ExpertsFMDan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460]
Invest Like the Best with Patrick O'Shaughnessy · February 24, 2026
5 Selected Lessons:
case · 2 reasoning_pattern · 1 heuristic · 1 framework · 1
Moat-durability triage for assessing AI's disruption of incumbent businesses
Source
Invest Like the Best with Patrick O'Shaughnessy — Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460] (February 24, 2026) · at 27:22 ▸
When this applies
when a powerful new general-purpose technology arrives and you must triage which incumbents' moats survive
The takeaway
Rank threatened businesses in phases (first commoditizable point-solutions, then more entrenched systems of record, then physical/robotics-protected industries) and judge each by whether switching costs exceed the ease of a cheap AI-built replacement — but assume all must still adapt, since standing pat is not a safe strategy even for currently protected incumbents.
Why it works
Systems-of-record businesses persist because switching core infrastructure (ERP/CRM) carries high operational risk/cost even when AI can build point-solution alternatives cheaply; but incumbents must still integrate AI into their own offering or lose relative position, the way e-commerce forced retailers to absorb a painful margin hit rather than ignore it.
One level up
When facing a disruptive general-purpose technology, your moat analysis must separate 'can be replicated cheaply' from 'can be safely ignored' — few businesses get to do the latter.
Where it stops holding
This stops working when the pace of AI capability improvement is slower than assumed, or when switching costs for core systems erode faster than expected (the speaker explicitly labels this view 'fairly low conviction').
Detail
procedure:
- Identify which company functions are pure digital intelligence (most exposed first)
- Identify which are systems of record with high switching costs (protected for now but must still integrate AI)
- Identify which are protected by physical/robotics barriers (protected until robotics matures)
- Assume all incumbents must evolve their business model (absorb margin hits) like Walmart did with e-commerce, rather than assuming safety
benchmarks: {"software_shorts_pre_2026":"basically none existed as a category before 2026 per the letter"}
rationale: Software was the first domino because coding tools entered mainstream awareness; the same logic should cascade to other domains as capability spreads, though the exact order and timing is uncertain.
In their words
I think that'll be the case with software, too, where companies that have really great distribution and great business models and their systems are record for companies.
Public vs private markets: same skill, different competitive structure
Source
Invest Like the Best with Patrick O'Shaughnessy — Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460] (February 24, 2026) · at 2:42 ▸
When this applies
when choosing between a highly liquid, crowded arena and a less liquid, less crowded one for allocating scarce analytical effort or capital
The takeaway
Competitiveness in a market is not just about headcount of competitors — it's about whether competitors are playing the same game as you; a crowded market where most players pursue different goals can be easier to beat than a thin market where everyone is aligned on the same objective.
Why it works
Public markets have many participants but they pursue different objectives (short-term flows, mandates, irrational behavior), creating noise to arbitrage; private markets have fewer participants but they're all solving the same long-term-value question, so skill differentiation matters more and liquidity/sentiment distortions are absent.
One level up
Edge comes from identifying where your objective function diverges from the market's dominant behavior, not merely from counting competitors.
Where it stops holding
This stops working when private market access itself becomes the binding constraint (you can't get allocation into the best private deals regardless of analytical edge), or when public market irrationality disappears and all participants become long-term fundamental investors.
Detail
approach: Classify competitors not by number but by whether their incentive structure is consistent with long-term intrinsic value investing; where it isn't, look for arbitrage opportunity.
signals:
- many participants pursuing different mandates/timeframes = crowded but exploitable
- few participants all aligned on same question = thin but hard to beat
- access/allocation becoming the scarce resource rather than analysis
rationale: Because private investors largely agree on business quality and compete mainly for allocation, while public investors disagree on objectives entirely, the nature of competition — not its size — determines where alpha persists.
In their words
I would say that one difference that equalizes it a bit is that you don't have this dynamic on the private side of people doing things that are economically irrational because they're focused on the short term, or they're just their business model is not consistent with investing based on long-term intrinsic value, where you have in the public market.
Focus vs. breadth trade-off in scaling a platform business
Source
Invest Like the Best with Patrick O'Shaughnessy — Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460] (February 24, 2026) · at 17:08 ▸
When this applies
when any company with a broadly applicable core asset/technology must decide how many end markets to pursue simultaneously
The takeaway
When a core asset is broadly applicable, the temptation to monetize every market at once is real but historically rare to pull off — err toward focus unless you have truly exceptional, overlapping talent density.
Why it works
Spreading a fixed-cost asset over more revenue streams improves amortization economics, but organizational focus (one A-team, one culture) tends to produce best-in-class execution; splitting attention risks mediocrity everywhere.
One level up
The more generically useful your core technology/asset is, the more disciplined your go-to-market scoping must be, because breadth without focus degrades execution quality across the board.
Where it stops holding
This stops working when a company genuinely has deep enough talent and capital to run multiple A-teams in parallel without diluting any one, which the speaker allows may be possible with today's AI labs' concentrated talent.
Detail
rule: Default to focus (one end market, one culture) over simultaneous multi-market pursuit.
threshold: Amazon took ~7 years post-IPO before diversifying into enterprise (AWS) — a historical data point for how long focus preceded broadening.
rationale: Culture and team orientation (consumer vs enterprise) are hard to run in parallel; Anthropic's narrow focus produced a market-leading position in coding/enterprise, while OpenAI's broad strategy remains unproven though staffed by exceptional talent.
In their words
I rarely have seen any company succeed trying to go after multiple end markets at the same time. Usually you have an A-team, that A-team is focused on one thing.
Rebuilding investor trust after the 2021 GameStop drawdown
Source
Invest Like the Best with Patrick O'Shaughnessy — Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460] (February 24, 2026) · at 33:54 ▸ · at 37:07 ▸
When this applies
when any leader faces a public failure that has damaged stakeholder trust and must decide whether to confront or avoid stakeholders
The takeaway
After a public failure, the highest-leverage near-term action is honest, direct communication of a changed approach to the people whose trust matters most — even though the trust itself will only be rebuilt slowly through results, not through the communication itself.
Why it works
Negative narratives about a person/firm cannot be disproven quickly with words; they are only reversed by sustained, visible, methodical performance over a long period, so the right immediate action is to be transparent about a changed process (not promises of quick redemption) and to invite scrutiny rather than avoid it.
One level up
Trust is restored by demonstrated behavior change plus time, not by rhetoric; the value of early communication is to set correct expectations and signal accountability, not to accelerate the timeline.
Where it stops holding
This stops working when the underlying process change isn't real or sustained — communication without follow-through would be seen through and accelerate departures instead.
Detail
situation: D1 suffered a severe, publicly visible drawdown after GameStop and broader 2021-2022 market turmoil; the firm's president suggested canceling the scheduled investor dinners to avoid a 'bloodbath.'
decision: Dan insisted on holding the dinners, opened by inviting LPs to 'ask me anything, criticize me,' and announced a shift to a much less risk-prone portfolio construction ('hit singles and doubles') even though taking more risk post-drawdown would have been the higher expected-value move.
outcome: Some LPs redeemed and Dan felt that was deserved given performance, but the firm retained enough committed LPs and rebuilt from there; he recalls the dinner as a pivotal moment where having an explicit plan, even amid broad doubt, 'changes the perspective entirely.'
lesson: Facing stakeholders directly with a credible changed plan during the worst moment, rather than retreating, is itself a decision point that restores internal conviction even before external trust returns.
In their words
I said, no, we have to do these dinners. And is the most important time to go out there and speak to our investors.
An amateur's online fraud write-up tanking a stock and launching a career
Source
Invest Like the Best with Patrick O'Shaughnessy — Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460] (February 24, 2026) · at 57:31 ▸
When this applies
when someone produces a genuinely rigorous piece of public work (analysis, research, code) as a credentialing move before having any formal credentials
The takeaway
A rigorous, public demonstration of real analytical skill — judged on its own merit rather than your credentials — can open doors that a traditional resume cannot, because the output itself proves capability directly.
Why it works
Posting genuinely rigorous original analysis publicly (even anonymously) lets its quality be judged purely on merit; if the analysis is sound and exposes something material (a fraud), the market itself validates it (stock price moves), and the portfolio of such work becomes a more powerful credential than a resume.
One level up
In fields where skill is verifiable from output, publishing genuine, high-quality work in the open is a bootstrapping strategy for credibility that substitutes for lacking formal pedigree.
Where it stops holding
This stops working when the work can't be externally validated (no market price or equivalent objective check), when the employer's incentives for using your work differ from what you assume (as in the interviewing firm's own motive), or in regulated contexts where sharing such analysis publicly creates legal/compliance risk (which it nearly did here).
Detail
situation: As a young analyst aspiring to join a hedge fund, Dan was given a case study assignment (analyze Orthodontic Centers of America) as part of an interview process; he found evidence of accounting fraud (capitalizing expenses that should be expensed) through deep financial statement analysis.
decision: Before his follow-up interview, he posted the anonymous six-page write-up on Value Investors Club to get feedback, despite being barred from trading the stock himself as a bank employee.
outcome: The stock cratered over the next day as hedge funds and mutual fund analysts (via word of mouth from friends he'd tipped) acted on the thesis; the interviewing firm was alarmed about compliance exposure and didn't hire him for that role, but the write-up itself became a traveling credential that got him his first hedge fund job.
lesson: A single piece of rigorous, verifiably correct public analysis can function as a far stronger credential than a resume, even when the immediate interaction around it goes sideways.
In their words
I now had this write-up that could go around to different hedge funds. And most of them already knew about it because they would short it after the write-up. And that's how I got my job.