The Human Genome Project, a $3 billion project that took place from 1990 to 2003, sequenced the human genome for the first time. With significant advances in computing capabilities, the cost of producing a high-quality ‘draft’ human genome sequence has drastically decreased, costing only $1,500 in late 2015 (Wetterstrand 2020). This is clearly an enormous decrease from the initial cost of $3 billion. High-performance computing (HPC), which began as a tool used by research scientists to perform complex calculations for academic projects just like the Human Genome Project, has become an integral part of the business world today. It allows companies to analyze large sets of data, calculate complex equations, and conduct other tasks that require more computational power than a traditional computer can provide. Initially, the power of HPC was available only to large corporations because purchasing a supercomputer was costly, and specialized training was required for employees to use the supercomputer effectively. To illustrate the utility of supercomputers, imagine a tree that fell across your driveway by your house. If the tree was small (a few inches in diameter), you could pick up the tree by yourself and move it. These are the kinds of tasks that a traditional computer can solve. If the tree is instead a foot or two in diameter, you are stuck. In this scenario, a supercomputer would be a chainsaw that would chop through trees eighty five yards in diameter. With computational power multiple orders of magnitude greater than a traditional computer, supercomputers are able to solve challenges on a different scale. In the endless effort to minimize costs, and for smaller firms to compete, companies developed the cluster-based system of HPC. This system interconnects multiple standard computers to provide computational capacity comparable to a supercomputer, albeit, at a much lower cost. In reference to the tree analogy, this strategy would call your neighbors and move the tree through a joint effort. One problem with both of these methods is that they require significant hardware investment that may be needed only for a project or two. This made HPC not particularly flexible, as more or less computational capability was needed for each project, which meant extra costs for all but the most computationally intensive project. To solve this problem, a cloud HPC service was developed, allowing companies to invest in less hardware and have the flexibility to purchase only the necessary computational resources for each job.

This newfound computational power hs been essential to the rise of cryptocurrencies such as Bitcoin. Only becoming popular in the last decade or so, Bitcoin has reached new heights recently, with an almost ninefold increase in the value of 1 bitcoin from March 30, 2020, to March 30, 2021 ($6,562 - $57,625)(Best 2021). While the applications of HPC in currency exchange and the stock market could be a whole other article unto itself, HPC has applications in the process of acquiring Bitcoin. Unlike traditional currencies, Bitcoin is obtained by solving sophisticated mathematical problems before others do. Like the 1849 Gold Rush, there is competition among miners for a limited supply of currency. Clearly, it is advantageous for a business to have the high-performance computing capabilities at its fingertips to make its algorithms for solving these problems quicker or capable of multitasking. Another tool used to solve these problems is an ASIC Bitcoin miner. These hardware pieces are specially designed to solve these problems, so they are usually more capable than an HPC setup. However, the only thing they can do is to mine for cryptocurrency, so they may not be worth it for everyone, especially since they do not come with a cheap price tag. Depending on the electricity cost, it can cost between six and eight thousand dollars to mine a single Bitcoin, but that price will only go up as Bitcoin becomes more scarce (Young 2020). HPC setups will not compete with ASIC miners. However they are more accessible to the public and much more versatile, making them a reasonable option for someone looking to get their feet wet rather than dive into cryptocurrency.

Among the many other applications of HPC, not many are as far-reaching as advertising. Advertising, especially on the web, has become ubiquitous in the twenty-first century. With the barrage of advertisements a user faces today, it is difficult for a company to separate itself from the rest and get the user’s attention. Figuring out how to do so provides enormous dividends because of the reach and accessibility of the internet. The data collection allows companies to make more accurate assessments of the target audience and create ads that are more likely to influence the user. However, not all data are valuable, and it is easy to be led down the wrong path by irrelevant or inaccurate data. As the noted statistician Nate Silver explains, “Every day, three times per second, we produce the equivalent of the amount of data that the Library of Congress has in its entire print collection. Most of it is…irrelevant noise. So unless you have good techniques for filtering and processing the information, you’re going to get into trouble” (Nichols 2013).

This is where HPC comes in. High performance computing can help a business at multiple stages in the ad development process. First, to expand on Nate Silver’s quote, HPC enables a business to accurately determine which variables are driving sales, the extent to which each variable does, and then which variables do not play a role in user engagement. If a user travels by a billboard ad, receives a mail promotion, or sees an ad while browsing the internet, and then purchases that product, the business wants to know how much each of these factors motivated the user to make the purchase, as well as the connection between these three variables. By looking at the activity of millions of users, the company can find patterns in the way a user may interact with each medium and possibly which medium ended up pushing them to make a purchase; or more importantly, what may move the user away from making a purchase. Next, having figured out what influences a user’s decisions, a business would like to develop a general strategy of spending their budget to effectively target as many users as possible to create the greatest profit. For large corporations, with budgets in the millions of dollars and hundreds of variables to consider, HPC is the best option for companies to find accurate results in a timely manner.

Once businesses have developed and implemented their plan, the final step is to watch what meaningful changes can be made. Data collection is a perfect evaluation tool. Again, HPC is a preferred method for collecting, cleaning, and preparing data because it allows companies to collect more data than other methods, leading to more precise insights, improving the designation, enhancement, and assessment process the next time around. Additionally, HPC gives the computing space and speed to conduct these processes simultaneously for multiple projects, clearly making a company more efficient.

Data is a valuable resource and one that can be collected in large amounts relatively quickly and painlessly. As a result, companies like Verisk Analytics and Palantir Technologies have started to spring up that make money by collecting data and selling it along with their analytical services to other companies.

Virtually any website keeps track of all kinds of ways the user interacts with their site, giving them nuanced information with various applications. For example, Twitter tracks how far down the user scrolls each time the app is used. This is not too revealing, but for a company that wants to make sure that their ads are being seen, they could pay Twitter for this information to strategically pay for ads. Clearly there are malicious possibilities with companies monitoring and collecting data on every action taken on the internet. A well-known example is the Facebook/ Cambridge Analytica scandal. Under the guise of a personality quiz, Cambridge Analytica acquired data on tens of millions of users’ data. The goal was to create voting profiles of the users for targeted political ads, all of which went against the warnings of their lawyers that this may violate United States election laws. This data has many positive applications. For one, companies can see how their customers interact with their website in real-time and improve their services constantly, leading to the consumers receiving improved products.

With an ever changing technological landscape, it is a fool’s errand to try to predict what the new cutting edge technology will be in even five years. Regardless, it is safe to assume that HPC capabilities will become more and more accessible to the general public as the cost decreases, creating a new world of possible applications. While these applications will generally enhance the world, Facebook will undoubtedly not be the last company to have a massive information mismanagement scandal.

Sources

Best, Raynor d. 2021. “Bitcoin price from October 2013 to April 28, 2021(in U.S. dollars).” Statista. https://www.statista.com/statistics/326707/bitcoin-price-index/.

Nichols, Wes. 2013. “Advertising Analytics 2.0.” Harvard Business Review. https://hbr.org/2013/03/advertising-analytics-20.

Wetterstrand, Kris. 2020. “The Cost of Sequencing a Human Genome.” National Human Genome Research Institute. https://www.genome.gov/about-genomics/fact-sheets/Sequencing-Human-Genome-cost.

Young, Joseph. 2020. “Why The Actual Cost Of Mining Bitcoin Can Leave It Vulnerable To A Deep Correction.” Forbes. https://www.forbes.com/sites/youngjoseph/2020/06/07/why-the-actual-cost-of-mining-bitcoin-can-leave-it-vulnerable-to-a-deep-correction/?sh=6ed8deb56067.