Jamison: AI Venture Bets Resemble Airlines With High Capital and No Moats
Travis Jamison · 30% IRR From Plumbers and Dry Cleaners? Here Is How · Jul 4, 2025 · at 4:46
Travis Jamison compares picking AI foundation model winners to investing in historical airline companies, referencing Warren Buffett's analysis.
0:00 / 0:26exact quote · 26.8s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“If you take all the profits of all the airlines over time, it's negative for investors. Like shareholders have had a negative return for that because they've just kind of competed themselves away. There's no giant moat to get those monopoly profits. And it's a really hard business. It's really, really capital intensive. AI is sort of looking like that a little bit to me from like picking the winners, you know, Picking those venture bets types of things. It's really hard to do, I would say.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Travis Jamison
Jamison: Venture capital's heyday is over due to excess capital
“And also when it comes to venture, I think venture's heyday is kind of over, at least for now. I think there's too much money has entered the space, so too many companies that shouldn't have been funded have been funded.”
Jamison: AI Will Not Materially Impact Physical Trades Like Plumbing
“AI is not going to change how, you know, someone remodels a house, how a plumber changes the pipes, like that type of thing, and at least not in a material enough way to make a difference.”
Jamison: AI foundation models will be commoditized without monopoly profits
“All of the big tech companies right now that are investing very, very heavily into building their own models. It's just going to be commoditized at the end. And no one really gets that monopoly profits where it's like, you know, a search, like Google got monop…”
Jamison: AI Model Niches Like Claude for Coding Lack True Differentiation
“I think even though the different companies are trying to niche down, you know, like Claude for coding as an example, I still don't think there's enough of differentiator to matter for that.”
Jamison: Creators can train LLM recommendations via self-published third-party content
“Cause you know, the LLMs are looking at Google results. We know that you can slowly but surely like train it on what is best and you have control that entire process. It doesn't have to be reliant on actual third party people. It can just be you on third party…”
Jamison: Holding 10-20% cash pays off by preventing panic selling during crashes
“There's a drag on your portfolio, but having, you know, you keep 10, 20% in cash or something like that, treasure bills, you know, there's going to be a drag on your portfolio, but if that drag lets you purchase when things crash, or even more correctly, proba…”