ExpertsFMJeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481]
Invest Like the Best with Patrick O'Shaughnessy · July 7, 2026
5 Selected Lessons:
heuristic · 3 reasoning_pattern · 1 framework · 1
Why SaaS gross margins are structurally eroding (strings vs. compute)
Source
Invest Like the Best with Patrick O'Shaughnessy — Jeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481] (July 7, 2026) · at 45:42 ▸
When this applies
when a product's underlying cost structure shifts from near-zero marginal cost to per-use marginal cost
The takeaway
When a product's marginal cost structure changes from near-zero to significant-per-use, expect margin profiles to compress industry-wide and competitive advantage to shift toward scale rather than margin.
Why it works
Traditional software had near-zero marginal cost (selling a 'copy of a string'), enabling high gross margins; AI-driven products require real compute per use, reintroducing meaningful marginal cost, so the old high-margin SaaS model structurally breaks and returns shift to whoever achieves scale at thin margins.
One level up
Re-derive a business model's expected margin and competitive dynamics from its true marginal cost structure rather than from the margin norms of its prior category.
Where it stops holding
This stops working for software products that remain genuinely zero-marginal-cost (e.g., not requiring per-inference compute), which can retain the old high-margin SaaS economics.
Detail
procedure:
- Classify the product by marginal cost: near-zero (string-copy) vs. significant (compute-per-use).
- Expect gross margins to track that classification: high for near-zero marginal cost, compressed for compute-heavy products.
- Expect competitive returns to accrue to the largest-scale providers because low margin + low scale is a bad business.
- Anticipate 'mom and pop' SaaS providers getting squeezed by scale players, a 'Walmart effect' in software.
benchmarks: {"prediction":"will see $10 trillion market cap companies as scale compounds with thinner margins"}
rationale: Software's historical high-margin era was a 'downstream coincidence of selling strings'; selling compute instead breaks that coincidence.
In their words
Now I think we're in an era where we're selling compute and selling compute, can't write the prompt once and then sell copies of the output, you have to do the compute every single time and so the marginal cost obviously is not zero.
Giffon's speculative long-run case for why AI job displacement fears may be overstated
Source
Invest Like the Best with Patrick O'Shaughnessy — Jeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481] (July 7, 2026) · at 31:48 ▸
When this applies
when a new automating technology threatens to eliminate a large category of existing work
The takeaway
A technology that automates existing work may not eliminate the underlying demand for human economic activity if that activity was never tightly coupled to scarce survival necessity — it could instead push the activity into newly invented categories of want.
Why it works
Giffon's reasoning: most white-collar jobs are not contingent on basic survival necessities (food, shelter, medicine) but exist because capital is inherently inflationary and must be deployed, and because human wants/desires tend to be unlimited — so he speculates that when automation removes one category of 'made up' work, the same underlying forces (unlimited desire, capital needing deployment) could generate new categories of work to replace it.
One level up
Distinguish work tied to irreducible scarcity (survival) from work generated by elastic, self-renewing demand (status, novelty, capital deployment) when forecasting the long-run effect of automation.
Where it stops holding
This stops working in the short-to-medium term, where Giffon explicitly says displacement 'could very well be bad' and cause real volatility and despair before new work categories emerge.
Detail
approach: Separate survival-necessary work from 'made up' white-collar work (including his own allocator job) before predicting displacement effects.
signals:
- work-from-home viability as evidence most knowledge workers don't have 40 hrs of real work
- four-day-week trend as 'soft launch'
- capital's inherent need to be deployed productively generating jobs around it
rationale: Giffon suggests we've largely 'solved' survival-level problems and believes the economy runs on inventing new things to want and new jobs to service that want, which he predicts won't run out — though this remains his own forward-looking speculation.
In their words
We're gonna have unlimited wants and desires and our economy is solely driven by our unquenchable desire to consume things. So we're gonna come up with new things to consume.
Narrative flexibility for old-but-newly-growing companies
Source
Invest Like the Best with Patrick O'Shaughnessy — Jeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481] (July 7, 2026) · at 2:09 ▸
When this applies
when any venture or asset has strong recent performance but an unflattering long history clouds the story
The takeaway
The same fundamentals can be priced very differently depending on which narrative clock you start, so actively choose and reset the framing rather than let default history dictate it.
Why it works
Investors/narrators anchor on company age and cumulative story rather than recent trajectory, so an identical business performs worse under an 'old company' frame than a 'young hot company' frame.
One level up
In markets driven by story more than fact, controlling the frame is as valuable as controlling the fundamentals.
Where it stops holding
This stops working when the underlying numbers are so weak that no reframing can survive diligence, or when counterparties insist on verified history regardless of framing.
Detail
rule: If a business's story is 'hurting' it (e.g., too old, too slow historically) but recent performance is strong, deliberately re-narrate the timeline rather than accept the default framing.
threshold: Company is several years old (e.g., 6-7 years) but has recently inflected (e.g., 200% growth) and is still being judged on cumulative absolute numbers (e.g., $8M revenue).
rationale: Narrative, not just fundamentals, prices capital; the same facts with a different starting clock read as a wholly different opportunity.
In their words
Whereas I think literally if you just change the name and told a different story and just sort of arbitrarily started the clock two years ago, that company would actually be like really hot.
Entertainment is the actual job-to-be-done of all media
Source
Invest Like the Best with Patrick O'Shaughnessy — Jeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481] (July 7, 2026) · at 25:14 ▸ · at 27:44 ▸
When this applies
when evaluating any information-consumption habit you've told yourself is 'productive' rather than leisure
The takeaway
Don't let a product's self-justifying framing (useful, productive, an asset) substitute for honestly assessing its actual function (entertainment, consumption).
Why it works
Media is produced, selected, and edited specifically to be entertaining because that's what algorithms and creators optimize for; framing consumption as 'productive' or 'investment' (versus 'liability') changes willingness to spend time/money on it even though the underlying function is unchanged.
One level up
Strip away marketing/self-narrative framing to see the true functional category of any product or habit before deciding how much of your scarce resource to allocate to it.
Where it stops holding
This stops working when the specific content genuinely transmits actionable, rare information not obtainable from secondhand discussion — but Giffon argues this is rare enough that the default assumption should be entertainment.
Detail
rule: Treat all consumed media as entertainment by default rather than as useful or productive work; decide deliberately how much entertainment you want rather than deceiving yourself.
threshold: n/a
rationale: You don't really miss anything important if you're not a complete hermit — you can get filtered takes secondhand from people who already absorbed the 'radiation.'
In their words
I think my takeaway from that would simply be that one should not fool themselves that they are looking for anything other than entertainment in all the media that they consume.
Professional constraints make beating the market harder for managers than amateurs
Source
Invest Like the Best with Patrick O'Shaughnessy — Jeremy Giffon - The Billion Dollar PDF - [Invest Like the Best, EP.481] (July 7, 2026) · at 50:25 ▸
When this applies
when any professional's structural constraints (mandates, client management, career risk) are compared with an amateur's freedom to act on conviction
The takeaway
Don't generalize an institutional base rate (professionals underperform net of fees) into a universal claim about the difficulty of outperformance — institutional constraints, not pure difficulty, often explain the gap.
Why it works
Professional managers carry mandates, client-facing obligations, and career/explainability risk that amateurs lack, meaning the professional's effective opportunity set and risk tolerance is narrower even if their raw skill/knowledge is equal or greater — so 'can't beat the market' empirically observed in professional-fund data doesn't imply skilled amateurs can't outperform.
One level up
Separate structural/incentive explanations from capability explanations before accepting a widely-repeated 'it's just hard' narrative.
Where it stops holding
This stops working for amateurs who lack even Buffett/Lynch-style conviction and simply chase recent performance without real edge — the heuristic favors people with genuine insight and patience, not any amateur.
Detail
rule: Buffett's 'put it in the S&P' advice is meant for the average person, not as proof active investing can't beat the market; professionals' structural constraints (not general difficulty) explain most of the underperformance base rate.
threshold: Contrast cases: amateurs who bought Bitcoin, Tesla, or Apple stock (as consumers of the products) often outperformed professional funds that structurally couldn't run concentrated, conviction-driven books.
rationale: A professional manager has mandates, business obligations, and client relationships that amateurs don't, making concentrated conviction bets much harder to hold.
In their words
it is really hard to beat the market because you have all these other factors that the average person doesn't have and this is the Peter Lynch argument.