Everest. The tallest mountain in the world standing at an incredible 29,032 feet (8,848 meters). It takes incredible strength and dedication to climb this mountain, and there is peril at every turn. But it still draws thousands of people every year. The end-goal is to get to the top of that famous mountain, but this takes months, and sometimes even years of preparation and dozens of unsung heroes to do so.
You might be asking yourself, why on earth is a taxonomy company talking about climbing mount Everest? There are a lot of similarities between the Everest Mountain climbing experience and your AI projects. AI is an extremely hot topic in not only the tech industry, but in everyday conversations. Companies who are starting an AI project may feel like they are staring Everest in the face: there is a high chance for reward, but also a high chance of failure. The stakes are serious, but so is the glory and payoff if you succeed. So how can your company rise to the top and reach that “AI Everest Summit”? Just like the climbers, there are many steps to take to train and prepare for this endeavor. Let’s first look at one of the essential pieces of an Everest climb and then compare it to your AI project.
One of the most essential, but often overlooked, heroes of the climb to Everest has to be the base camps. There are two base camps: one on the Nepal side of the mountain and one on the Tibet side. In recent years, the Nepali base camp is the only one open to foreigners, so we will focus on it. Sitting at 17,598 feet (5,364 meters), the South Base Camp hosts about 40,000 people every year.
Those who want to ascend to the highest point on the planet, must start at the base camp.
Base camps are an extremely important part of the Everest climb because they allow you to acclimate to the thinner air and train your body for the most challenging and arguably most rewarding part of the journey. The base camps themselves are like small temporary cities, with numerous outfitters offering their own amenities. Some groups have spa-like sanctuaries while others offer a simple resting place to get out of the elements and prepare for the next part of the climb. For every climber at the base camp, there are four to five workers preparing food, running activities, providing medical care, and generally caring for the people who spend time there. These base camps live at the limits of human life – any higher and life is not permanently sustainable.
You can think of the basecamps as your data models for your AI project. Probably not the most glamorous part of the AI project, but an extremely important step you must take to train your AI. If you fail to spend enough time with your data models (or acclimating at the basecamps), you run the risk of failing to reach the top and succeed in your project. (You might even find yourself facing hallucinations from the altitude sickness ;)) Trying to climb the mountain without first spending enough time with your data models will end in failure of some kind.
As is the case on the mountain, there are many kinds of accommodations or data models to choose from, each offering their own advantages and disadvantages. Which method you decide to use depends on your use case, variables, and limitations. But before you even make it to “base camp,” you have a decision to make.
There are two main routes to the South Base Camp: hiking or flying. Both begin at Lukla Airport. If you choose to hike, the trek takes about a week and covers almost 40 miles (65 kilometers) and gains about 8,000 (2440 meters) in elevation. This grueling journey takes hikers through some of the toughest terrain on the planet. The other option, however, involves taking a helicopter from Lukla Airport to the South Base Camp. This route takes a mere 35 minutes and drops you right at the foothills of the tallest mountains on earth.
The hiking route is similar to building your own data model from scratch: it requires a lot more time, preparation, and room for failure or error. Many people choose to take this route and enjoy the process of making their way to the base camp on their own two feet. Others, though, prefer to leave the hard work of getting to base camp to someone with more experience and tools (mainly, a helicopter). These people choose to use their time to acclimate and train at the base camp before summiting Everest.
In our data model example, the helicopter ride could be employing the use of something like WAND’s prebuilt Common Data Models. These Common Data Models (CDMs) get you to that base camp in a fraction of the time it would take to do so manually and guarantee a strong starting point for your AI training. With a CDM, you can more quickly and easily get your system ready to do the hard part and prepare for deployment and the imminent glory of summiting AI’s Everest in your company.
No matter what route you choose, and what kind of data model you end up using to train your AI, the journey of joining the world in AI creation and deployment is sure to be a challenging one. Some companies will succeed, others will fail. It’s up to you to put in the training and make the strategy for your AI Everest Summit.
