Momentum is not equally effective in every market regime. O'Shaughnessy says that after declines of 40% or more, the prior worst performers can beat momentum leaders because strong businesses were indiscriminately pushed to unusually low valuations.

Severe bear markets can invert momentum
High-quality companies may be sold indiscriminately
Recovery buyers gravitate toward deeply discounted survivors
Factor behavior should be examined by regime

“momentum inverted”
Jim O'Shaughnessy · 08:30

“a lot of really great stocks during the bear market got pushed way low in price”
Jim O'Shaughnessy · 08:45

#momentum#bear markets#factor investing

Myth Buster43:14

Why Industrial-Era Metrics Misprice Intangible Businesses

Traditional measures such as book value were designed for an economy dominated by physical assets. Modern companies derive substantial value from brands, software, patents, trademarks, and research, requiring models that estimate assets conventional accounting does not capture well.

Book value reflects physical assets better than intellectual property
Brand and research investments may be economically valuable but poorly represented
Industrial-era economic metrics have lost explanatory power
Valuation models can incorporate proxies for intangible value

“I think there's more than some truth to that.”
Jim O'Shaughnessy · 44:32

“They were designed for manufacturing. They were designed for physical things”
Jim O'Shaughnessy · 45:17

#valuation#intangibles#technology#accounting

Hot Take· 3

Hot Take49:56

AI and Phones Are Breaking Hollywood’s Production Gatekeepers

O'Shaughnessy expects inexpensive cameras, editing software, AI, and online distribution to unlock filmmakers excluded from traditional Hollywood infrastructure. His film venture pairs those tools with transparent economics and ownership participation for contributors.

Professional-looking work can now be produced on phones
AI can augment low-budget filmmaking
Online platforms reduce distribution barriers
Transparent revenue sharing can improve contributor incentives
Narrative-driven films need not depend on expensive effects

“the people who couldn't make movies in the past are going to be able to make them in the future.”
Jim O'Shaughnessy · 50:10

“Everyone who comes and works on one of our films is going to own a piece of that film”
Jim O'Shaughnessy · 52:14

#filmmaking#ai#creator economy#media

Hot Take1:08:28

AI Could Accelerate Drug Discovery—and Catalog What Does Not Work

AI can search molecular combinations, repurpose existing compounds, and perform preliminary in-silico testing at a scale humans cannot match. O'Shaughnessy also proposes generating and publishing likely null hypotheses so researchers can avoid duplicating unproductive work.

AI can explore vast molecular search spaces
Existing drugs may have useful applications elsewhere
In-silico testing can reduce early human and animal experimentation
Scientific incentives discourage publishing negative findings
A shared null-result database could prevent wasted effort

“AI can go into all of those spaces that we humans simply can't do and find the connections”
Jim O'Shaughnessy · 1:09:14

“there's a wealth of information in the stuff that doesn't work.”
Jim O'Shaughnessy · 1:11:24

#drug discovery#ai#science#research

Hot Take1:18:12

The AI Future May Be Many Private Agents, Not One Omniscient Model

Rather than expecting one model to solve everything, O'Shaughnessy anticipates many specialized agents customized to individual users. Privacy is central because people will grant richer calendar, location, contact, and preference access only when they trust that the information remains protected.

Specialized agents can outperform a universal interface on personal tasks
Personal data determines an agent’s usefulness
Privacy encourages users to share richer context
AI enables mass customization across services
Current systems remain early and comparatively crude

“it's the quality of the data that you train your AI on that determines its value to you.”
Jim O'Shaughnessy · 1:21:29

“I see it as a lot of smaller but incredibly useful AI agents doing specific things for each of us.”
Jim O'Shaughnessy · 1:22:04

#ai agents#privacy#personalization#mass customization

Explainer· 4

Explainer17:53

Microcaps Offer Diamonds in the Rough—But a Broad Index Can Be Dangerous

The microcap universe contains potentially exceptional investments alongside thousands of weak businesses. O'Shaughnessy argues that investors need disciplined screening tools rather than indiscriminately buying the entire category.

Microcaps can provide private-equity-like exposure
The universe contains many low-quality companies
Systematic screening helps isolate stronger candidates
Selection matters more than broad category exposure

“you would not want to buy an index of microcap stocks.”
Jim O'Shaughnessy · 18:32

“there are so many diamonds in the rough in microcap”
Jim O'Shaughnessy · 18:41

#microcaps#stock selection#quant investing

Explainer20:02

Custom Indexing Can Reflect Each Investor’s Own Values

Canvas was designed to avoid imposing a universal ESG definition on investors. Advisors can tune granular exclusions or preferences around a client's religious, ethical, employment, or social priorities while retaining a systematic portfolio structure.

Values differ substantially between investors
Preset ESG portfolios impose someone else’s definitions
Custom screens can exclude or emphasize specific activities
Advisors understand individual client requirements better than product manufacturers

“give you, as the advisor, full control over what your client portfolio wanted to look like.”
Jim O'Shaughnessy · 20:32

“we wanted you to be able to reflect your client's unique needs.”
Jim O'Shaughnessy · 25:43

#custom indexing#values investing#esg

Explainer54:17

AI Can Expand a Book Across Languages, Audio, and Local Promotion

Infinite Books aims to replace legacy publishing economics with higher author shares and AI-enabled services. The envisioned stack includes rapid translation, multilingual synthetic narration, and agents that identify geographically relevant promotional opportunities.

Authors can retain a larger share of revenue
AI translation can open multiple language markets
Voice models can narrate translated audiobooks
Marketing agents can target local podcasts and publications
Authors retain control over whether AI services are used

“the current publishing industry is still playing under nineteen twenty rules, not twenty twenty rules.”
Jim O'Shaughnessy · 54:31

“all of these tech advantages are being left just lying around on the floor”
Jim O'Shaughnessy · 56:03

#publishing#ai#books#creator economy

Explainer1:11:35

Why Open Model Weights Can Trigger a Creativity Explosion

O'Shaughnessy supports open-source foundational models because widely available weights allow independent builders to modify and improve them. He views Stable Diffusion's release as evidence that distributed experimentation can produce rapid creative and technical progress.

Model weights are necessary for deep modification
Open releases reduce concentrated control
Distributed builders address diverse problems
Stable Diffusion prompted rapid creative experimentation
Linux provides an analogous open-source success

“They released the whole thing, full open source, driven source, fully transparent”
Jim O'Shaughnessy · 1:12:53

“when you allow all of these clever people the ability to play with it, to tinker with it, you get a much better model.”
Jim O'Shaughnessy · 1:13:16

#open source#ai models#stable diffusion#innovation

Story· 1

Story58:53

The Fellowship Finding Brilliant People Outside Traditional Networks

O'Shaughnessy Ventures gives grants and fellowships to overlooked creators, scientists, and founders without requiring equity or first-refusal rights. Examples include an Indian musician, an African biodegradable-products venture, a refugee-services app, and open-source quantum-computing research.

Geography and credentials historically concealed talent
Digital networks make global discovery possible
Fellowships provide $100,000 over a year
Funding carries no contractual strings
Cognitive diversity encourages cross-pollination

“We're gonna find and fund them and see what comes from that.”
Jim O'Shaughnessy · 1:00:16

“The fellowships are one hundred thousand over a year, no strings.”
Jim O'Shaughnessy · 1:00:29

#fellowships#talent#funding#innovation

Frameworks in this episode

Custom-Index Tax-Loss Harvesting

Evidence-First Factor Investing

Human-Nature Investing Lens

Human-Plus-Machine Co-Creation Loop

Insights & moments

⚡Myth Buster· 2

Why Momentum Can Reverse After a Severe Bear Market