Anxiety and Hope in a World Full of Algorithms

Like 70% of Americans, I suffer from climate anxiety, and like many sufferers, the more time I spend on social media, the more my anxiety grows.

While funny cat videos can help temporarily quell the panic there aren’t enough of them to distract from the images of suffering animals and climate reports that pop into my feed in between. When I see a heartbreaking picture of a dying polar bear sitting on a thin piece of ice the only relief is a compilation of cats knocking stuff off of countertops, and as soon as its over I’m right back into a report about how Bitcoin’s energy consumption is equal to that of the entire country of Switzerland.

I felt utterly helpless, until an advisor told me about how significant an impact algorithms were having on climate change. Those bits of math and code that allow us to unlock an iPhone with our faces, or power self-driving cars, or make automated trades on the stock market need an enormous amount of compute power. Picture giant servers sitting in a huge warehouse, running day and night to build these algorithms. In fact, one expert predicts that if nothing changes algorithms will use 10% of the entire worlds energy by the year 2025. That energy need translates directly into C02 emissions, which translates directly into my anxiety.

But this time I didn’t feel entirely powerless, because it just so happens that I’m building technology that deals directly with algorithms. For the first time I was in a position to actually do something meaningful to reduce my climate anxiety, and help others contribute meaningfully as well.  

You see, most companies aren’t able to access existing algorithms because of security and procurement challenges, and because they’re just plain expensive. Instead, many choose to build their own, setting up those huge warehouse full of servers running all day and night to build them. According to a University of Amhurst study, the resulting C02 emissions for making a single algorithm can be equal to five times the amount produced by the entire lifespan of the average American car. Even worse, nine out of ten of them fail, according to a recent study by Gartner.

To be fair, many algorithms are truly unique to the company that’s building them, and those companies should absolutely build them internally. The majority, however, are completely transferrable, yet companies are constantly building nearly identical algorithms all the time.  

Some companies do, in fact, buy algorithms from others. There is also a community of data scientists and engineers that build and give a ay algorithms for free on the Internet in what we call the “open source community.” 

That might work for a bootstrapping startup, but large corporations trying to purchase the latest and greatest often run up against very real security and procurement delays. When they can’t get access to what they need fast enough to compete, they build their own.

But what if it were easier for companies to acquire algorithms in an open marketplace?

It happens in other industries all the time; Etsy is a marketplace for interesting craft goods (I myself bought a picture of a cat with a top hat and manacle). Jeep Wrangler as hundreds of companies building custom additions to its vehicles. Even Photoshop, the popular design software allows people to create and sell plug-ins.

Data scientists and engineers already give away thousands of algorithms for free through open source platforms. The problem is that while some of them are very good others remain incomplete or just proof of concepts, or lack the security required by large companies. We believe that with a little push in the form of a monetary incentive, those volunteer algorithm builders would be willing to take their creations all the way to the finish line. That’s what my company, gravityAI, is trying to facilitate.

Our platform wraps the algorithms insecurity and makes it easy for companies to buy and integrate them. While our marketplace may drive the creation of new algorithms, it is my hope that it will also prevent companies from building the same algorithms over and over again. If we’re successful, it will have the same effect as taking thousands of cars off the road.

But that’s not all. We also provide free access to tools and other algorithms that can do things like reduce the data needed to create them, or better predict what’s happening with the climate, or power things like smart thermostats that reduce the energy requirements for homes.

If algorithms can be part of the problem, they can also be part of the solution; we just need someone to build them.   I believe people will, and not just for the monetary potential. At least 70% of the thousands of brilliant data scientists and engineers out there experience the same anxiety I suffer from every day. Together with a little hope, some cold hard cash, and the right platform, we can make climate anxiety a thing of the past.