Investing in Startups

The episode is packed with substantive, research-backed observations about seed investing that challenge conventional wisdom - power laws, check sizing paradoxes, adverse selection dynamics, and why being contrarian is nearly impossible at seed. The density of novel claims is high throughout (e.g., small checks outperforming, idiosyncratic signals mattering zero), though some sections drift into explanation of already-stated concepts rather than introducing fresh ideas.

“seed is actually relatively pretty efficient. The company raising at 75 is probably about 15 times better than the company raising at 5”

“the most expensive seed deals should actually be the highest performing because they're essentially aggregating a bunch of little alpha edges from all the positive signals”

The research findings are genuinely counterintuitive and move beyond recycled venture wisdom: small checks outperforming large checks, the asset class behaving more like vintage Rolexes than municipal bonds, and the discovery that larger portfolios should outperform concentrated ones at seed due to power law properties. However, the framing leans on existing academic concepts (Markowitz, power laws, adverse selection) and the host-guest dynamic doesn't push into truly unexplored territory - it validates and explains rather than break entirely new conceptual ground.

“small checks from VCs outperform other checks they write”

“seed does look a lot more like, it does look a lot more like the, the market for vintage Rolexes or cool Air Force One shoes than it does for municipal bonds”

Abe Othman is a legitimate researcher at AngelList with seven years of experience analyzing one of the largest seed-stage datasets in existence. He has published peer-reviewed work, conducted original analysis, and speaks with earned authority about data he directly manages. This is not a career podcast guest or generic thought-leader - he's a practitioner-researcher with direct access to the signal being studied, giving him credibility that shows in specifics.

“I'm a researcher at Angellist”

“I have the privilege of working with”

The episode includes concrete data points and specific frameworks (alpha parameters, MOIC metrics, check size ratios like 2x typical checks, the 13x valuation gap example from $5M to $65M preseeds). However, the researcher frustratingly avoids naming specific companies, founder stories, or detailed case studies that would ground claims further. Metrics are referenced but rarely quantified beyond 'we looked at hundreds of GP year examples' - the underlying dataset is proprietary and largely unavailable to the listener for verification.

“someone's doing a Pre seed at 5 versus someone's doing a seed at 75, there's not much difference”

“the hit rate of like you know, 10x MOIC seed deals...you can kind of see an up into the right slope as relative valuations increase”

Joe Maker asks substantive follow-ups and attempts to stress-test Abe's claims (e.g., probing the small vs. large check tension, asking whether results net out via adverse selection, requesting emotional reactions to findings). However, the host rarely pushes back on conclusions or presses Abe on unsupported leaps. The conversation is collaborative rather than adversarial; Abe's framings are largely accepted rather than challenged. Joe could have pressed harder on why proprietary data constraints prevent independent verification or whether the vintage Rolex analogy might be self-fulfilling.

“Is it possible that the lack of edge or alpha or outperformance on the large check could be some kind of middle ground between”

“You kind of alluded to a couple times your opinions change on different ideas, perspectives over time...What does that look like?”