Principle
Valuation is a disruption-risk forecast, and AI has massively raised the risk
The central task of long-term investing is assessing disruption risk, and AI has made that the hardest and most important variable.
He ties it to Buffett, who could not perceive internet-era disruption (Wikipedia vs World Book). With AI the problem is more complicated, and he expects some of the most dominant-seeming businesses to simply disappear.
Before any long-duration bet, write down explicitly how AI could make this business obsolete and how wide the moat really is.
Principle
Design the vehicle so you are never forced to make decisions by circumstances
Permanent capital converts market panics from existential threats into buying opportunities.
Pershing Square Holdings, Howard Hughes and Pershing Square USA each have anchor employee stakes and are permanent-capital vehicles, enabling buying during the COVID crash or the financial crisis.
Match the duration of your capital to the duration of your strategy; eliminate redemption or refinancing triggers that could force action at the worst moment.
Principle
Define success as the greatest beneficent impact on the largest number of people
A clear impact-maximizing definition of success guides where to deploy wealth and time in later life.
He hopes AI and better health extend his life; the institute is the vehicle he expects to have huge impact.
Write a one-line definition of success that is measurable in reach; use it to decide between competing commitments.
Principle
A leader working for modest pay sets the compensation ethos for everyone
Owning a large stake and forgoing fees can be worth more than the fees because of compounding and culture.
Buffett ran a partnership taking 25 percent over 6 percent, but tired of investors moving money and lacking permanent capital.
If you own a large stake, consider whether cutting your own fee buys more value via culture and alignment than it costs.
Principle
Unsexy structures are copy-proof
The Berkshire model is simple but rarely replicated because it is unglamorous and requires rare control positions.
He calls the Howard Hughes position a creature of history: it is difficult to get to a 47 percent stake of a public company.
Look for durable strategies others avoid for status reasons; the lack of imitation is part of the moat.
Principle
Long-tenured teams make candor verifiable
Low turnover is an investment-process asset because it lets the decision-maker take team input at face value.
In an industry where top people get 100 million dollar bonuses to move, Pershing has essentially no turnover, a culture of transparency and candidness.
In judgment-heavy teams, treat retention as a decision-quality lever, not just an HR metric.
Principle
Pay no one on individual positions so the whole team is aligned on the portfolio
Compensation tied to the whole portfolio, plus wide equity ownership, produces candid investment debate.
Performance drives the management company because compounding of underlying assets grows the fee stream. He cites alignment, the right people and principles on a stone tablet as why Pershing runs itself.
In shared-judgment teams, reward collective outcomes and distribute equity widely rather than paying for individual wins.
Principle
When everyone has the same AI, the edge is doing what has not been done before
Shared AI tooling commoditizes analysis; outsized returns come from unprecedented, outside-the-box positions.
Examples cited: buying credit default swaps before the pandemic, buying the stock of a bankrupt company, shorting the credit of a AAA company before the financial crisis. AI would not have told you to do any of those things.
Use AI to accelerate research, but locate your edge in judgments that have no precedent in the training data.
Principle
AI efficiency gains accrue only to businesses with pricing power
AI-driven cost reduction is not automatically a profit gain; who keeps it depends on pricing power.
His example: AI that rewrites legacy COBOL at big financial institutions in days instead of months. Everyone will be forced to become more AI native, so the question becomes whether savings stick.
When modelling AI efficiency upside, separate cost savings from retained margin and haircut savings in commodity industries.
Principle
Underwrite every position as if the market will close for ten years
Treat liquid investments with the discipline of an illiquid ten-year commitment.
He links this to the present-value view: the early years of a business's life matter more, but you need high confidence about 10, 20, 30 years out to make an illiquid long-term investment.
Ask of any holding: would I be comfortable owning this if I could not sell for 10 years? If not, it is a trade, not an investment.
Principle
Public-market quality means the business survives a CEO change; private quality means it does not
The degree of CEO dependence is a defining difference between public and private investing.
Public markets: super durable growth companies, dominant, highly profitable, strong balance sheets, pricing power. Venture: pre-revenue or loss-making companies where you bet they become cash generators.
Calibrate how much you underwrite management by stage: business durability for mature companies, founder for early ones.
Principle
In venture, bet on the person over the idea because the original idea usually changes
Early-stage investing should be underwritten on founder quality, not plan quality.
He backed Bom Kim despite thinking Groupon was a bad model, because Kim made a powerful case for South Korea and Ackman liked him. The final business had nothing to do with the original plan.
In early-stage deals, explicitly score the founder's adaptability separately from the business plan and weight it higher.
Principle
Short selling is asymmetry in reverse
Prefer instruments where the payoff is large relative to capital at risk; plain shorts structurally invert that.
The bond-insurer bet was primarily CDS. Aversion to short selling was added to the investment principles after the painful short experience.
If you have a negative view, express it through an instrument with capped loss and large payoff, not an unhedged short.
Principle
Reputation and track record turn activism into dialogue
Activism is a substitute for credibility; once you have credibility you rarely need it.
He notes three of five or six recent new positions produced letters from CEOs inviting engagement, and argues you do not run a proxy contest against someone with 3 million Twitter followers.
Invest in a long-term reputation early; it lowers the cost of every future negotiation.
Principle
Bubble discipline is a function of time horizon
Avoiding bubble FOMO requires accepting short-term underperformance for long-term discipline.
He defines a bubble as human nature: a lot of money being made, FOMO, more money into the same trade until overvaluation, then turbulence and a burst.
Decide your horizon before the mania; if it is long, accept looking foolish for a while.
Principle
Delegate idea generation and operations so the principal can do strategy
A founder can run multiple public entities by owning final decisions and strategy rather than ideation and operations.
A partner runs the business elements, boards of the funds are independent, and he treats new ventures like portfolio companies: install the CEO and then stay out of day-to-day operations.
Move yourself from generator to final arbiter once the team is trusted and principles are codified.
Principle
Every hard thing produces something good if you act on it
Near-death experiences, survived and learned from, leave operators materially better off.
He applies this to his daughter's recovery and to the brain institute as the good coming from the crisis.
After a setback, deliberately codify what changes; the improvement is the payoff for the pain.
Principle
You do not need to make the money back the same way you lost it
After a broken thesis, redeploy into the highest-certainty alternative rather than holding to recoup.
He notes the tax loss has some value to Pershing and the capital went into something with higher certainty.
When exiting a loss, judge the replacement on its own merits; never hold a position just to get back to even.