Alphabet (NASDAQ:GOOG) (NASDAQ:GOOGL) is best known for its industry-leading search engine, Google. But even though this tool has seen widespread adoption across the globe, most people probably don’t consider the technology under the hood. Google Search leans on artificial intelligence to understand language and deliver more accurate results. In other words, it tries to interpret what you mean, not just what type into the search bar.
However, there’s more to Alphabet than Google and it’s still pioneering new use cases for artificial intelligence across its various businesses. In this Backstage Pass video, which aired Sept. 27, 2021, Motley Fool contributor John Bromels discusses a few ways in which Alphabet uses artificial intelligence.
John Bromels: Talking about artificial intelligence, the first name that comes to mind for me is Alphabet. Ticker symbol is GOOG and GOOGL. This is a little handy fellow. That’s actually what I look like when I’m not projected on a screen, this yellow fellow over here. We think of Google as being a search engine, as having business suites and various other things. But Google is actually doing a lot behind the scenes, Google specifically and Alphabet.
The other companies part of Alphabet in general, actually doing a lot behind the scenes, especially in artificial intelligence. To set the stage actually for a lot of modern and contemporary artificial intelligence work, you actually have to go back to 1997 when IBM‘s Deep Blue first beat International Grandmaster Gary Kasparov in Chess. This was a big deal considering that, and people don’t remember this, the year before Deep Blue had actually failed to do that. It had played Gary Kasparov and lost. It had a rematch in ’97 and was able to win. Of course Deep Blue is IBM’s thing. IBM also shocked the world in 2011 when it’s Watson defeated Ken Jennings and Brad Rutter on jeopardy. But Google actually made headlines a few years later in 2017, and this was the big prize. It’s AlphaGo computer defeated a Chinese Go Master three games in a row in 2017.
Go is so much more complex than chess in terms of the number of possible moves and the number of possible iterations that Google didn’t do what IBM did with Deep Blue. In 1997 Deep Blue, literally the programmers had Deep Blue consider every possibly move and go out to examine every single possibility and then go back and pick whichever one had the most possible mean combinations going off of that branch that it set up because Go is thousands and thousands of times more complex with so many more possible moves. I mean, I believe there’s something like trillions of possible moves or possible sequences in any given Go game. What AlphaGo, which was Google’s project did, they taught it by having it play itself. Playing itself against games to figure out what the best practices and best strategy were. Out of that process, this actually, the Chess and the Go things launched Google and Alphabet’s interest in solving these AI and machine learning problems. It would take a long time to go through all of the things that Google and Alphabet are trying to do in AI and machine learning because they have numerous projects. But I just wanted to highlight one of those.
DeepMind is the name of the Alphabet subsidiary, much like Google is an Alphabet subsidiary, DeepMind is the Alphabet subsidiary that specifically is looking at AI and machine learning. Google turn this to the question of protein folding. This is a very specific process in biotech. When a protein forms from a bunch of amino acids, it takes these amino acids, which are all shaped like wool strings wool chains and it folds them forms them into this sort of 3D structure. Much like almost like holding a piece of paper into Origami bird or other animal. The thing is, it uses many of the same immuno acids, it can form them into this structures but each structure determines what that protein can do. Think of it as you could take the same piece of paper and fold it into numerous Origami animals. The same is true with protein.
However, if your protein is a little bit off, if it gets folded just slightly wrong, it can cause a whole host of genetic issues, including things like cystic fibrosis. That’s a result of a protein that is just formed slightly differently. It can take years and the tons and tons of money and research to try to analyze a single protein’s structure and fold and be able to predict how it’s going to work in the lab.
Alphabet’s DeepMind decided to try to solve this problem in 2016, they start working on AI in 2016. In 2018, they have this program called AlphaFold, it squeaks out a win just barely in the biennial cash competition, which is a biennial competition to see if we can predict how proteins are going to form based on the amino acids that go into them essentially replicating by prediction these years and years of lab work and all of this money.
Then, just last year in 2020, DeepMind came back with AlphaFold 2, two years later, and essentially, announced this problem has been solved because AlphaFold 2 was able to replicate by simple prediction modeling. Was able to basically replicate what it takes scientists years and years to do in the lab, looking at the actual protein. It was able to predict, “Yes, this is what it’s going to look like.” This has incredible implications for biotech and for genetic work and other disease and condition treatment fields. That’s like just literally one thing that Google and Alphabet are doing out of the dozens of things that they are doing in this field.
This article represents the opinion of the writer, who may disagree with the “official” recommendation position of a Motley Fool premium advisory service. We’re motley! Questioning an investing thesis — even one of our own — helps us all think critically about investing and make decisions that help us become smarter, happier, and richer.