Steve Eisman brings a gold rush analogy to Triggernometry
The investor portrayed by Steve Carell in The Big Short asks whether the companies building AI can make the economics work.
Big Short Investor: Is the AI Bubble About to Burst? - Steve Eisman
Up next
Why this story? The channel's weekly summary: what went up this week, what is drawing views now, and where the channel stands among similar channels.
Steve Eisman starts with a distinction that gets lost in arguments about AI: predicting the next word isn't the same as booking a theatre ticket. On Triggernometry, he tells Konstantin Kisin and Francis Foster that a language model does the former, while an AI agent can use the model to carry out a task like the latter. Some viewers picked up on that distinction, asking whether it matters if a model can think as long as it gets the job done.
Eisman, who made his name betting against debt tied to the US housing bubble, is more worried about the cost of building AI than about it wiping out jobs. In his gold rush analogy, Nvidia supplies the picks and shovels. The companies spending heavily on data centres and building models have a different problem: Eisman argues that cheaper alternatives make it hard for model developers to protect their business. The episode includes paid ad segments.
"Big Short Investor: Is the AI Bubble About to Burst? - Steve Eisman" drew about 165,000 views in nine hours, making it the most watched of four new uploads and the fastest-moving video on the channel. The Off the Fence round-up on Jensen Huang's resignation and the Trump-Xi-Putin summit was next for daily momentum. Returning guest Brianna Wu drew more views overall than that round-up with her argument that Democrats should read the midterms as a verdict on the cost of living, not a sweeping mandate. She also had a separate member Q&A.
Subscribers held steady over the week. Across similarly sized commentary channels, Triggernometry's weekly views were close to Rebel News's, while Occupy Democrats and The Bulwark had much busier weeks. Triggernometry's latest daily viewing was below half its usual pace, even with the strong start for Eisman's interview.
An American investor who profited by betting against securities tied to the U.S. housing bubble before it burst in 2007–2008.
More on Wikipedia ›A U.S. technology company that designs graphics processors used for artificial intelligence and gaming.
More on Wikipedia ›Published on Trari
Where next?
Things we can learn from the video
Press to jump
- 3:07Eisman argues that language models predict words rather than think.
- 13:40Big cloud providers are taking on much heavier capital costs as they build data centres, Eisman says.
- 20:05Eisman compares fears of AI job losses to earlier predictions that machine learning would eliminate radiologists.
- 42:25Eisman sees a bubble risk: model developers lack durable protection from cheaper competitors whose model weights are publicly available.
- 51:38Eisman points to differences in energy, permitting and state financing when comparing US and Chinese AI development.
A professional in this field? Want to write about Money & self-improvement on Trari? Tell us ›
What viewers say
Was cutting grass and this came up. I am a regular listener of The Real Eisman Playbook, and I have listened to triggernometry for years. When I saw Steve was the guest, I was thrilled. Big fan!
@dprestons0318See it on YouTubeWhy the preoccupation with AI being able to "think" or not? If AI can achieve results which mirror intelligence, the underlying mechanism does not matter.
@zachmallet47See it on YouTubeI like how you guys are not agreeing. We need more of that. It was also handled well, Steve is a great guest. Cheers.
@ApprimeableSee it on YouTube
1 more comment
Have this guy back on ASAP, excellent conversation
@michaelscandiffio549See it on YouTube
Public comments from the video's page on YouTube
Want to know when somebody answers you?
The videos drawing views on Triggernometry right now
- Channel
- Triggernometry
- Subscribers
- 1,960,000
- Channel views, last 7 days
- 1,398,723
- Videos published this week
- 4
- Video published
- November 21, 2026
- Video views so far
- 164,812
The numbers
Here are the week's views and subscriber totals, along with how the uploads are moving now.
💡 Test your knowledge
What is Anthropic's flagship family of AI models called?
Claude is the name Anthropic gives its flagship family of language models.




