Andrew Zimmern and IBM Want To Put Meat In Your Cocktail

10 BEST PODCASTS OF THE YEAR
New York Times
BEST FOOD PODCAST
James Beard Awards & Webby Awards

This episode is no longer available. Please stay tuned for more information as we work to make our archives accessible. If you have any questions, you can reach us at hello [at] sporkful [dot] com.

Andrew Zimmern and IBM Want To Put Meat In Your Cocktail

Posted by Dan Pashman

Apr 13, 2015

Andrew Zimmern and IBM Want To Put Meat In Your Cocktail

This week's episode of The Sporkful podcast is up! Listen through the player or subscribe in iTunes.

When IBM's Watson dominated on Jeopardy a few years ago, the question most people asked was, "What does IBM want to do with this thing?"

Now Watson is turning its attention to food, using its powers to identify surprising new flavors and foods.

Here’s how it works. IBM fed Watson 10,000 recipes from Bon Appetit. Watson uses those recipes to learn which flavor compounds go well together. Then it makes new suggestions for less obvious pairings. IBM and Bon Appetit test the suggestions, give it feedback, and it LEARNS.

Of course, when it comes to unusual flavors, there is one human who can go toe-to-toe with the data-crunching likes of Chef Watson: Andrew Zimmern (shown above), host of Bizarre Foods on Travel Channel. On his TV show, Andrew goes all over the world looking for new tastes. In other words, he’s sort of like the human version of Watson.

This week on The Sporkful, I sit down with Andrew and the people behind Chef Watson to discuss life at the frontiers of flavor.

IBM, Bon Appetit, and the Institute of Culinary Education (ICE) have collaborated to test and refine Chef Watson. This week they released a cookbook and an app in beta.

That's right, eaters. Chef Watson could be coming to a kitchen near you.

The cocktail section in the new cookbook exemplifies the app's sometimes bewildering culinary innovations. Take cider muddled with pancetta (shown above) -- a meaty cocktail unlike anything I've ever imagined. I have to be honest -- that's one pairing I'm not sure I would endorse.

But the good news is, Watson doesn't just process ones and zeros -- it can learn from flavor pairings that don't quite work..

"It used to be that you had to program a computer for it to do something useful," says Florian Pinel (shown above, black t-shirt), software engineer at IBM Chef Watson. "Now we have these computers that can learn and reason a little bit like humans."

But Bon Appetit's Dawn Perry (shown above with IBM's Steve Abrams), who tested the Chef Watson recipes, found that even a super-smart computer has its limitations in the kitchen.

"I joke that Watson and I are like roommates during culinary school. And the sitcom is like me and Watson and me doing a lot arm crossing and 'Waaat-son.'" she says. "That’s why I think the human is still so essential because our personal preferences do matter."

As for Andrew Zimmern -- he's convinced Chef Watson will change the way we eat for the better.

"We have such a narrow definition of what is edible," he says. "What I admire so much about Watson is I think it’s going to change our food lives."

After talking about Chef Watson, I wanted to try it out for myself. So I went to see my friend Manoush Zomorodi, host of WNYC's technology and culture podcast New Tech City.

Manoush and I (shown above) both like to eat. But while I love to cook and always go the extra mile for deliciousness, Manoush is all about mealtime efficiency. We set out to test Watson on two metrics: taste and convenience.

We rummaged through Manoush's fridge and pulled out banana, avocado, Brussels sprouts, garlic, candied ginger, and jalapenos:

Then we typed those ingredients into Watson and picked a recipe from among the options.

It involved cooking avocado.

We were skeptical, but we forged ahead -- washing, chopping, and cooking. It turned out to be a quick, easy meal to prepare.

The bad news: Chef Watson's recipe for "Irish Jalapeno Pepper Ginger Avocado Banana Sauté" (outlined below or get full details here) definitely had some bugs.