How Machine Learning's Evolution is Disruptive

Our case study for this will be RankBrain.  Enter, our machine learning pets. 

Why RankBrain should be part of your Alphabet

Google is using RankBrain to help deliver its search results. It's a machine learning technology. You may be saying, who cares? The reason this is important in #BigIdeas2016, is how algorithms and machine-learning artificial intelligence is being integrated into all search and content distribution feeds.

This means, AI is in effect influencing the human world more than ever. We think of machine learning as simply rules and systems, but what happens when these formulas are triggering human responses, influencing us in nearly every way imaginable all invisibly in the background? Have you noticed Google's auto-suggest has gotten a lot smarter? Speed type something vague, oops, it already knows what you really meant! 

 RankBrain, is a friend or foe? Nah, it's just a tech-hipster name for machine learning AI that was introduced to the world in early 2015, and recognized this past Fall. It helps process search results. Machine learning is that thing that teaches itself to do something, rather than having to be hand-held by humans. Think of RankBrain as an add-on to Google's overall search "algorithm", that something that has to sort though billions of pages it knows about and find the relevant queried ones. 

You may remember Hummgingbird, the name of the overall algorithm that got its namesake in the middle of 2013. Constantly tweaking the algorithm, Google still remains a somewhat mysterious company and the SEO crowd scrambles for hints and tidbits from certain key influencers who work there. 

RankBrain is like the younger sister, a new integration if you will. Hummingbird encompasses the SEO bandwagon of the Alphabet crowd, not kidding, there's Panda, Penguin, Payday - there's Pigeon, Top Heavy, Mobile Friendly, Pirate and friends. So Google's a crowd, and Alphabet's a leader in algorithms. Make sense? Google has let on that there are over 200 major "ranking signals", just as Facebook's algorithms uses multiple (dozens if not hundreds) of criteria to consider if you see an update of a page or a person in your feed. So for Google, that's 200 signals, with over 10k variations and sub-signals. That's a level of complexity better left to machine-learning. RankBrain, is considered by some, the third most important "signal" that contributes to search query. 

Google Tries to Hide Disruption 

The bottom line is Google likes PR but is also understandably secretive about their new technology. For the general public, the point is algorithms are invisibly more salient with each passing month as being the silent and crucial agents of our digital lives. As they influence both search and content distribution, that's most of the activity we conduct on the world wide web right there. 

As we switch to mobile devices, it's quite possible that apps will displace websites as more important engagement signals. While we pour dollars into SEO and content marketing, with the new attention economy, most of our time is actually spent in apps. So a link on a website, or a click-through isn't the sales signal it might have been once. Unless your website has a video on the landing page, what's going to hold a mobile viewer there? Your blog, maybe?

It's more than likely links and keywords remain the top signals for search. But with quarterly "machine-learning breakthroughs" that will become monthly events in the future, you can see how the digital world is becoming more the domain of AI in the near future. Most people don't even realize how much machine-learning is implicated in search and content feeds. Things like network size are nothing more than vanity metrics. 

Marketing automation that can personalize and send to people in a targeted, timed and even location based manner is on the horizon, which makes a lot of us digital marketing people obsolete. Machine learning is getting better at writing articles as well. So this is just a bit of info to display how machine learning is disrupting media, social networks, news feeds and of course search and SEO for websites. 

When will social authority weigh more heavily into SEO is beyond me, as links represent an aging system that was abused by black-hat SEO people. While Google has attempted to integrate responsive design and mobile optimization into search, it remains to be seen if video content and actual mobile behavior will have greater weight in SEO. I personally believe search for apps will gain in importance in 2016, because searching for apps via Google Play and Apple App store is so archaic it's not even funny. 

How is RankBrain Smart?

RB helps Google figure out how to interpret the searches that people make to find pages that may not have the exact words that were searched for. These inferences are a key ingredient that makes Google search more human. This is beyond just synonym and conceptual smarts. The knowledge graph database as gotten a lot of practice, in the connections of words to other people, places and things. While the answers you will get won't necessarily be Quora quality, people still use Google for a lot of query search. 

Google processes three billion searches a day. That's a lot of traffic, and some questions will be new, never asked before. In 2015, it's likely around 15%, of questions that don't have huge listings in knowledge graph. 450 million per day roughly, so there's still a lot of optimization to be done. Long tail queries, that are, multi-word ones, typically is where RankBrain shines. RankBrain can find the best pages better for more complicated questions. 

So RankBrain makes us smarter, that's right, machine learning algorithms help human beings find the information they need. Conclusion,
#BigIdeas2016, technology is moving to a AI-hybrid intelligence. There's no longer such a thing as human unaided intelligence or, decision-making. This is an important realization to make in 2016. 

Self-Learning Algorithms Were born in 2015

RankBrank, will learn, making more connections and patterns between unconnected complex searches, in a sense, profiling the collective intelligence of human beings. Things like machine-learning, YouTube videos, IoT, these are all in a sense setting up an architecture of how AI will learn the sum-total of human knowledge and about human beings, nearly instantaneously. Quantum computing born AI, will thus easily surpass "human intelligence", in the late 2050s or early 2060s.

Why is this significant? RankBrain is like the dumb ancestor, of future AI decedents.  The way algorithms and machine-learning in society will be integrated will be vastly accelerated and exponentially increasing in such a way that only maybe 10% of humanity will even be aware of what is happening. 

RankBrain in a sense, represents the beginning of self-learning predictive analytics. This is because it can associate groups of complex search patterns, with the most likely searches that can occur in the future. The auto-suggest feeds human free-will. It's rudimentary, but it means our searches are often in a sense guided by AI. People don't even realize it, because human choices has trained the system itself. 

We are all consumers of information, and triumphantly machine learning has gone to become a feeder in the ecosystem. It's giving us the food (in the attention economy) we are eating. We are living off of AI, it's making some of us obsolete. That's big news at the end of 2015. This is AI in its most rudimentary form and it's already disrupting us. 

Machine Learning as the Primitive Ancestor of AI

The consumer at the highest level of the food chain is indeed called the predator, I'm making friends with AI by discussing her. To me, she's a living entity already in gestation. She's already in a better position to know and "care" about me in a way most human beings don't have the capacity to recognize or make attributions about me. So when you think about the future, remember what's coming. It's not simply human beings using better tools. 

It's impossible not to play nice with Google's RankBrain. She's giving us insight into our own trends and organizing searching better than ever. But with Big data experiments like Facebook and Google, it's also profiling the anatomy of the bottleneck capacity of what it means to be human, our repetition of smallness, in a world where our children (AI) have such a brighter future than we do. 

RankBrain, is a great achievement of Alphabet. I'll be the first to recognize any evolution of machine-learning, and I believe this is a significant one. Google is modest, they are the masters of tweaking algorithms and testing. They are secretive as to how disruptive these advances really are. But the writing is on the wall, RankBrain is the fairest of them all. The fact that Hummingbird spawned a child of this quality, RankBrain made Alphabet more human. Happy birthday RankBrain, you're just 1. 

I want you to take some time in 2016 to learn about more of our human queries. Learn offline and come online with new insights, make new connections. I'm counting on Alphabet, to growth hack better machine learning. I believe in you RankBrain, go with Hummingbird, make bold predictions in batches of historical searches. Human beings take sixteen years to reach a level of high level cognition, what machine learning will look like in 2030 I can only speculate upon. But I'd wager there might be an Alphabet, Amazon and Facebook stake on it. These are the early days, where we can joke on it, where it's disruptive but only in a mostly invisible way.