Animal Spirits Podcast Episode 452: Is AI a Mistake?

Date: February 18, 2026
Hosts: Michael Batnick (“A”) & Ben Carlson (“B”)

Episode Overview

This episode grapples with one of the most pressing questions of our time: Is AI a mistake? Michael and Ben wade into the rapidly evolving landscape of artificial intelligence and its impact on jobs, markets, and society, mixing their trademark mix of skepticism, humor, anxiety, and pragmatism. Along the way, they tackle the week’s tumult in software and wealth management stocks, the realities of tech disruption, the durability of SaaS businesses, prediction markets, the future of work, and some classic Animal Spirits cultural asides.

Key Discussion Points & Insights

1. The Great AI Debate: Hype, Fear, and Middle Ground

Timestamps: 03:17 – 13:54

• AI discourse is polarized: Is it billionaire-hyped vaporware, or an existential threat to white-collar jobs?
• Michael references Derek Thompson’s framing: people are either dismissing AI as hype or panicking that it’s 12 months from replacing all white-collar work.
• Ben observes, “There’s some head-in-the-sand people… I think that’s where that line of thinking comes from.” (04:38)
• Both hosts position themselves in the “middle ground”: AI is real and rapidly evolving, but claims of instant job apocalypse are overblown.
• Ben: “What if AI is a mistake? What if everything these tech people are saying is true and…a majority of people in the future think: why did we do this? I think more people are going to hate AI than like it going forward.” (05:50)
• Michael pushes back, suggesting over long timeframes, the “AI mistake” take may age poorly: “By 2080, or like 100 years from now, I think that take would age poorly, like every other take does.” (06:36)

2. AI’s Real Productivity Gains – and Who Gets Left Behind

Timestamps: 07:22 – 13:54

• Ben sums up viral “AI is moving faster than you think” essays by Matt Schumer and others, which claim weeks of work are now doable in hours.
• The hosts note most writing about AI disruptions is by tech workers for tech workers—extrapolation isn’t always reality.
• Michael: “So much of the population is so oblivious to what's happening... there is a lot of extrapolating being done.” (08:22)
• Current fears in the market (e.g., software stock re-ratings, job risk for tech workers) may be a massive case of recency bias.
• They debate how easy it would actually be for AI to replace all knowledge work and whether it will lead to more inequality.

3. Inside the AI Labor Shock: First-Person Accounts

Timestamps: 14:42 – 17:12

• Read an email from a 31-year-old data scientist describing terror at the pace of AI’s skill acquisition:
  • "This is task completion time horizon and it’s doubling every seven months... AI is decimating headcount and time to market... I’m terrified. I don’t know what’s going to happen over the next 12 to 24 months. I don’t know if I’ll have a job.” (Listener email at 15:14)

4. AI Company Moats, Power, and Surging Revenue

Timestamps: 17:32 – 21:29

• Discussed the baton-passing from OpenAI to Anthropic (Anthropic’s revenue skyrocketing, major VC raises).
• Questioned what the true "moat" is for AI firms when tools and research seem easy to copy.
• AI compute costs are enormous, limiting how fast even top players can scale. (Ben: “The costs are so astronomically high… we can’t push the RPMs all the way to the red line.” 17:58)
• Market share stats: “1 in 5 businesses on Ramp now pay for Anthropic… the overlap with OpenAI is 79%.” (19:34)

5. Are Markets Overreacting to AI’s Disruption?

Timestamps: 22:02 – 31:10

• Michael points to irrational market moves—e.g., Schwab stock falling 10% on Altruist’s new Hazel AI tool—calling it “massive recency bias, overreaction.” (23:29)
• Ben: “This tech is very real… But this technology is going to make us so much more efficient. Our financial advisors’ lives will be easier.” (23:59)
• Demographics and productivity: With falling birth rates, AI is coming at "the perfect time" to fill the economic gap.
• Michael calls out “career risk” and organizational inertia as significant moats for incumbents—even as code becomes easier to replicate.
• Notable quote on enterprise software: “When you buy software for your company, you're not just buying features, you're buying someone to blame when things go wrong...You pick the option that if it fails, you can defend to your boss. ‘We went with Salesforce’ is a defensible sentence in any boardroom in America. ‘We went with an app I vibe coded over the weekend’ is a resignation letter.” (Michael, quoting Finbar Taylor, 27:00)

6. Valuations, SaaS Moats, & Stock Market Volatility

Timestamps: 28:09 – 37:40

• Market is compressing multiples—not always in anticipation of revenue collapse, but an end of "premium" multiples as moats dissolve. (Nicholas Bustamanti, 29:08)
• Re-rating of SaaS stocks is “appropriate,” but the market typically overreacts.
• Notable: Intense single-name stock volatility even as indices remain near all-time highs. (Ben, 32:12)
• Michael: “The bubble we were promised is not so bubbly... This is a weak-ass market compared to the dot-com bubble, obviously.” (33:46)

7. International Markets & Economic Contradictions

Timestamps: 39:50 – 43:08

• Japan’s stock market is outperforming the S&P 500 over five years, with rising rates seen as a positive sign.
• International small cap value stocks are also outperforming mega-cap US tech, an underappreciated trend.

8. Prediction Markets Take Center Stage

Timestamps: 54:42 – 64:01

• Surge of interest in prediction markets for both economic/data forecasts and “truth-seeking.”
• Vanity Fair & NYT: Average accuracy of prediction markets rivals or outperforms expert forecasters and Wall St. analysts.
• Thomas Peterffy (Interactive Brokers founder): “I am extremely bullish on the prediction markets. I think this is going to be the biggest thing that happened in our business in the last hundred years… By assembling these forecasted facts, we are mapping the future with high resolution.” (Michael reading at 60:05)
• Concerns about prop bets, leaks, and guardrails so the wisdom of the crowd isn’t undermined by manipulation.

9. AI, The Future of Work, and Human Psychology

Timestamps: 44:45 – 50:56

• "Prime age" (25-54) labor force participation is at all-time highs—AI’s labor reckoning isn’t here yet.
• Michael posits two possible future scenarios: sudden labor shock vs. slow absorption and business adapting without mass unemployment.
• Ben repeats, “The course of human history is: we figure it out. We're resilient, we survive, we're dynamic.” (47:07)
• Leisure time is increasing—are we happier? Debated whether more time to think makes us more miserable or is simply a continuity of human psychology.

10. Society, Nostalgia, and Everyday Life

Timestamps: 51:13 – 69:26

• Cultural asides: 80s movies, Dad life milestones, middle-aged humor, being roasted by kids.
• Ruminations on the shift from frugality as a path to wealth to recognizing high income (not saving on coffee) is more impactful.
• Discussed housing vs. portfolio wealth and the generational real estate divide.

Notable Quotes & Memorable Moments

• Ben (About the “AI mistake” fear): “What if everything these tech people are saying is true and…a majority of people in the future think: why did we do this?” (05:50)
• Michael (Enterprise adoption): “You pick the option that if it fails, you can defend to your boss.” (27:00)
• Listener email: “I can now accomplish what previously took a team of four… I’m terrified. I don’t know if I'll have a job.” (15:14)
• Ben (On social media vs. AI): “I think if we could snap our fingers and get rid of social media, society would be better.” (46:27)
• Thomas Peterffy, via Michael: “I am extremely bullish on the prediction markets...By assembling these forecasted facts, we are mapping the future with high resolution.” (60:05)
• Ben (AI and writing): “I just don't think people are going to want to read AI stuff. Maybe I'm wildly off base here. I feel like you can still tell what is written by AI.” (36:20)
• Ben (On market sentiment): “If you're pounding the table that AI is not going to do anything…you're wrong. That's the only thing I can say definitively right now.” (72:43)

Segment Timestamps

• 03:17 – 13:54: The AI spectrum: hype vs. apocalypse; the middle ground
• 14:42 – 17:12: Listener email on real productivity shocks & fears
• 17:32 – 21:29: AI business moats & Anthropic’s rise
• 22:02 – 31:10: Market reaction/overreaction & tech sector volatility
• 31:31 – 37:40: SaaS moats, enterprise inertia, and valuation
• 39:50 – 43:08: Global markets – Japan, small cap value
• 44:45 – 50:56: AI, participation rates & human adaptability
• 54:42 – 64:01: The ascendance and promise/dangers of prediction markets
• 66:04 – 69:26: Frugality, wealth-building, and generational shifts

Original Tone & Takeaways

• The episode is equal parts anxious, skeptical, and hopeful. The hosts steer clear of doomer gloom and utopian hype, repeatedly landing on “thoughtful middle ground.”
• Their big-picture advice: Don’t ignore AI, but also don’t believe every apocalyptic or messianic take—historically, technological progress ultimately helps more than it harms, but the ride is always bumpy.
• Markets are currently pricing in extreme outcomes and may be overreacting—both in AI and other sectors.
• Prediction markets and the wisdom of crowds are highlighted as potentially society-changing, though not without pitfalls.
• On career risk, enterprise inertia, and the human element, Michael and Ben stress that “knowing people” still matters, and organizational change is slow—even as tools get smarter.
• The episode remains rich with personal anecdotes, market war stories, and pop-culture detours, ensuring the financial analysis remains relatable and lively.

For questions or feedback, email animalspirits@thecompoundnews.com.
See you next week!