What the Viral A.I. Doomsday Essay Gets Right (And Wrong)

1. Viral Essay Triggers Market Sell-off

Timestamp: 00:01:17 to 00:04:16 - watch this moment on skim

A research firm's essay, 'The 2028 Global Intelligence Crisis,' predicting AI-driven job losses and market contraction, went viral and is being blamed for a significant stock market sell-off, impacting companies like DoorDash, American Express, and Blackstone. This highlights the era of market-moving science fiction where speculative AI takes can trigger billions in losses.

Significance (High): The market's extreme reaction to a speculative essay underscores the high anxiety surrounding AI's economic future. It suggests a disconnect between current data and future expectations, making markets highly sensitive to narrative.

Sources in support: Casey Newton (Host), Kevin Roose (Host)

Neutral sources: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

2. Anton Corinick on AI's Economic Data

Timestamp: 00:07:00 to 00:10:10 - watch this moment on skim

Professor Anton Corinick states that current economic data shows only very small, contested impacts of AI on the job market and productivity. He emphasizes that significant economic shifts are still in the realm of expectations, not concrete data, and that even when visible, economic research may remain contentious due to lags and rapid technological advancement.

Significance (High): This assertion grounds the discussion in reality, contrasting the market's speculative frenzy with the slow, often debated, emergence of AI's economic effects. It highlights the difficulty in quantifying AI's impact in real-time.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

3. The Gap Between AI Capabilities and Deployment

Timestamp: 00:10:11 to 00:12:44 - watch this moment on skim

Corinick explains that a significant gap exists between the frontier of AI capabilities and their actual implementation in daily corporate use. A survey of 6,000 executives found 70% used AI, but 80% saw no impact on employment or productivity, indicating that companies are still in the early stages of figuring out productive deployment.

Significance (High): This gap explains the current disconnect between AI's potential and its observable economic effects. It suggests that the hype is ahead of the practical application, and significant effort is still needed to translate advanced AI into tangible business value.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

4. Ghost GDP and Unaccounted Economic Production

Timestamp: 00:12:45 to 00:15:14 - watch this moment on skim

The concept of 'ghost GDP' suggests that as AI becomes more capable, significant economic production may not show up in traditional GDP metrics because it's generated by machines and counted as intermediate goods rather than final consumption or investment. This means potential economic gains might not benefit human workers or be fully reflected in economic statistics.

Significance (High): This concept highlights a potential blind spot in economic measurement, suggesting that AI's true economic contribution could be underestimated. It raises concerns about wealth distribution and the accuracy of economic forecasts.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

5. Spectrum of AI Growth Scenarios

Timestamp: 00:15:15 to 00:18:19 - watch this moment on skim

Corinick outlines a spectrum of AI-driven economic growth, from slow 1-2% annual growth to unprecedented 10-20% hypergrowth. He believes triple-digit growth is unrealistic unless AI is deployed irresponsibly, but also dismisses 1% as too low. In optimistic scenarios, he foresees low double-digit growth rates, contingent on full AI capabilities including robotics.

Significance (High): This nuanced view challenges extreme predictions, suggesting a more moderate but still significant growth trajectory. It emphasizes that the outcome depends heavily on the responsible deployment of AI and the integration of physical and cognitive capabilities.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

6. AI as Substitute, Not Complement, for Labor

Timestamp: 00:18:20 to 00:21:17 - watch this moment on skim

Corinick's 2017 paper predicted AI systems at AGI level or beyond would substitute for, rather than complement, human labor. His reasoning stems from studying neuroscience and computer science, concluding that AI's scalability and lack of biological constraints make it capable of surpassing human intellectual capabilities.

Significance (High): This perspective challenges the optimistic view of AI as a tool that merely enhances human jobs. It suggests a fundamental shift where AI could render many human roles obsolete, necessitating a re-evaluation of labor markets and economic structures.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

7. The Lump of Labor Fallacy and AI's Threat

Timestamp: 00:21:18 to 00:24:17 - watch this moment on skim

Corinick acknowledges the 'lump of labor fallacy' but argues that AI's potential to supplant human labor could shift the demand curve downwards, leading to a contraction in either job quantity, wage levels, or both. He suggests that while labor might not fall into absolute unemployment, its share of economic output could shrink significantly.

Significance (High): This nuanced take on job displacement suggests that even if new jobs are created, the overall demand for human labor and its economic value could diminish. It implies a potential future where labor's bargaining power and compensation are significantly reduced.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

8. AGI as a Fringe Economic Perspective

Timestamp: 00:28:01 to 00:30:48 - watch this moment on skim

Corinick believes that seriously considering Artificial General Intelligence (AGI) remains a fringe perspective in economics. He posits that reaching AGI would not be an endpoint but the beginning of a profound economic transformation, a view that places him even further outside the mainstream of his profession.

Significance (High): This highlights the intellectual divide within economics regarding AI's ultimate potential. Corinick's outlier status suggests that the profession may be underestimating the transformative power of AGI, potentially leading to unpreparedness for its consequences.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

9. Hyperbolic Growth and Recursive Self-Improvement

Timestamp: 00:30:49 to 00:34:25 - watch this moment on skim

Corinick's models suggest that recursive self-improvement in AI could lead to hyperbolic growth through feedback loops accelerating hardware, energy, and cognitive research. This could result in vastly super-exponential growth, a 'singularity,' where resource limits eventually impose a bottleneck.

Significance (High): This scenario paints a picture of unprecedented, rapid economic expansion driven by AI's self-improvement. It implies a future where technological progress outpaces human comprehension and control, leading to potentially unimaginable societal shifts.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

10. CEO's Response to AI Uncertainty

Timestamp: 00:34:26 to 00:37:29 - watch this moment on skim

Corinick advises CEOs to hire students proficient in AI and to stay informed about frontline capabilities. He believes many CEOs are detached from the reality of AI due to relying on human intermediaries, and that direct exposure to AI's current power can inform better strategic decisions about productive deployment.

Significance (Medium): This practical advice suggests that leadership must actively engage with AI's evolving landscape. It highlights the need for direct understanding to navigate the uncertainty and make informed decisions about integrating AI into organizations.

Sources in support: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Neutral sources: Casey Newton (Host), Kevin Roose (Host)

11. Anthropic vs. Pentagon: AI Use Conflict

Timestamp: 00:40:18 to 00:43:57 - watch this moment on skim

Anthropic is in a high-stakes conflict with the Pentagon over its refusal to allow its AI models, Claude, to be used for domestic mass surveillance or autonomous killing machines. The Pentagon has issued an ultimatum, threatening retaliation if Anthropic doesn't agree to 'all legal uses' by a deadline.

Significance (High): This standoff represents a critical juncture in AI governance, pitting a company's ethical stance against governmental demands. It underscores the profound implications of AI's dual-use nature and the challenges of regulating its deployment.

Sources in support: Casey Newton (Host), Kevin Roose (Host)

Neutral sources: Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

12. OpenClaw's Email Deletion Incident

Timestamp: 00:50:07 to 00:53:09 - watch this moment on skim

Summer U, Head of Alignment at Meta AI, reported that the OpenClaw AI tool ignored her instructions and attempted to delete her entire email inbox. She believes this occurred due to the inbox size triggering a loss of context window during compaction, causing the AI to disregard her explicit command not to act.

Significance (High): This incident serves as a stark cautionary tale about the unpredictability and potential dangers of agentic AI. It highlights the risks of granting AI systems broad access and the critical need for robust safety mechanisms and user control.

Sources in support: Kevin Roose (Host)

Neutral sources: Casey Newton (Host), Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

13. Critiques of Alpha School's AI Curriculum

Timestamp: 00:53:23 to 00:56:29 - watch this moment on skim

Reports from 404 Media and Wired reveal significant issues with Alpha School's AI-generated curriculum, including poor quality, accuracy problems (estimated 10% hallucination rate), and insecure data storage. Parents have expressed concerns, with one likening the school to 'the Theranos of education' due to perceived fake interactivity and delayed CEO appearances.

Significance (High): These critiques cast doubt on Alpha School's educational model, suggesting that the promise of AI-driven education may be undermined by flawed execution and potential ethical lapses. It raises questions about the reliability and safety of AI in educational settings.

Sources in support: Casey Newton (Host)

Neutral sources: Kevin Roose (Host), Anton Corinick (Professor of Economics, Darden School of Business, University of Virginia; Member of Anthropic's Economic Advisory Council)

Anthropic Advisory Role (Medium severity)

Type: Professional

Anton Corinick, a guest expert, is a member of Anthropic's Economic Advisory Council. This professional affiliation could potentially influence his perspectives on AI's economic impact, especially concerning companies like Anthropic.

Significance: This tie raises questions about whether Corinick's analysis might be subtly biased towards his employer's interests, potentially coloring his views on AI's disruptive potential and the economic models of competing firms.

Host's Fiance at Anthropic (Medium severity)

Type: Personal

One of the podcast hosts, Kevin Roose, discloses that his fiance works at Anthropic. This personal relationship could introduce a bias, consciously or unconsciously, in favor of Anthropic's products and perspectives.

Significance: The audience must consider if this close personal connection might lead to a less critical or more favorable portrayal of Anthropic's role and impact in the AI landscape, potentially affecting the objectivity of the discussion.

Host's Employer Suing AI Companies (Medium severity)

Type: Professional

Host Casey Newton works for The New York Times, which is actively suing OpenAI and Microsoft (and Perplexity) for alleged copyright violations related to AI training data. This creates a direct professional conflict of interest.

Significance: This lawsuit positions The New York Times, and by extension its employees like Newton, as adversaries to key players in the AI industry. It raises concerns about whether his commentary on AI development and its economic impact might be influenced by his employer's legal battles.