
Aug 17, 2026 · 1h 7m
Alec Litowitz argues adaptability beats certainty in investing
E417: $20B Investor on Risk, Uncertainty & Adaptability
As AI and algorithms make information faster but not necessarily better, Litowitz explains how investors can make decisions without mistaking confidence for knowledge.
- 1Investors must separate measurable risk from uncertainty, where probabilities and reliable models may not exist.
- 2Strong opinions become useful when held provisionally, tested against missing information, and separated from personal identity.
- 3Adaptable institutions prioritize exceptional people, recoverable experiments, feedback, and learning over individual trade-making or vindication.
Don't miss
Litowitz explains why separating decisions from identity lets investors run recoverable experiments, revise strong opinions, and learn without treating error as personal defeat.
The brief
Alec Litowitz opens with the distinction between measurable risk and uncertainty, using early convertible-arbitrage and merger investments—including an AbbVie and Shire failure—to show where models stop helping.
The founder of Magnetar argues that good decisions require provisional beliefs, Bayesian updating, and a willingness to ask what trusted information or missing perspective could overturn a thesis.
Litowitz treats culture as an investment system: hire exceptional people, demand high standards, separate identity from decisions, and reward learning rather than personal vindication.
The conversation extends that framework to AI and algorithms, which can amplify emotional extremes or produce average answers unless used to challenge assumptions and expand possibilities.
His closing advice returns to beginner’s mind: uncertainty is not merely a threat, and fear of mistakes can prevent the experimentation that makes adaptability possible.
What was said on this episode
18 statements · 11 positive · 4 negative · 1 mixed · 2 neutral
Resolving uncertainty before others can create better investment opportunities.
“if I can go into those spaces and I can resolve the uncertainty, let me try to model where that world's going before everybody else does, then I think that you can structure much better upside, downside opportunities.”
Listen at 0:37
About 92% of announced merger deals complete, versus market pricing of 87%.
“about 92% of announced deals go through. But the market trades at like 87% goes through.”
Listen at 1:56
Investment edge is greater in unresolved uncertainty and disappears once widely understood.
“The edge is usually greater there. As soon as everybody else figures out if it's going through or not, that edge is gone.”
Listen at 4:31
Litowitz made a large bet that the AbbVie-Shire merger would be approved.
“we had a very large bet that the deal would go through.”
Listen at 6:05
The AbbVie-Shire tax-policy uncertainty could not be resolved externally.
“that uncertainty either couldn't be resolved because it was in the heads of one or two people at Treasury.”
Listen at 6:49
Investors should target uncertainty that can be resolved.
“I want to go where uncertainty is, but I want to go where it's resolvable.”
Listen at 7:15
SpaceX’s adaptive cognitive culture contributes to Musk’s success.
“that is a culture of cognition and adaptation that is quite unique. And that's why he's been so successful, in my opinion.”
Listen at 14:01
Attaching ego to adaptability encourages changing views when evidence changes.
“If you force people to attach their ego to being adaptive, it requires them to want to change their mind if the evidence changes.”
Listen at 26:37
Organizations should optimize for adaptability rather than being right.
“You make the production function adaptability and not about being right.”
Listen at 26:50
Magnetar’s non-siloed structure enabled pursuit of uncertain opportunities.
“We created a very non-siloed firm because We specifically didn't know where the opportunity was going to be”
Listen at 29:39
Adaptability will be the most important trait in the future.
“The most important trait coming up, in my opinion, is going to be adaptability.”
Listen at 31:55
AI will outperform humans in mathematics and coding.
“AI is going to outperform you in math and coding.”
Listen at 35:24
Social media can become a major influence on children’s upbringing.
“social media has created part of the environment in which, if you allow your kids to have access, in which they are raised, not by you. but by something else.”
Listen at 57:46
Social-media engagement incentives amplify extreme opinions and polarize users.
“engagement in social media meant the extreme opinions were more engaged. And therefore, we polarized people by giving these really wide opinions.”
Listen at 58:08
Social-media algorithms highlighted extreme viewpoints.
“algorithms there highlighted the extremes.”
Listen at 58:35
AI systems tend to provide the average expected answer.
“AI now, you go, you get an answer. What does it give you? The mean. It gives you the best expected answer out of the distribution.”
Listen at 58:41
AI should expand users’ possibilities rather than simply provide answers.
“How do you augment? Give me things I didn't think about. How would I expand the set of possibilities as opposed to going, just give me the answer.”
Listen at 59:25
AI is most powerful when used to augment human thinking.
“AI is the same thing. My book is saying, here's how you use it to augment.”
Listen at 1:01:07
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.
