Custodes Futurifor the advancement of humanity

Conditions that foster future pioneers

The goal is to raise the number of people who get both the ability and the access to become pioneers. That means turning the lucky encounters that saved Faraday, Ramanujan, and Einstein into systematic ones.

This page details the ideal conditions that would produce more pioneers over the long term. For near-term, actionable goals that address specific barriers, see the individual barrier pages, which are listed under each of the nine factors.

The central finding: "lost Einsteins"

Bell, Chetty, Jaravel, Petkova, and Van Reenen (2019) linked patent, tax, and school records for over a million U.S. inventors.

  • Children from top 1% income families are ten times as likely to become inventors as children from below-median families.
  • Childhood exposure to innovation has significant causal effects. These effects were gender-specific. Girls were more likely to invent in fields where more women invented locally.
  • Early ability explains little of the income gap. In the conference version of the study, third-grade math scores account for less than one-third of it (Bell et al., 2018). By eighth grade, test scores explain half, because low-income children fall behind over time.
  • The authors project that the rate of innovation in America would quadruple if women, minorities, and children from low-income families invented at the same rate as white men from top-20% income families. This is a projection, not a measurement.
  • Raising childhood exposure may raise innovation more than raising financial incentives such as tax cuts.

Ten conditions at a glance

Each label gives the overall strength of the evidence for that condition. Exceptions are listed in the detail below.

Conditions in detail

1. Biological and health conditions Biological

What the evidence shows

Early health sets limits that later opportunity cannot always fix. Raising blood lead from 2.4 to 30 µg/dL was linked to a loss of 6.9 IQ points (Lanphear et al., 2005). Low birth weight raises the risk of infant death (Maruyama and Heinesen, 2020). Wollstonecraft died of the childbed fever that Semmelweis showed was preventable. Children who die young are lost before any ability can develop.

Ideal conditions

  • Prenatal care and safe childbirth with infection control
  • Early nutrition
  • Newborn screening and treatment for genetic disorders
  • No lead or neurotoxins, and clean air and water
  • Control of malaria and other infectious diseases
  • Lifelong healthcare
  • Safety in research with new hazards

Evidence strength

Strong for most components, though some are weaker (see the strength ratings). Contested for birth weight effects on IQ and earnings.

2. Physical environment Environmental

What the evidence shows

Heat and pollution have direct, measured effects on learning. Without air conditioning, a 1°F hotter school year cut that year's learning by 1 percent (Park et al., 2020). Children who grew up with the least nearby green space had up to 55% higher risk of various psychiatric disorders, though this is correlational (Engemann et al., 2019).

Ideal conditions

  • Climate-controlled learning spaces
  • Green space and safe places to explore
  • Stable housing without chronic noise or crowding

Evidence strength

Mixed. Strong for heat and pollution, and correlational for green space.

3. Early childhood and education Educational

What the evidence shows

Early inputs strongly affect how productive later inputs are (Heckman, 2006). A 35-year follow-up of gifted students concluded that special educational opportunities enhance talent (Lubinski and Benbow, 2006). Later costs concentrate among children whose behavior problems persist. These costs trace largely to early disadvantage (Caspi et al., 2016, and Rivenbark et al., 2018). Rigid gates nearly shut out Einstein and Ramanujan despite their exceptional ability.

Ideal conditions

  • High-quality early childhood programs, especially for disadvantaged children
  • Free or low-cost books, courses, and advanced material
  • Public libraries and shared access to books and tools
  • Early, non-stigmatizing help for persistent behavior problems
  • Early talent identification and advanced-pace learning
  • Lasting support for high achievers from low-income families, because the income gap opens as low-income children fall behind (Bell et al., 2019)
  • Flexible pathways and second chances instead of single high-stakes gates
  • Affordable higher education, including access for low-income students to the most innovative colleges, where students from low- and high-income families patent at similar rates (Bell et al., 2019)

Evidence strength

Mixed. Evidence on early childhood returns is limited (the key trial is small and unreplicated, rated C). Evidence on access is strong. Evidence on gifted education models is limited (rated C for the claim that opportunities enhance talent). The concentration of later costs is rated B. The interventions that reduce it have not been reviewed.

4. Exposure, mentorship, and door-openers Social

What the evidence shows

Childhood exposure to innovation has significant causal effects. The effects were gender-specific. Girls were more likely to invent in fields where more women invented locally (Bell et al., 2019). Pioneers often reached their field through one chance contact. Hardy read a clerk's letter, Davy answered a bookbinder, and Faraday's path began at a public lecture. Bell et al. (2019) estimate that the gender gap in innovation would fall by half if girls were exposed to female inventors as often as boys are to male inventors. However, their analysis does not identify which programs work best. Franklin built a library, an academy, and a learned society for others, and he declined patents. Patents also reward invention, though, so the site does not judge which approach produces more progress overall.

Ideal conditions

  • Regular meetings with working scientists, engineers, and inventors, especially people who share the child's background, with exposure programs aimed at children from underrepresented groups who excel early in math and science
  • Mentorship that reaches beyond elite schools and families
  • Open channels for outsiders, such as review of unsolicited work
  • Public lectures and outreach
  • Pathways that pioneers build for others

Evidence strength

Mixed. Evidence is moderate to strong for exposure effects (rated B). The mentorship pattern is consistent but anecdotal. The value of institutions built by pioneers rests on historical cases only.

5. Economic conditions Economic

What the evidence shows

Economic conditions matter most through security, time, and the freedom to take risks. The best-measured economic barrier is growing up in a lower-income family. Children of top 1% families become inventors at about ten times the rate of children from below-median families (Bell et al., 2019). Whether a society's overall inequality reduces mobility by itself is contested (Corak, 2013). So the evidence most directly supports reducing childhood disadvantage. Reducing overall inequality involves value judgments the data do not settle. Funding that tolerates failure may help. HHMI investigators produced high-impact papers at a much higher rate than similar NIH grantees, though the authors caution it may not scale (Azoulay, Graff Zivin, and Manso, 2011). Support helps only if it leaves thought free. Beethoven's patrons, who asked only that he stay in Vienna, show this. So does Spinoza's refusal of a professorship that required him not to disturb the established religion. The claim that scarcity itself impairs thinking is contested and rated D (Mani et al., 2013).

Ideal conditions

  • Economic security
  • Stable income with time to think
  • Long-term funding that tolerates failure, for people rather than narrow projects
  • Multiple funding sources so one rejection is not fatal
  • A low cost of risk-taking
  • Support that does not require giving up freedom of thought

Evidence strength

Mixed. Evidence is contested for scarcity's effect on thinking (rated D). It is limited for funding structure (rated C: one unreplicated study of elite scientists). It is strong for the gap by family income (rated A for the measured gaps). It is contested for the causal effect of overall inequality (Corak rated C for causal claims). The conditions attached to support rest on historical cases only.

6. Social conditions and collaboration Social

What the evidence shows

Wuchty, Jones, and Uzzi (2007) analyzed 19.9 million papers and 2.1 million patents. They found that teams increasingly dominate nearly all fields. Teams now produce the exceptionally high-impact work that solo authors once did. Franklin's Junto and Einstein's Olympia Academy were small communities of this kind.

Ideal conditions

  • Communities for free cross-field discussion
  • Easy ways to find complementary collaborators
  • Spaces that allow chance encounters
  • Fair credit-sharing norms

Evidence strength

Mixed. Strong for the rise of teams (descriptive, rated A, from a near-complete census of papers and patents). Historical cases only for discussion communities such as the Junto.

7. The growing burden of knowledge Educational

What the evidence shows

Jones (2009) found that as knowledge builds up, age at first invention, specialization, and teamwork all rise. Bloom et al. (2020) found that research productivity is declining.

Ideal conditions

  • Faster ways to bring learners to the frontier
  • Structures that combine specialists
  • Tools that shorten the time it takes to master a field, where some researchers suggest AI may help, which is speculative

Evidence strength

Moderate. The trend is moderate to strong, and solutions are not yet established.

8. Cultural conditions Cultural

What the evidence shows

Bell et al. found that women and disadvantaged youth are as underrepresented among high-impact inventors as among inventors overall. This suggests lost opportunity rather than lower ability. Wheatley's authorship was doubted because she was an enslaved Black woman. Wollstonecraft had no access to higher education as a woman. Julian was refused positions because of his race.

Ideal conditions

  • A culture that values curiosity and questioning
  • Tolerance for nonconformity
  • Respect for evidence
  • Inclusion of historically excluded groups
  • Fair credit for contributions

Evidence strength

Moderate. The exclusion data are strong.

9. Political and institutional conditions Societal

What the evidence shows

Moser, Voena, and Waldinger (2014) found that U.S. patenting rose 31% in the fields of Jewish chemists who fled Nazi Germany. It rose mainly because those fields drew in new researchers. Institutions matter (Acemoglu and Robinson, 2012). Narcissism predicts who emerges as a leader, largely through extraversion. It does not predict who leads well, although a moderate level was linked to the most effective leaders (Grijalva et al., 2015). So choosing leaders by track record rather than confidence may protect institutions. These points describe conditions the evidence supports. They do not take a position on how to achieve them, which is often contested.

Ideal conditions

  • Peace and stability
  • Rule of law
  • Protection from persecution
  • Freedom of inquiry and expression, including freedom of science from political control
  • Openness to talent from anywhere
  • Leaders chosen for their record rather than their self-confidence
  • Broad access to opportunity
  • Public investment in education, health, and basic research

Evidence strength

Mixed. Evidence that persecution and instability drive talent away is limited (the main study covers one field and one historical episode, rated C for general claims). Evidence on leader selection is moderate (one meta-analysis, rated B). The general claim that institutions matter is weak (one book, rated C). The cost of political control of science rests on historical cases only (Vavilov under Lysenkoism). How to put these into practice is debated.

10. Psychological conditions Psychological

What the evidence shows

Self-Determination Theory finds that autonomy, competence, and relatedness support intrinsic motivation (Ryan and Deci, 2000). Brief interventions such as growth-mindset training show small effects.

Ideal conditions

  • Environments that encourage questions and self-directed exploration
  • Experiences of mastery
  • Supportive relationships
  • Norms that treat failure as learning
  • Mental health support

Evidence strength

Moderate. The supporting research is largely correlational, and mindset interventions show small effects.

A pipeline for pioneers

Life stageKey conditions
Before birth and early childhoodNutrition, healthcare, safe childbirth, newborn screening, no toxins, clean air and water, protection from malaria and other childhood diseases, stable housing without chronic noise or crowding, economic security
ChildhoodQuality early education, climate-controlled learning spaces, green space and safe places to explore, libraries, books and tools, exposure to innovators, freedom to explore, early help for persistent behavior problems, lasting support so high achievers from low-income families do not fall behind
AdolescenceAdvanced learning, mentors, flexible exams, second chances
Young adulthoodAffordable higher education, access to the most innovative colleges regardless of family income, faster ways to reach the frontier of knowledge, research communities, first jobs that leave room to think
CareerLong-term funding that tolerates failure, collaborators, teams that combine specialists, tools that shorten the time to master a field, freedom of inquiry, open channels for outsiders, support without conditions that limit thought
Throughout lifePeace, stability, inclusion, rule of law, freedom of belief and expression, openness to talent from anywhere, leaders chosen for their record, public investment in education, health, and basic research, norms that treat failure as learning, mental health support, good health, safety

Limits of this analysis

  • Survivorship bias. People with the same conditions who did not become pioneers are rarely documented.
  • Adversity is not a requirement. Bell et al.'s data suggest hardship mostly filters people out.
  • Trade-offs exist. HHMI-style funding may not scale. Resources spent one way are not available for other uses.
  • "Ideal" involves values. Evidence shows which conditions raise the odds. Priorities and funding involve value and political choices.
  • Luck will always matter. The aim is to make fortunate encounters far more common.

References cited on this page

  1. Bell, A., Chetty, R., Jaravel, X., Petkova, N., & Van Reenen, J. (2019). Who becomes an inventor in America? Quarterly Journal of Economics, 134(2), 647-713. Oxford Academic
  2. Heckman, J. J. (2006). Skill formation and the economics of investing in disadvantaged children. Science, 312(5782), 1900-1902. DOI
  3. Lubinski, D., & Benbow, C. P. (2006). Study of Mathematically Precocious Youth after 35 years. Perspectives on Psychological Science, 1(4), 316-345. DOI
  4. Mani, A., Mullainathan, S., Shafir, E., & Zhao, J. (2013). Poverty impedes cognitive function. Science, 341, 976-980. DOI
  5. Azoulay, P., Graff Zivin, J. S., & Manso, G. (2011). Incentives and creativity: evidence from the academic life sciences. RAND Journal of Economics, 42(3), 527-554. Wiley
  6. Wuchty, S., Jones, B. F., & Uzzi, B. (2007). The increasing dominance of teams in production of knowledge. Science, 316(5827), 1036-1039. DOI
  7. Jones, B. F. (2009). The burden of knowledge and the "death of the Renaissance man." Review of Economic Studies, 76(1), 283-317.
  8. Bloom, N., Jones, C. I., Van Reenen, J., & Webb, M. (2020). Are ideas getting harder to find? American Economic Review, 110(4), 1104-1144. AEA
  9. Moser, P., Voena, A., & Waldinger, F. (2014). German Jewish émigrés and US invention. American Economic Review, 104(10), 3222-3255. AEA
  10. Acemoglu, D., & Robinson, J. A. (2012). Why Nations Fail. New York: Crown.
  11. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68-78. DOI
  12. Park, R. J., Goodman, J., Hurwitz, M., & Smith, J. (2020). Heat and learning. American Economic Journal: Economic Policy, 12(2), 306-339. AEA
  13. Caspi, A., Houts, R. M., Belsky, D. W., Harrington, H., Hogan, S., Ramrakha, S., Poulton, R., & Moffitt, T. E. (2016). Childhood forecasting of a small segment of the population with large economic burden. Nature Human Behaviour, 1, 0005. DOI
  14. Rivenbark, J. G., Odgers, C. L., Caspi, A., Harrington, H., Hogan, S., Houts, R. M., Poulton, R., & Moffitt, T. E. (2018). The high societal costs of childhood conduct problems: evidence from administrative records up to age 38 in a longitudinal birth cohort. Journal of Child Psychology and Psychiatry, 59(6), 703-710. Duke Scholars
  15. Grijalva, E., Harms, P. D., Newman, D. A., Gaddis, B. H., & Fraley, R. C. (2015). Narcissism and leadership: a meta-analytic review of linear and nonlinear relationships. Personnel Psychology, 68(1), 1-47. DOI
  16. Corak, M. (2013). Income inequality, equality of opportunity, and intergenerational mobility. Journal of Economic Perspectives, 27(3), 79-102. DOI
  17. Durlauf, S. N., & Seshadri, A. (2018). Understanding the Great Gatsby Curve. In M. Eichenbaum and J. A. Parker (Eds.), NBER Macroeconomics Annual 2017, 32, 333-393. University of Chicago Press