Automation at Work: An Interview with Hilary Mason

Hilary_Mason_SquareAs the founder of a machine intelligence research company, Hilary Mason knows a thing or two about automating her work. The Fast Forward Labs founder is the ideal co-host for our upcoming webinar exploring the technology behind our Wordsmith platform for automating the writing process.

The free January 26th webinar still has some space available, so claim your spot today!

Ahead of the webinar, we asked Hilary about other ways that automation and technology are changing the way that work gets done.

How has automation changed the way you live and work over the last few years?

Right now we have fantastic tools for organizing information and automating tasks in very specific domains. For example, TripIt is a great tool for tracking all travel-related information, warning you when a flight is delayed or a fare changes, and making sure that wherever you go, you know where you’re supposed to be and how you’re supposed to get there. When you travel a lot, this is really valuable.

Similarly, Goodreads has a wonderful book recommendation algorithm, which looks at what you’ve read and recommends other similar things you might like. Of course, I still ask for recommendations from friends, because that’s much more fun.

Does knowing that technology will continue to advance change the way you think about work? In other words, do you act differently knowing that your day-to-day processes will continue to evolve?

I generally try to automate any trivial task that I have to do at least a few times. I love writing little Python scripts to assist with common tasks like this one, which I wrote a few years ago to automatically write e-mails introducing two people (only after they have both opted-in to the introduction, of course).

When you grow up as a computer scientist you tend to look for opportunities to write code to handle things you need to do. This is because programming is much more fun than just doing these things, even if it takes the same amount of time.

Do you build time into your day/month/year to evaluate new tools? How do you carve out process improvement time from the day-to-day necessities?

I don’t think this is the kind of thing you can put on your calendar. I always have an eye out for tools and processes that can make my work more efficient.

When you’re hiring, what employee traits do you look for knowing that specific tasks and competencies could become automated in the near future?

I think machine learning research may be one of the last things to be entirely automated! That concern aside, we hire for curiosity and the ability to learn quickly. I look for people who thrive on the chance to learn something new and have the creativity and energy to experiment and try to new approaches, and who also bring a large set of intellectual tools and prior experience to their work.

Technology doesn’t usually replace human tasks on a one-to-one basis; it usually opens up new possibilities for how to work. For example, natural language generation technology doesn’t just automate repetitive writing tasks – it can create new types of writing that adds new value. Can you think of other examples where technology has given you a totally new perspective or  process?

This idea is related to the cost of doing a task. For example, if a person has to write a report, they must spend a certain number of hours per page writing that text. But if a machine can do it, it might takes second instead of hours, and suddenly you can apply that technology to areas where you previously could not justify the cost, such as writing reports for each individual person on a team.

At Fast Forward Labs, we are always looking for algorithmic capabilities that provide these orders-of-magnitude improvements. NLG is one area, and we are also looking into probabilistic methods for stream data (which provides a similar cost reduction in the CPU and memory necessary for running calculations on the data in real time).

I’m particularly excited about text summarization algorithms. It’s no big deal if you don’t have to read an entire article because you have a summary, but imagine what happens when you can get a robust summary of 10,000 articles? That’s when we’ll be giving people capabilities that we just haven’t had before. And that’s really exciting.

Are there any tasks you prefer to do manually even though they could be automated?

I like baking, even though I’m sure someone could build a wonderful cookie-making robot, and I can always buy cookies at the market. The process is fun, and it makes the results more enjoyable.

What’s a task you do every day that you’re pretty sure will be automated in three years?

I would love for the process of taking notes during meetings to be automated, and I think it’s approaching possible.

What task do you do every day that you’re pretty sure won’t be automated any time soon?

Travel! No matter how good our technology gets, there’s nothing like meeting someone face to face.

Learn more about natural language generation in our free webinar, “Natural Language Generation: New Automation and Personalization Opportunities,” on January 26th. Claim your spot today!