Alexa just read me Monday’s baseball results. It was great for two reasons. First, the A’s knocked out the Mariners 9-0 in the second game of a double header. Second, it did so in Samuel L. Jackson’s voice. With Amazon I’ll assume you’re glad I focused on the latter fact for the rest of this post.
Alexa leads Toni Reid and Rohit Prasad took to the stage at Disrupt today to discuss the history, major obstacles, and future of the intelligent assistant. They also took the opportunity to announce that the highest-earning actor of all time now has his own wake-up word. After installing Samuel L. Jackson̵
I used this skill on an Echo show this week and it’s a fun addition to Alexa who swears at times. In fact, it comes with a warning upon installation that it’s not particularly family-friendly (like the real Samuel L. Jackson, he uses MFer a fair bit more than your standard voice assistant). Most of the commands are just standard Alexa stuff in Jackson’s voice, but there are some gems like asking questions about reptiles on airplanes. He’s a little less familiar with some of the deeper cuts from his filmography.
The addition of Jackson for Alexa follows Google’s own introduction of celebrity cameos from John Legend and Issa Rae. It’s clear that both companies agree that a little star power goes a long way in keeping users occupied with a virtual assistant. Frankly, Jackson is one of the best anyone can imagine, both in terms of name recognition and novelty. It’s hard to imagine too many names that could top it all (Obama? Oprah? Pee-Wee Herman?).
One of the stories Amazon wants to tell here is how surprisingly difficult it was to create a secondary wake-up word. The Samuel L. Jackson skill was previously available, but essentially required the user to say, “Hey Alexa, ask Samuel …”. Essentially a smart phone game for assistants. The novelty wears off pretty quickly to be honest.
“The Alexa wake-up word has billions of interactions every week,” says Shiv Vitaladevuni, senior machine learning manager at Alexa, in a news-linked post. “However, there was a lack of training data for the wake word ‘Hey, Samuel’. To develop the multi-wake word model for ‘Hey Samuel’ and Alexa, we had to develop new training and data modeling techniques, drawing on lessons learned from the past. “