Fail fast, fail often, learn quickly: that’s the secret to winning the fight against climate change

In December 2015, the world came together for the Paris Agreement, a global pact to reduce the threat of climate change. The agreement was a breakthrough in international political collaboration, with close to every single country pledging to take steps to lower greenhouse gas emissions. To achieve this goal, the Paris Agreement asks each country to:

“undertake rapid reductions thereafter in accordance with best available science, to achieve a balance between anthropogenic emissions by sources and removals by sinks of greenhouse gases in the second half of this century.”

In other words, the signatories of the agreement set the goal as (1) dramatically reducing CO2 emissions within the next 30-50 years, and simultaneously (2) pulling existing CO2 out of the atmosphere. But with today’s technology, the former is difficult, and the latter is practically impossible.

Let’s start by looking at the option for pulling existing CO2 out of the atmosphere. Unfortunately, the rate of CO2 captured by natural sinks such as the ocean and plant life is already saturated. And at present, there are no practical, manmade technologies to remove gigatons of carbon from the atmosphere. There’s a reason for this: the inviolable laws of physics dictate that it takes lots of energy to concentrate and remove CO2 from the atmosphere – it is akin to unscrambling an egg.

So even our best ideas for dealing with this problem often assume a free energy source, which is a bit ironic given CO2 is a byproduct of generating energy in the first place. For example, take one idea, called bioenergy with carbon capture and sequestration. This involves burning plants (naturally fed by “free” sunlight) for power, but capturing and storing the emitted CO2 before it can escape the smokestack. This approach requires an unfathomably large area of arable land and water before it can make a dent in the Earth’s carbon budget. Absent new technological breakthroughs to make this practical, our children may face the choice between roasting and starving.

Rapidly reducing carbon emissions to zero is the only other option. Today, the dominant renewable or zero-carbon technologies are still hydroelectric power and nuclear power. Newcomers solar and wind have massive growth potential but still need revolutions in energy storage and energy system technologies (such as smart grids) to distribute their volatile and intermittent power outputs.

Either way, we need technological breakthroughs for the goals of the Paris Agreement to become a reality.

Unfortunately, energy transitions – for example, the amount of time it takes for a new energy source to move from 5% of the market to 25-30% – have historically taken many decades. It took crude oil and natural gas 30 years each to rise to significant shares of the total energy supply in the United States. Globally, such energy transitions are even slower, taking about half a century.

But we don’t have that long. So how can we accelerate breakthroughs in energy technology?

Let’s look at an example of a breathtakingly rapid innovation from the Fourth Industrial Revolution that involves both hardware and software – self-driving cars.

Just 12 years ago, the Defense Advanced Research Projects Agency (DARPA) of the United States government held their first ever Grand Challenge, a driverless car competition with a $1-million-dollar prize.

The driving course was in the Mojave Desert, with no other vehicles on the road, no traffic signals to obey, and no pedestrians to avoid. Fifteen teams participated in the final race. The top-performing robot car, engineered by the Red Team from Carnegie Mellon University, navigated less than 5% of the course before it snagged onto an obstacle and its wheels caught fire. Other teams were thwarted even earlier by obstacles such as a football-sized rock, wild bushes and fences. In short, all participants failed spectacularly.

Instead of quitting, the participating teams took what they learned and innovated. And instead of cutting back, DARPA announced the day after the disastrous first competition that they would hold a second competition and further doubled the prize money.

Only 18 months later, when DARPA held the competition again, 5 teams completed the full course, with the Stanford Racing Team taking the prize.

Today, Google’s fully autonomous self-driving car has logged more than 1 million miles. Many of the engineers and technology behind this achievement trace directly back to the teams and their efforts in the DARPA Grand Challenge.

DARPA and the teams they brought together were both willing to try many, many approaches and were not afraid of failure. Paradoxically, the failures of the first competition were a great triumph for innovation.

Technological breakthroughs are difficult, and even more so in energy. The upshot is that energy breakthroughs are transformational for human civilization. The Second Industrial Revolution was centred around the use of electricity, which took off after the invention of electromagnetic generators such as dynamos and alternators. And steam power famously kicked off the First Industrial Revolution.

We are now in the era of the Fourth Industrial Revolution, but about 80% of the world’s electricity is still powered by steam turbines driving electromagnetic generators. Thus, a true technological breakthrough in electricity generation would require inventions that compare to the First and Second Industrial Revolutions combined. That the steam turbine is still the dominant generation technology after 130 years may be why many leaders call for an energy miracle – the phrase hints at the difficulty of the endeavour.

My company, Modern Electron, is developing technology to revolutionize thermal electricity generation. But ultimately, electricity is just one contributor to the global energy and climate dilemma. We need hundreds of companies attacking energy from many different new angles. These efforts will need to be pushed in conjunction with innovative R&D at research laboratories and public-private partnerships.

Governments also need to change how they innovate, and we the people need to change our expectations of governance. For example, one of the advantages of the United States is that there are 50 states each free to try different ideas – the equivalent to A/B testing for statecraft. So instead of taking a monolithic approach, governments in the 21st century may instead be more productive by becoming a platform for many competing programmes. Governments could even learn from venture capital, where investors hunt for unicorn companies which earn 1,000X payoffs that offset the multitudes of other failures.

We need to make it politically acceptable for governments to undertake risky initiatives, even if many of them initially fail, like DARPA’s first ever Grand Challenge on the self-driving car. Maybe then policy-makers will feel free to innovate.

The author is a Global Shaper in the Seattle Hub. The Annual Curators Meeting of the Global Shapers community is taking place in Geneva, Switzerland from 19 to 22 August.