Algorithm Creates Beer Recipes, Adjusts Them As User Feedback Comes In

“As customers drink them, they’re encouraged to visit a website and share their feedback.”

It really depends on what feedback is received. If the majority of US beer consumers took the time to provide feedback the beer would indeed be a Bud Light type beer. If the website was swamped by BAs who have extreme exuberance the beer would be a ‘NE’ style IPA. Or perhaps a Barrel Aged Imperial Stout.

Cheers!

 

A beer built by committee - what could go wrong?

 

If the feedback came from folks like me it would be a Lutefisk Nose Goblin Wild Ale with extra B.O.

#MessWithIt

 

The best idea certainly, can't allow any of that free thinking to run loose.

 

Only if there was a just single copy owned by ABInBev or another of the other mass market breweries.

But the nice thing about software is that you can have multiple copies owned by different breweries who use it only with their customer base.

So for a Wild Ale only brewery the result would be different.

 

Well since a lot of the content on this site is "built by committee" you can make an educated guess about what could go right or wrong. Or you can take a look at Wikipedia which has lots of pages "built by committee."

 

Garbage in, garbage out.

 

Do we need to watch that extreme exuberance to keep it from morphing into irrational exuberance?

 

That's where statistics and "fuzzy" algorithms would come in to play.

 

There are too many flaws with this. The first of which is assuming the drinker can determine what it is about a beer that they don't like. Many people confuse the buttery flavor of caramel malt with diacetyl. Does the algorithm know the source of the problem? One drinker might associate "hoppy" with bitterness while another associates it with hop flavor. Can it interpret what the consumer really tastes? What about aromas or flavors that would normally carry a negative connotation but are prevalent in descriptors of certain beers, for example "horse blanket" or "sour?"

Aside from this many techniques like barrel aging, continuous hopping, doing a hopback, dry hopping, using a coolship, doing a decoction mash...these were all innovated throughout history. Using an algorithm might get you closer to honing in a beer that you've already made but I seriously doubt the computer is going to spit something out that says "shoot the yeast into space and make a beer with that because it will be the next big thing" or "throw a bunch of doughnuts into that mash."

 

Why not!?!

Cheers!

 

No reason; Alan Greenspan may wish it though! As to the topic at hand, my feedback might be useless. I find something to like about most beers.

 

Beer geeks like us wouldn't suggest improvements to a beer because we know every beer is unique and we embrace the variety. Mass-market beer drinkers don't know how to suggest improvements to beer-flavored water, and they don't know that they could make it any better.

 

Plus you would never want that. Let the brewer make the beer, you might be surprised. Like ellicottville likes to make fruit beer. They have made some that added tea and made one of the best stand alone hefeweizens even thought it added blood orange. Genesee has come up with traditional German styles from their brew house and so let the guy do that. Upstate that makes the kentucky common also makes a nice dipa . =)