Link)
Innovation outpaces institutions, and our regulatory system isn't built for exponential response. AI benefits will be slowed by regulatory infrastructure.
Every mile driven in a Tesla, every post on X, every prompt in Grok, every human experience captured by Neuralink, and every country gaining internet access with Starlink diversifies and trains the autonomy engine. And dismantles outdated structures.
Recorded music wasn’t invented by Thomas Edison until the 1870s. Before then, the other option for music at home was piano lessons.
Nearly 46% of all U.S. trips are under 4 miles; less than 1% are by plane or rail. California’s high-speed rail is estimated to cost $100B+, connecting the 520 miles between San Francisco and LA.
The latest measles, mumps, and rubella (MMR) vaccine rate (2023) was over 91.3%, down from 92.4% in 2020. In 1992, the MMR vaccination rate was approximately 66%. With an Ro of 12+, the target vaccination rate is 95% to avoid an epidemic, though rates as low as 80% may suffice. MMR was introduced in 1971.
Approximately 1 in 6 acres of global cropland (14-17%) is contaminated with toxic metals above safety thresholds. In California's Salinas Valley, where soils naturally contain elevated cadmium, some spinach samples have tested at levels exceeding California's daily safety limit by up to 25 times per serving.
Price Discovery: Human IVF donor eggs typically cost between $2,795 and $2,995 per cycle, donor sperm generally costs between $500 and $1,000 per vial. Registered Wagyu semen: $150 to $500+
When something is important enough, you do it even if the odds are not in your favor.
Elon Musk
Some Data and Links
→ The Case for E-Bikes Gary Winslett’s Substack, “The Normie Case for E-Bikes”, puts numbers behind a quiet revolution: nearly half of all trips in America are under 4 miles. Yet we build infrastructure for speed and distance, not practicality. An e-bike replaces a second car for short trips, at a cost of ~$800/year versus $6,000–$12,000 for a car. Protected bike infrastructure costs ~$1–2M per mile, compared to $70M/mile for urban highway or $300M/mile for light rail. We don’t need high-speed rail to transform transportation. We need four-mile solutions. If even a fraction of American households replaced a second car with an e-bike, we could redirect $5,000/year into better food, family health, or retirement.
→ The spatiotemporal distribution of human pathogens in ancient Eurasia: (link) A study of ancient DNA reveals that many of the most dangerous human pathogens, including plague, did not emerge with early farmers as previously thought, but rather arose later with nomadic pastoralist groups around 6,000 years ago. As these Eurasian steppe nomads began raising large herds of livestock, zoonotic pathogens such as Yersinia pestis (plague) began to spread widely. The resulting epidemics were intense enough to alter the genetic makeup of these populations, favoring immune-related mutations that, while protective, may have also increased susceptibility to autoimmune diseases, such as multiple sclerosis. This immunological advantage may have contributed to the nomads' westward expansion into Europe around 4,500 years ago, where indigenous populations had limited resistance to the new pathogens, potentially facilitating large-scale population replacement.
Massive Pathogen Screening: Researchers screened DNA from 1,313 ancient human remains spanning 37,000 years, identifying 5,486 microbial DNA hits from 492 species, including 3,384 known human pathogens.
Zoonotic Disease Emergence: Zoonotic pathogens were first detected around 6,500 years ago and reached their peak around 5,000 years ago, coinciding with the domestication of livestock.
Steppe Migration and Disease Spread: Steppe pastoralists (Basically nomads) not only spread cultural and genetic traits into Europe but may have brought infectious diseases that devastated local populations with little immunity.
Yersinia pestis (Plague) Timeline Extended: Plague DNA was found as early as 5,700 years ago across Russia, Central Asia, and Siberia, pushing back the timeline and range of early plague cases. This challenges earlier beliefs that plague was isolated to later historic outbreaks.
Coinfections Were Common: Some individuals were found to carry multiple pathogens, including combinations like hepatitis B virus with plague or relapsing fever, revealing a high burden of disease in ancient populations.
Pathogen Evolution Evidence: The study shows that ancient microbial genomes underwent changes under immune pressure, and some mutations that aided pathogen resistance may have increased the risk for autoimmune conditions.
Disease Surveillance Potential: The paper demonstrates how large-scale ancient DNA analysis can provide new insights into epidemiological transitions, pathogen-host coevolution, and past human migrations. It lends credence for NIH concentrating some work on large longitudinal databses of human DNA, proteomics, and other data.
→ 3D-Printed Pancreatic Cells: Wake Forest University successfully 3D-printed functioning human pancreatic cells that respond to glucose and produce insulin. This bioengineering breakthrough could revolutionize the treatment of Type 1 diabetes. 3D bioprinting works by using a special printer to deposit living cells in precise layers, along with a gel-like material that mimics human tissue. The printed structure is then matured in a controlled environment until the cells grow together and function like real tissue. (link)
→ Grok 4’s debut marks a fork in AI strategy: OpenAI keeps refining a browser-based LLM that serves the crowd, while Musk is wiring a reasoning layer straight into Tesla dashboards, Starlink antennas, Optimus joints, and soon Neuralink synapses. That edge-first network has fewer users, but each sits on a torrent of real-time, context-rich data that Hayek would call the “local knowledge” on which markets run. Grok doesn’t need perfect prose; it needs actionable signal, and the messy telemetry streaming from 300 million rolling sensors and low-orbit relays may beat polished chat history every time. If Claude Shannon was right that information is surprise, Musk’s vertically integrated stack could generate the rarest surprises of all, those that let machines act autonomously in the physical world, turning the LLM itself into a commodity and the network into the moat.
→ Clearing Brain Sugar Stores Could Protect Against Dementia: Alzheimer’s may be linked to how brain cells manage sugar. Normally, our brain cells keep a small reserve of stored sugar (called glycogen) that they can tap into when stressed. This sugar isn’t burned for energy. Instead, it’s rerouted into a protective system that helps detoxify the cell and defend it from damage, kind of like using a rainy-day fund to fix a leak in the roof.
But in Alzheimer’s, researchers have found that this backup system breaks down. The tau protein, known for forming damaging tangles in the brain, also traps glycogen and blocks the brain’s ability to access this sugar reserve. Without it, neurons struggle to protect themselves from oxidative stress, a key factor in aging and brain degeneration.
Imagine your cells are like tiny factories. As they produce energy, they also release some “exhaust”, unstable molecules called free radicals. These molecules can damage DNA, proteins, and cell membranes if they’re not cleaned up.
Antioxidants neutralize this exhaust. However, when the system is out of balance (due to excessive free radical production or inadequate antioxidant defense), damage accumulates. That’s oxidative stress.
Researchers showed that stimulating the enzyme responsible for breaking down glycogen reactivated this protective pathway. In lab models, this reduced damage and improved brain health. Interestingly, dietary restriction and certain medications also appeared to enhance this process naturally. This work could explain why GLP-1 drugs, now widely used for weight loss, show promise against dementia.
In short, Alzheimer’s may not just be a disease of misfolded proteins; it could also be a disease of stuck sugar. And freeing up that sugar could unlock a powerful new path to treatment. (link)
→ The Only Winning Move Is Not to Play: In the 1983 film WarGames, a military AI simulates global thermonuclear war only to conclude that “the only winning move is not to play.” Today, large language models (LLMs) like those studied in the recent paper "Strategic Intelligence in Large Language Models" demonstrate a remarkable capacity for strategic reasoning. When tested in iterated game theory environments like the Prisoner’s Dilemma, these models not only adapted to uncertainty and retaliation but also developed distinct, persistent “strategic fingerprints”, some ruthless, some forgiving. Given these capabilities, it's plausible that such models could have been used to pre-run the 2022 invasion of Ukraine, surfacing scenarios in which even tactical victories would yield long-term strategic losses: economic isolation, NATO expansion, reputational decay, and domestic instability. A simulation might not have stopped the war, but it could have revealed, like in tic-tac-toe, that all viable paths end in a stalemate or Pyrrhic cost.
One more such victory and we are undone!
Pyrrhus of Epirus, 279 BC
Of course, LLMs are not clairvoyant. They lack access to classified intelligence, suffer from “black swan blindness,” and may mirror the biases of their training data. However, as red-team tools or strategic stress testers, they can expose gaps in policy assumptions, explore secondary effects, and generate adversary rationales at scale. Their value lies not in predicting precise outcomes, but in revealing that, in some games, like protracted wars of conquest, every move accelerates mutual loss. When strategy becomes simulation, the most rational act may be restraint. Startups would benefit from running this type of simulation to determine how they may fail. Using the results to mitigate the downside. (link)
This Weeks Charts
→ A historically high $7.4 trillion is sitting in U.S. money market funds: Ackman’s “bullish” stance likely assumes that this cash will rotate into equities or other assets, pushing prices higher, especially if inflation eases, the Fed pauses, or recession risk recedes. Very traditional view.
Dry Powder for Risk Assets: A huge cash stockpile on the sidelines can be seen as “bullish” because it suggests latent buying power. Investors may be waiting for a clearer signal to deploy capital into equities, credit, or alternatives.
Risk Aversion or Strategic Liquidity?: High balances might also indicate investors are fearful (e.g. of recession, interest rates, or geopolitics), parking funds in short-term instruments.
Higher Rates, Higher Yield: With short-term Treasury yields over 5%, money market funds have become attractive again for yield-driven reallocation.
A Setup for Volatility: Historically, when large amounts of capital sit on the sidelines, markets can whipsaw as sentiment changes. Once consensus shifts toward “risk-on,” the firehose of liquidity can fuel rallies, especially in oversold or under-owned sectors like small caps or tech.
If you're tracking deflationary tech and creative destruction themes, this could also be a setup: as innovation unlocks growth and lowers costs, the capital sitting idle today may become tomorrow's fuel for the next wave of systemic disruption. The KC Fed Reserve had a very interesting paper in 2023, linking the degree of innovation closely to interest rates. Recent high interest rates have significantly reduced venture investing, slowing almost all innovation, except AI. As the industry catches on and learns how to apply AI, and if interest rates decline, we could see new all-time highs for investment in innovation. A bit like the 2000 period. In this case, rather than a Y2K upgrade, the drive is a total factor productivity upgrade.
→ The Only Winning Move Was Not to Cut: Before coal, before oil, before electrons moved through copper wire, civilization ran on wood. A new NBER paper, "Wood and the Rise and Fall of Civilization", reminds us that timber was the core energy source of early economies. By some estimates, over 25% of U.S. GDP in the 1700s came from firewood: cutting it, hauling it, trading it, and burning it.
But wood doesn’t scale.
As populations and empires grew, forests thinned. The British Navy’s insatiable demand for tall, straight pine masts drove the Crown to reserve old-growth trees in the American colonies, branding them with the “King’s Broad Arrow” and outlawing their harvest. Colonists didn’t just resent the intrusion; they saw it as an existential chokehold on their future. One of the earliest symbols of resistance was a pine tree flag bearing the phrase: An Appeal to Heaven. It wasn’t just a theological claim, it was a declaration of resource independence. A rebellion over forest policy.
In typical British fashion, Parliament protected the mast trade. In typical American fashion, colonists chopped down the trees anyway.
Where other nations hit a “timber ceiling,” the U.S. broke through it. By the mid-1800s, the United States began shifting to coal, steam, canals, and then steel. What began as a rebellion over wood evolved into a system built for abundance. The U.S. didn’t just escape the scarcity trap. It commercialized its way out of it. That’s the essence of creative destruction. A finite resource triggers a constraint. That constraint forces innovation. The innovation builds a new foundation for growth.
But history repeats with different materials. What mast trees were to the 18th century, lithium, cobalt, and rare earths are today. We’re hitting new ceilings. The only winning move isn’t to stop growing, it’s to start reinventing what growth runs on.
And the Appeal to Heaven flag? It still flies in some corners today. Sometimes misunderstood, sometimes politicized, but initially, it was a banner for energy transition. While MAGA picked up the Appeal to Heaven flag recently, maybe the renewable industry would find the flag useful with the passage of the Big Beautiful Bill (Link)