The International Commission on Financing Global Education Opportunity has done something remarkable: they’ve got people talking about the future of the human race.

In this striking report, The Learning Generation, the Education Commission dissects the state of education investing today. We’re investing much too little much too late, and so unevenly that we’re on track to leave billions of children on the wrong side of an species-threatening wealth and opportunity gap.

TL;DR: We’re doing education wrong. Stability of human race and health of planet Earth at risk.

Our shortsightedness is throwing the next generation under a big yellow bus. If we stay on this track, only 1 in 10 young people in low-income countries will be on track to high school success by 2030. Meanwhile, the world of work — where our kids will take the education we give them — will be transformed by a sensor-based internet of things and artificial intelligence. Half of today’s jobs will be automated by 2050. HALF.

Remarkably, The Learning Generation provides actual specifics about how to solve this before its too late. Returns on smart education investing are stunning: every dollar invested in an additional year of schooling generates $10 in benefits in low-income countries. For the 63 million kids living in conflict-affected areas, every additional year of schooling reduces an adolescent boy’s risk of becoming involved in conflict by 20%. Educating every child is one of the best ways to close the huge opportunity and wealth gaps we’ve already got, and it just might help us avert the collapse of a civilization far more connected than Rome when big wealth gaps sped its collapse.

Fortunately, the Commission’s four angles of attack on this species-threatening problem make tremendous sense:

1. Performance: student attainment and success are what we measure most,
2. Innovation: rethink how and where learning happens,
3. Finance: spend more money, earlier and smarter, and
4. Inclusion: prioritize the most disadvantaged and marginalized.

The Commission’s done a great job describing the scale of the problem and highlighting four areas where we might focus our energy first. Now its our turn. How do we start?

Learner > Factory

We’ve got to stop doing school the way we did factories a hundred years ago. Industrial age concepts like “economies of scale” dominate modern thinking about school operations and finance. These ideas have led to massive schools built for the average student. But brain research, new approaches to personalization in other fields like medicine, transportation, and good old human decency tell us there is no such thing as the average student. Yet wealthy countries keep plowing billions into making outdated factory systems suck less instead of thinking of non-factory approaches. Meanwhile, poor countries struggle to catch up by building factories of their own.

We can’t even imagine that post-2050 world yet. What if even more than 50% of jobs are automated? What if white-collar, not blue-collar jobs, are automated?

In my work training early-stage education entrepreneurs to build towards this future instead of lubricate the existing factory system, success comes from thinking big and testing small. Having something as sacred as our children at stake requires big thinking manifested in baby steps of iterative R&D.

Smart investing in education addresses the massive scale of the problem and the intimacy of the solution simultaneously, along a continuum that starts with lots of little bets and keeps investing incrementally in the best ideas until they reach their optimal size, sometimes stopping short of factory-scale.

Here’s a base-ten, log-scale number line. Each multiple represents the size of the first investment we make in a promising idea:

$1 — $10 — $100 — $1,000 — $10,000 — $100,000 — $1M — $10M — $100M

The further left your investment strategy starts, the better chance it has at reorienting education around individual learners instead of the factory.

The way we get started is by making thousands, millions of little bets. And we stick with the ones that show potential.

At my organization, 4.0 Schools, we teach education entrepreneurs to run one-day pop-up versions of new learning spaces before they do anything else. After tweaking those pop-ups, they run extended ten-student pilots, getting feedback constantly from students and families. All while keeping their day job. Only after they prove something works at this scale do they quit their job and start a full-scale learning space or tool. This iterative, high-frequency process lets us explore 100X more ideas for the price of one reckless, factory-sized bet.

Now, let me throw the question to you: What do you think are the most high-impact ways to spend money in education? What’s it going to take (more money, policy reforms, technology, health/other holistic interventions) to create good quality education systems in developing countries? How else do we get more people testing ideas? And what can the United States’ education system learn from experiments in other countries?

Check out responses from the Commission, and join the conversation!

Matt Candler is founder and CEO of 4.0 Schools, an incubator of learning spaces and tools.