How Cloud Power Is Improving Written English

The city of Kiev in Ukraine is the unlikely birthplace of technology that is enabling people all over the world to express themselves in written English with a level of grammatical sophistication that previously eluded countless writers, native and non-native English speakers alike.

Kiev is the home of Alex Shevchenko and Max Lytvyn, the founders of a company called Grammarly that claims to have developed the world’s best grammar checker. I had the opportunity on Friday to speak at length with Brad Hoover, Grammarly’s CEO, whose story certainly grabbed and held my attention. A venture capitalist with a background in engineering, Hoover stumbled upon Grammarly in the course of searching for a tool to automate the proofreading of his own writing, and was so blown away by it that he contacted Shevchenko and Lytvyn to pursue what he saw as a hot investment opportunity. One thing led to another, and a couple of years ago he took the helm as Grammarly’s CEO, setting up an office in San Francisco to serve as co-headquarters along with the Kiev operation.

Hoover explained that what enabled Grammarly to raise the grammar checking bar is the emergence of cloud computing — Grammarly runs on Amazon’s EC2 cloud platform. This is how he described the development effort:

There were a variety of things that happened in this space since the last time this was done, primarily by Microsoft in the 90s. One was the evolution of cloud-based computing, and the evolution of new algorithms, particularly within the [natural language processing] field. The team has succeeded in building the world’s most accurate English language spelling and grammar checker, which is significantly more accurate than competing products on the market, primarily from Microsoft and Google. … The reality is our algorithms are incredibly computationally intensive, and cannot be run on most, if any, laptops and desktops today. With the emergence of the cloud, we’re able to run much, much larger machines on demand, which allows us to run these powerful algorithms in a relatively cost-effective way.

For users, the service is cost-effective relative to hiring a proofreader, but with an annual subscription costing $140, it isn’t cheap. Hoover freely acknowledged that:

Our goal is to perfect written English. There are between 1 billion and 2 billion English writers in the world, and every one of them has a need for Grammarly’s product. The reality is we have a very high price point today, but we’re working at ways to reduce that, and in many cases, even eliminate it, so we remove what we think is the last barrier to reaching all 1 to 2 billion. … The price point is high because the development of this product is a very expensive undertaking. As you probably know, this area of computer science is one of the most cutting-edge areas. It relies on different types of expertise, and as a result we have a very sizable, very qualified, and very expensive team who is developing this. So essentially, we as a board plow all of that revenue back into building the product so that we can, in fact, accomplish the goal of perfecting written English. That’s a very ambitious goal. It’s a very, very challenging technical problem. But we’re constantly advancing our technology and building our team to do that.

I find myself rooting for Hoover and Grammarly, because their endeavor is a noble one. I’ll share more about why I drew that conclusion in a subsequent post.