Why you should stop reading the news | Bryan Caplan

News Consumption and Personal Well-being

• Rob (Host) and Rolf de Belli (author of Stop Reading the News) advocate that reducing news consumption significantly improves well-being, with Rob reporting a ~90% reduction in daily news intake for the past 10 months.
  • Rob notes that eliminating constant news feeds has reduced anxiety, stabilized mood, and increased productivity by freeing up ~10% of waking life previously spent reading headlines.
  • Brian Kaplan argues that following the news rarely provides personal utility, citing examples where alternative information sources (e.g., friends' emails, Wikipedia) would have sufficed to prevent harm or inform travel decisions.
• The media ecosystem is described as prioritizing emotional engagement over factual accuracy, creating a distorted view of reality.
  • Kaplan compares the news to a school that only reports true but overwhelmingly negative stories about a specific group, creating a false impression of prevalence and danger through selection bias.
  • The primary drivers of this bias are identified as:
    • Negativity Bias: News outlets prioritize "bad" stories (terrorism, plane crashes, scandals) because they generate clicks, despite these events being statistically rare.
    • Ideological Homogeneity: Kaplan estimates a 20:1 left-to-right ratio among US journalists, leading to an "overwhelming left-wing view" in mainstream media, even if the specific stories are factually true.
    • Sheer Innumeracy: Media focuses on vivid, low-probability risks while ignoring high-probability, low-salience issues (e.g., aging, antibiotic resistance).
• Consumption of news reinforces cognitive biases that lead to inaccurate world models.
  • Availability Bias: People overestimate the frequency of dramatic events (e.g., terrorism) because they are constantly reported, leading to poor risk assessment and policy preferences (e.g., strict building codes that skyrocket housing costs to prevent rare disasters).
  • Hindsight Bias: News coverage often simplifies complex events into neat narratives after the fact, reinforcing the illusion of predictability.
  • Oversimplification: The format encourages forming opinions on complex, distant issues without the necessary depth or data, as demonstrated by a video on South Africa that lacked statistical backing despite claiming a decline in standards.
• The social dynamics of news consumption are critiqued as a "status game" where sharing bad news confers a sense of superiority or "insider" knowledge.
  • Kaplan suggests that the "frenetic energy" of the news is a form of "negative flow" that replaces genuine human connection, which is the primary driver of happiness.
  • The "illusion of empathy" leads people to feel they have contributed to the world by merely reading about tragedies, whereas true empathy requires action or financial support.
• Kaplan supports the argument that media coverage directly incentivizes terrorism.
  • A study cited by de Belli found that increased media coverage of terrorist attacks correlates with a higher rate of subsequent attacks, suggesting terrorists act specifically to capture news attention.
  • Kaplan notes that while total bans on terrorism coverage are difficult due to the potential for underground dissemination, the lack of such bans is a failure of the media industry compared to similar voluntary restraints taken regarding teen suicide and mass killer naming.

Artificial General Intelligence (AGI) and Future Risks

• Rob (Host) holds an anxious view of AI, estimating a >10% probability of extreme positive or negative outcomes, whereas Brian Kaplan remains skeptical of rapid, transformative change.
  • Kaplan admits his view on AI capabilities shifted after GPT-4 earned an "A" on his labor economics exams (compared to a "D" on GPT-3), but he maintains that base rates suggest radical transformation is unlikely.
  • Kaplan argues that physical constraints (materials, energy, legal systems) prevent AI from being a "silver bullet" for civilization-altering tasks, regardless of its intellectual output.
• Specific risk scenarios were evaluated by Kaplan to test the threshold of concern.
  • Bioweapons: Kaplan argues that while AI may lower the barrier to entry for designing pathogens, the physical constraints of obtaining materials and the high risk of self-inflicted harm make a "solo" or "small group" bioweapons attack unlikely without state sponsorship.
  • Autonomous Agents: Kaplan assigns a <1% probability to machines having independent, unalterable goals by 2060, citing the deep-seated human psychological aversion to granting moral agency to machines and the ease of "killing" or resetting them.
  • Cyber Warfare: Kaplan views the future as an arms race between offensive and defensive AI capabilities, where defense is likely to keep pace with offense, preventing total systemic collapse.
• Kaplan challenges the "base rate" arguments used by AI optimists/pessimists regarding economic growth.
  • He rejects the idea of accelerating "eras" of growth (pre-human -> hunter-gatherer -> farming -> industrial -> information) as "math mysticism," noting that the "information age" has not yet delivered the predicted economic boom due to the "crooked timber" of human bureaucracy and regulation.
  • Kaplan argues that human bottlenecks (legal systems, cultural resistance, slow adoption of new technologies like electricity) will persist, preventing exponential growth in the near future.
• Regarding cultural upheaval, Kaplan is less concerned about AI than the Internet, arguing that wealthy populations are increasingly risk-averse.
  • He cites Steven Pinker's "Enlightenment Now" thesis, suggesting that as global wealth rises, people become less willing to die for ideological causes, reducing the likelihood of major conflicts driven by AI-induced cultural shifts.

Political Irrationality and Voter Behavior

• Kaplan's core argument in Voters as Mad Scientists is that voters suffer from "rational ignorance" and "rational irrationality," leading to systematically bad policy outcomes.
  • Rational Ignorance: The individual probability of a single vote changing an outcome is near zero; therefore, it is rational for voters to invest little time in learning policy details.
  • Rational Irrationality: Because the cost of holding false beliefs is negligible (one vote doesn't change the world), voters have no incentive to correct their biases, leading them to prioritize "what sounds good" over "what works."
• Evidence for irrationality is found in the gap between political rhetoric and betting behavior.
  • When asked to bet real money on their political predictions (e.g., migration stats, economic impacts), voters' confidence plummets, and many retract their extreme claims, revealing that their public assertions are often performative rather than factual.
  • Kaplan argues that "social desirability bias" drives voters to lie or hold beliefs that align with their group identity, even when those beliefs are demonstrably false.
• Policy failures are attributed to the mismatch between voter incentives and effective governance.
  • Universal vs. Targeted Redistribution: Governments universally redistribute wealth (inefficient) because it is more politically palatable ("sounds good") than means-tested aid, despite the latter being more effective for poverty alleviation.
  • Nuclear Power: Political pressure from news-driven fear (e.g., after accidents with no fatalities) leads to the premature decommissioning of safe technologies, despite the economic and safety benefits.
• Kaplan suggests that the path to better policy is not to demand more voter rationality, but to identify policies that are both effective and easily "sellable" as noble or emotionally appealing.
  • He notes that the "costs" of good policies (e.g., paying for regulations) are invisible and unpopular, whereas the "benefits" of bad policies (e.g., safety from rare events) are highly visible and emotionally resonant.

Specific Policy and Personal Views

• Space Resource Allocation: Kaplan advocates for a "finders keepers" system for space resources.
  • He argues that private property rights incentivize discovery and development, whereas government ownership leads to "wasteland" preservation and stagnation.
  • He rejects the idea of military-based allocation or international treaties, citing the failure of such models in Antarctica and the success of mineral rights regimes on Earth.
• Homeschooling: Kaplan, who homeschools his four children, argues that the opportunity cost is low due to efficient curriculum management.
  • He reports that his older children were motivated and spent less time on academic work than in traditional school, while the younger children receive personalized attention that accelerates learning.
  • He frames homeschooling as a "junket" that enhances family bonding (e.g., travel, shared learning) rather than a burden, though he acknowledges the difficulty of the college application process.
• AI in Education: Kaplan anticipates that Large Language Models will eventually substitute for teachers, particularly in providing personalized feedback and explaining concepts, though he notes current models still require significant fine-tuning and human oversight for broader educational contexts.