VALUExBRK 2026 | Bill Ackman, Mohnish Pabrai, Tom Gayner, Chris Bloomstran & More

ELI5/TLDR

This is the annual value-investing conference Guy Spier throws in Omaha the day before the Berkshire Hathaway meeting. A parade of famous fund managers and business people each gets fifteen minutes on stage, then takes questions. The two threads running through the whole day: how do you build something that compounds for decades, and what does AI actually do to investing. Spier himself has a brain tumour, which gives the whole thing an unusually warm, mortality-aware tone — everyone keeps circling back to the idea that the point isn’t getting rich, it’s stewardship and good relationships.

The Full Story

Steven Myhill-Jones — How Charlie Munger hand-picked him to run the Daily Journal

The Daily Journal is a tiny NASDAQ company that has published legal newspapers in California since 1888. The newspaper part is dying; 80% of revenue now comes from Journal Technologies, which builds software to run courts. And during the 2008 crisis it accidentally made $28 million publishing foreclosure notices, which Munger then invested into stocks that grew to several hundred million — so a big chunk of the company’s value is just a stock portfolio Munger built.

Myhill-Jones, a Canadian software founder, kept refusing to even meet Munger because he “never wanted to run a public company.” He flew down anyway, sent ahead a seven-page memo arguing why he was unsuitable, and a few days later got a call: “Stephen, I’d like for you to become CEO. And before you say no, I’m almost 98 and it’s probably time for me to retire. So I’d also like you to succeed me as chairman.” He nearly drove off the road.

His real point was about incentives. He agrees with Munger’s “show me the incentive and I’ll show you the outcome,” but thinks people read it too narrowly as money.

“The goal isn’t to hope that people’s values override their incentives. [It’s] to build organizations where incentives and intrinsic motivation are aligned.”

On AI: he runs a hard business — selling complex software to courts with long procurement cycles and heavy regulation. Things that once felt like burdens now look like moats, because nobody is going to replace a court system with “a few weekends of Claude Code.”

Adam Mead — Berkshire by the numbers

Mead wrote the definitive financial history of Berkshire (Buffett bought copies for the directors). His talk was a slideshow of jaw-dropping statistics. In the last five years Berkshire added more equity capital than in its first 52 years — that’s compounding. Operating earnings drove 58% of all capital growth. Insurance float ended last year at $173.5 billion, and Berkshire has had an underwriting loss in only a handful of years since 1967 — meaning it has mostly been paid to hold other people’s money.

A few sharp facts: Buffett paid himself $100,000 a year and took a real-terms pay cut every year. In Berkshire’s whole history, no acquisition was ever less than 15% of equity capital — so a 15% deal today would be a $110 billion check. The theme across six decades is “a willingness to bet big when it makes sense” — whole companies, huge stock positions, even bonds (he made 16% buying distressed Washington Public Power bonds).

Brian Lawrence — Two AI agents named Marvin and Vicki

The most interesting talk of the day. Lawrence (Oakcliff Capital) built two AI agents running on Mac minis using “open Claude.” Vicki is a shopping bot, deployed to stress-test investments. The lesson: the internet is now a spy-versus-spy war zone. Zillow, Compass, OpenTable all block her; she had to rent a $5/month residential proxy to “pretend she’s a lady from Queens.” Useful intelligence about how little autonomous web agents can actually do right now.

Marvin is a research analyst that applies the firm’s investing frameworks, monitors the portfolio, and has a “compounding memory.” The crucial insight came from Marvin himself, commenting on his own limits:

“My memory compounds correct and incorrect beliefs equally. Human discussion is the quality filter. Without discussion with you, I drift towards confident mediocrity.”

That phrase — confident mediocrity — is the warning. The agent will give you a beautiful, fluent answer that is wrong, and over time bad inputs stack into bad conclusions. Lawrence’s frame: a Google search optimizes for engagement; what he wants Marvin to optimize for is what is “knowable, important, and true.” His predictions: knowledge workers will each have their own agents, agents will be worth 100x their token cost, but they’ll also “marinate you in your own biases,” and people will over-rely on them and get dumber.

His business filter for AI risk: a company is safe if it sells something denominated in human time but its costs are not tokens. Dangerous if it sells human time/attention while its cost of goods is “increasingly denominated in tokens that are going down 99% a year.”

Bill Ackman — Permanent capital, Israel, and creativity AI can’t copy

Ackman took questions only, no slides. On AI in markets: it’ll make the fast-money crowd (Citadel, Millennium) even faster and more short-term — an arms race to read a 10-K and trade in milliseconds. But that helps patient investors, because great companies will get dumped at silly prices on short-term noise. And AI can’t be creative: it only knows the past.

“Many of our most successful investments… were things that had never happened before.”

On Israel (he owns 5% of the Tel Aviv Stock Exchange with his wife): think of Israel as “Silicon Valley if it were a country,” and owning the exchange is “owning a royalty on the success of capitalism in a country.” He likes Meta (“extremely cheap” — investors are nervous Zuckerberg goes “off on a journey,” but his long-term track record is rational).

The meatiest section was on permanent capital. The reason isn’t “money can’t leave” — it’s that his strategy requires it. After the Valeant disaster (down 90%, his “worst mistake”) and the Herbalife short (“when they make a movie about your short position, that’s when you get out”), he handed back all open-ended money eight years ago and has had his best run since. He explained discovering the 40-Act closed-end fund structure — one layer of tax, can hold control stakes and minority public stakes and derivatives — and how he’s rebuilding Howard Hughes into a mini-Berkshire: an undervalued asset-rich company, buy an insurer (Vantage), put float into treasuries, invest the surplus in stocks. He owns 47%, “about what Buffett had.” Would he be Treasury Secretary? “No.”

Tom Gayner (Markel) — Choosing a race you can win

Gayner runs Markel, a Berkshire-style insurance compounder ($8/share at 1986 IPO, ~$1,800 now). His central metaphor: he’d lose a 100m dash to Usain Bolt, and a mile, and probably a marathon — but a foot race from Key West to Seattle?

“That’s a bet. Because it is no longer about pure athleticism… It’s about will. It’s about determination. It’s about the proper pacing and stick-tuitiveness that can last for decades. That’s my edge.”

That’s the case for permanent capital and long time horizons against faster, smarter, better-funded competitors. He addressed the Jana Partners activist campaign diplomatically — agreed Markel’s insurance underperformed and they’ve changed management, but disagreed that the diversified “Markel Ventures” businesses are a distraction; they provide cash when insurance is weak. His governance answer borrowed a Scottish colleague’s pledge of “hyper honesty and hyper communication” — control the process, not the outcome. On his nightstand: War and Peace (he loved how Kutuzov beat Napoleon by strategic retreat — letting the enemy punch himself out) and Wright Thompson’s The Cost of These Dreams.

Eric Markel (Nightview Capital) — Stewardship and the 1,300-year-old hotel

A genuinely moving talk about Hoshi Ryokan, a Japanese inn founded in 718 AD, still run by the same family — now the 47th generation, the first woman to lead it in 1,300 years. He’s writing a book (Outlast) about the world’s oldest businesses. The 47th-generation owner doesn’t consider herself an owner at all — she’s a “caretaker,” a gardener. Her only goal: “to hand the business to her nephew in better condition than she inherited it.”

“What if instead of calling ourselves CEOs, chief investment officers, research analysts, we simply called ourselves caretakers of the businesses we steward? How would our actions change?”

He tied it to tikkun olam (Hebrew, “repair a broken world”) and a Greek proverb: “a society grows great when old men plant trees in whose shade they shall never sit.” His punchline: longevity is the wrong goal. “Longevity is what happens when you stop thinking about longevity altogether” — it’s a byproduct of caring intensely about quality right now.

Gizela Bauer — 30 years of knowing Warren Buffett

A German financial journalist who, in the pre-internet 1990s, decided Germany “needed a role model like Warren Buffett,” wrote him an enormous letter, and got a fax back inviting her to the meeting. She’s now friends with him for life and made a film with him. Her lessons: his memory was a “computer in a time where nobody had one”; he builds a “precious collection of people,” deciding in one second who he lets into his orbit; and he cuts complexity with simple sentences — “America’s best days lie ahead of us” — “a big scissor for all the micro-prediction worries.” The trick to judging people, which she learned naturally as a journalist: be curious, ask questions, don’t talk about yourself. Did Buffett become a hero in Germany? Yes — so much so he once joked, “if you bring more German shareholders asking questions, I don’t want to see you here again.”

Luca Dellanna — Incentives are weak, habits are strong

A crisp, contrarian talk. Incentives matter, but people overestimate them. If you find a wallet, whether you return it doesn’t actually depend on how much cash is inside — culture and habit decide. A sales bonus motivates people who are already good at selling; it can’t create skill. The distinction: incentives shape one-off decisions, but repeated behavior is governed by habits (“if intentions were enough, everyone would go to the gym”). How do you build habits faster than slow repetition? Two levers: competence (you won’t form a gym habit if lifting feels awkward and ineffective) and immediate feedback (he cited Rory Sutherland’s point that toothpaste is minty because without the mint-tingle you’d get no feedback and quit). DuPont’s safety culture was his case study — and a costly signal seals belief: his boss’s boss once sent a whole chartered bus back to the depot, stranding 17 executives for 40 minutes, because it had no seatbelts.

Tom Russo — Brands, and why the “quality of thought” matters

Mostly a warm 41-year reminiscence with Spier, centred on Weetabix — the inedible British cereal Russo invested in when it was “a mountain of cash at four times earnings.” On the current slump across branded goods (luxury, spirits, food): “it’s all of the above” — pandemic working-capital damage, Nestlé losing its self-critical culture, inflation. His enduring insight on luxury: it sells “a badge that allows them to tell others who they are.” And: “the richer you get, the fewer opportunities you have to distinguish yourself as special” — so it comes down to the quality of your thinking.

Chris Bloomstran — Berkshire as the Russian army in retreat

Famous for letters longer than Berkshire’s entire annual report. He opened with a stat: Berkshire’s stock could fall 99.3% and still have beaten the S&P since 1965. (Munger’s reaction when told: “that’s just nothing but simple compound interest. There’s nothing impressive about that.”) $100 in Berkshire in 1965 became $6.1 million at 19.7%; the S&P turned it into $45,000.

His real argument, riffing on Gayner’s War and Peace theme: Berkshire’s genius is retreating — pivoting when others charge. Buffett closed the textile mill, wound up his partnership in 1969 when ideas dried up, and cut overpriced Coke from 30% to 14% of assets via the GenRe deal. Right now Berkshire is doing it again in insurance: deliberately writing less premium because there’s far too much capital in the industry. GEICO’s combined ratio “worsening” from 81% to 85% will be read as bad news, but it’s healthy — they spent an extra billion on advertising and grew policies 5%. Reinsurance is writing a dollar of premium for every dollar of capital (“crazy — should be 50-60 cents”). Berkshire has pulled $100 billion upstream out of insurance reserves. The crowd will read shrinking underwriting profit as failure; Bloomstran reads it as “the Russian army retreating in a period where everybody else is punching” — and if the stock drops, Berkshire just buys back more.

On the $370 billion cash pile: not a market-timing call, just a small opportunity set. Buffett trimmed Apple 80% partly because the 21% tax rate was a gift. Greg Abel’s real job is to “spend money and spend it in earnest” in the next crisis. On macro: total credit market debt at 350% of GDP terrifies him, but “it’s hard to let the macro enter your thinking” beyond keeping balance-sheet integrity high.

Lauren Templeton — Fairfax as “Berkshire 30 years ago”

A Fairfax board director (and Sir John Templeton’s great-niece) made the explicit pitch: Fairfax has the same DNA as Berkshire — insurance float, value investing, founder control (Prem Watsa), decentralized — but at ~$40 billion market cap it’s “small enough to move, large enough to matter.” Since inception (1985) book value has compounded 18.7% a year; the stock ranks 7th of all US-listed companies since 1985 (ahead of Berkshire at 49th). She walked through where Fairfax sits today versus the year Berkshire hit each milestone — and it lines up to roughly 1996–2002 Berkshire. She declined detailed questions (earnings released that morning) and closed with a lovely story: she once asked Buffett to phone her frugal value-investor father for his birthday; he did, twice (Dad hung up the first time thinking it was his brother), refused any donation in return — “that would cheapen what I did.”

Brandon (New Money) — Why YouTube is a great business

A former physiotherapist who built the world’s biggest value-investing YouTube channel. Useful facts: YouTube takes more US TV-watch-time than Netflix (12.7% vs 8.4%); 2.7 billion monthly users vs Netflix’s ~325 million subscribers; 500 hours uploaded per minute across 114 million channels. The model that makes it special: anyone can make anything, so every niche (slime-squishing, sardine reviews) finds an audience, and YouTube pays creators 55% of ad revenue — a self-reinforcing flywheel no other platform replicates. His channel did $1.4 million in pure AdSense. His advice to the room: put organic (not paid) content at the core of your fund/business, reverse-engineer the questions your ideal client is searching, and “have the courage to suck” — bad videos just don’t get clicked, so there’s no downside to starting. Shorts? “They suck” for monetization; the money is in long-form.

Mohnish Pabrai — The closer, and a Munger story

Pabrai took only questions. On AI and international investing: very early days; as long as humans are in markets they’ll swing between fear and greed, which is all the opportunity an investor needs. He’s optimistic AI is an equalizer, not a divider — and told a beautiful story of two sisters in a remote North Indian village who learned tailoring (and how to fix their own sewing machine) entirely from YouTube, pulling their family up after the father was paralyzed. Jugaad — making do — as a force more powerful than government or infrastructure.

On Constellation Software (he recently bought): the SaaS selloff over AI fears is overdone; incumbents get a tailwind, and Mark Leonard’s acquisition engine (touching ~100,000 vertical software companies twice a year by phone and twice by email, à la Buffett’s “if you ever decide to sell, think of us”) is uncloneable because “humans are very poor cloners.” On currency debasement and gold: don’t waste brain cells. A can of Coke has value regardless of currency because people will trade their labour for it; he holds 70% of his portfolio in Turkey, where the currency fell 90% and his dollar returns were still spectacular. On the lone investor vs the committee: “we don’t know too many committees who won Nobel prizes” — and small managers keep the advantage of fishing in tiny opportunities the big funds can’t touch.

He closed with an unprintable-but-printed Munger story about a 1960s distressed-debt deal involving a maverick entrepreneur who built sales by shooting holes in car engines, his grieving widow, and a very well-endowed blonde nurse-mistress — ending with “Mr. Dignified Munger’s” line about the lunch at the blue-blooded California Club. The room lost it.

Key Takeaways

Myhill-Jones: Intrinsic motivation, properly aligned with incentives, beats relying on values to override pay. Complex public-sector software (courts) is more defensible in the AI era, not less.
Mead: Berkshire added more equity in the last 5 years than its first 52 — pure compounding. Operating earnings = 58% of all capital growth. No Berkshire acquisition was ever under 15% of equity. Buffett paid himself $100k and took a real pay cut yearly.
Lawrence: AI agents “compound correct and incorrect beliefs equally” — human discussion is the only quality filter against “confident mediocrity.” Safe businesses sell human time but don’t pay costs in tokens; endangered ones do the reverse.
Ackman: Permanent capital is a strategic necessity, not a vanity. Rebuilding Howard Hughes into a mini-Berkshire (owns 47%, buying insurer Vantage). Likes Meta (“extremely cheap”) and the Tel Aviv Stock Exchange (“a royalty on capitalism in a country”). AI can’t be creative — it only knows the past.
Gayner: Pick a race long enough that will and pacing beat raw talent — “Key West to Seattle.” Control the process (hyper-honesty), not the outcome.
Eric Markel: Stewardship over ownership; “caretaker, not owner.” Longevity is a byproduct of caring about quality now, never the goal itself. Hoshi Ryokan, founded 718 AD, 47 generations.
Dellanna: Incentives shape one-off decisions; habits govern repeated behaviour. Build habits via competence + immediate feedback. Costly signals (the seatbelt bus) cement belief.
Russo: The branded-goods slump is cyclical and structural (pandemic, inflation, lost culture at Nestlé). Luxury sells identity/tribe badges. “The richer you get, the fewer ways you have to distinguish yourself.”
Bloomstran: Berkshire’s edge is disciplined retreat — shrinking insurance premium when capital floods in, like Kutuzov letting Napoleon exhaust himself. GEICO’s “worse” 85% combined ratio is actually healthy. The $370B cash isn’t market-timing; Abel’s job is to deploy it hard in the next crisis.
Templeton: Fairfax ≈ “Berkshire 30 years ago” — same model, 1/25th the size, can still move the needle. Compounded book value 18.7%/yr since 1985; ranks 7th of all US-listed companies (Berkshire 49th).
Brandon: YouTube beats Netflix on US TV-watch-time; 55% creator revenue share drives an uncopiable niche flywheel. Money is in long-form, not shorts. “Have the courage to suck.”
Pabrai: Constellation Software’s selloff is overdone; Mark Leonard’s deal engine is uncloneable. Currency debasement is a waste of worry — great businesses transcend it (70% in Turkey, currency down 90%, returns still great). Lone investors keep an edge in small opportunities. Explain a thesis to a 10-year-old in four sentences or pass.

Claude’s Take

This is a genuinely good use of five hours, but you don’t need all five. The signal is concentrated in maybe four talks: Lawrence on AI agents (the freshest, most concrete thinking in the room — “confident mediocrity” is the line that’ll stick), Bloomstran on Berkshire’s insurance retreat (a real, contrarian read on numbers everyone else will misinterpret), Ackman on permanent capital (a clear-eyed account of why the structure matters, not just that it’s nice), and Pabrai for the texture. Eric Markel’s stewardship talk is the emotional core and worth it on its own terms even though it offers no investable idea.

The BS-filter notes: there’s a lot of Berkshire hagiography and mutual back-slapping — half the talks are “things I learned from Warren.” Templeton’s Fairfax pitch is exactly that, a pitch, from a director who conveniently couldn’t take questions; treat the “Berkshire 30 years ago” framing as marketing, even if the numbers are real. Brandon’s YouTube talk is partly an ad for his book and agency. And the whole event runs warmer than usual because of Spier’s illness, which is moving but also means the tone occasionally drifts from analysis into eulogy.

Score: 8. It loses points for length and repetition, but the density of genuinely-original operators saying specific, checkable things — Lawrence’s agent architecture, Bloomstran’s reinsurance-capital math, Ackman’s 40-Act mechanics — is high enough that an 8 is fair. This is the rare conference where the famous names actually said something instead of recycling platitudes.