What’s the biggest and most urgent problem in the world? | The 80,000 Hours career guide (2023)
- The selection of a specific problem is the single most significant driver of career impact, potentially increasing effectiveness by over 100 times compared to typical charitable efforts.
- High-impact problems are defined by three criteria: they must be large in scale, highly neglected, and solvable.
- Most charitable donations in rich countries (5% in the US) and graduate career choices (40% in teaching and health) focus on domestic issues, which the authors argue are not the most urgent.
- Resources directed at the world's poorest populations go approximately 20 times further due to the low cost of living and the high marginal utility of money in extreme poverty.
- The US spends $1.7 trillion annually on welfare compared to only $200 billion in total overseas development aid, despite there being far fewer people in poverty in the US (40 million) than globally (650 million).
- Oral rehydration therapy for diarrhea has saved 3 million lives annually over the last five decades, an achievement surpassing the combined death toll of all wars and political famines (approx. 2 million per year) in the latter half of the 20th century.
- Historical data from 1960 to 2001 shows immunizable diseases killed 5 million fewer people annually by 2001, diarrheal diseases dropped from 4.5 million to 1.5 million, malaria from 4 million to 0.5 million, and smallpox was eradicated entirely.
- Health interventions in poor countries show extreme variance in cost-effectiveness; the top interventions are over 10 times more effective than the average and 15,000 times more effective than the worst.
- Donating $100 to top global health charities (e.g., Against Malaria Foundation) can save a year of healthy life for an infant, whereas the US NHS spends over $30,000 for the same outcome, suggesting a 300-fold difference in efficiency.
- A career focused on top global health interventions could generate 100 times the impact of a career focused on typical rich-country issues when combining the 20x resource multiplier with the 5x intervention effectiveness multiplier.
- The authors decided against focusing on factory farming (which kills 100 billion animals annually) because they believe risks to future generations offer even greater potential impact.
- Future generations are considered highly neglected because they cannot vote, have no economic power, and their suffering is abstract and hard to visualize.
- The scale of impact for future generations is potentially millions of times larger than present concerns, with projections suggesting up to 1 million future generations could exist if humanity survives.
- Climate change receives tens to hundreds of billions in investment but is estimated to be less existentially risky than biosecurity threats, which kill fewer people in the present but pose a higher risk of permanent civilizational collapse.
- The estimated annual spend on fighting climate change is over $600 billion, compared to only $1–10 billion spent on biosecurity.
- A pandemic killing over 100 million people in the next century is estimated to have a probability similar to or greater than nuclear war or runaway climate change.
- Artificial Intelligence (AI) is identified as a potentially more pressing issue than biosecurity due to its rapid acceleration and extreme neglect.
- Current AI safety efforts receive minimal funding compared to the billions spent on increasing AI capabilities, with the alignment problem (ensuring AI goals match human values) remaining unsolved.
- The number of full-time researchers working directly on AI alignment is estimated at around 300, making it over 10 times more neglected than biosecurity.
- Survey data from 2022 indicates a median estimate of 2032 for a 10% chance of human-level machine intelligence, with a mean estimate of 2042.
- Experts estimate a 50% chance of human-level AI arriving by 2052, with a median estimate of 2086 for a 90% chance.
- Unlike the Industrial Revolution which automated manual labor, the AI revolution risks automating intellectual labor, creating a scenario where systems are smarter than humans in all respects.
- The "alignment problem" is characterized as complex and difficult to solve, as no one currently knows how to program moral behavior into computers.
- When uncertain about the most pressing problem, the authors recommend three "meta" strategies: global priorities research, broad interventions (like improving governance), and capacity building (promoting effective altruism).
- Giving What We Can has achieved a multiplier effect where every $1 invested led to over $9 in donations to top charities, raising over $3 billion in pledges.
- Career advice organizations can multiply impact by helping thousands of people find high-impact careers, potentially generating thousands of times more impact than working directly on a problem alone.
- The distribution of impact across problems is highly skewed; most problems are low impact, while a few select problems offer 10 to 100 times more impact.
- Personal fit is a critical variable, as working on a less pressing problem where one has exceptional skill and motivation can yield 10 times more impact than working on the top problem without fit.
- Early-career professionals should prioritize building career capital and skills relevant to their high-impact targets rather than immediately solving the most urgent problem.
- Individuals are encouraged to spend time identifying their top three personal global problems and determining what information gaps prevent them from acting on those issues.