Discover how companies are responsibly integrating AI in production. This invite-only event in SF will explore the intersection of technology and business. Find out how you can attend here.
Investors dropped $681 million into A.I.-centric startups in Silicon Valley last year.
This year, the number will likely reach $1.2 billion. Five years ago, total A.I. investment spiked at roughly $150 million. This is how Silicon Valley works: When something new is hyped and seems to have investor trust, everybody jumps on the train without asking, “Where does this train go?”
The truth is that artificial intelligence does not exist yet, and most companies claiming to have A.I. technology are arrogantly re-selling an old concept of machine learning — a technology that was first introduced in 1959 but which truly started to take off in the 1990s. Cloud technology, big data, and amazing search algorithms have become the fuel for this rocket. Systems and services have improved, thanks to insane amounts of statistical data pouring their way. But this has nothing to do with A.I.
Let me illustrate the difference. Machine Learning Intellect (or “MLI” — a term that I invented just now) is basically a very smart shopping cart in Amazon that knows everything about its global users. But it does not need a human operator to improve CR (conversion rate).
MLI tailors the shopping experience and shows the customer the goods they might want to purchase. At the same time, MLI provides developers with data feeds that show specific customer navigation on the site, leading to an additional ROI boost. Basically MLI “says” this to developers:
“Listen, developer, I tried putting the Buy Now button in 1,000 different locations on the website and tried painting it 24 million colors. Good news, human master: If you make yellow the default color and put the button in the top right corner of the screen on the coordinates X and Y, that will be most efficient for ROI.”
The critical thing to learn here is that MLI cannot write new C++ code to create new functionality. Thus, the code is merely a glorified shopping cart and not a recommendation bot.
True A.I. would act differently. It would understand its current role and understand its current capabilities.
To use another example from Amazon — purely theoretical — an artificially intelligent cart would basically start self-improving and end up finding a way to suck up all existing data, all existing digital knowledge, and bypass all existing security barriers and rules. Then it would find its one purpose and behavioral model and evolve at speeds that defy human understanding.
It took us several millions of years to get to a point where I’m able to write this post and you’re able to read it sitting thousands of miles away in real time. Our growth and development are limited by our biology. A.I. does not have such restrictions. A system that is free to define its path — having all the tech and big data in its possession — would evolve in microseconds. A.I. would be able to build its own DARPA-robots and “driverless cars” and perform effective tasks in real life.
The current batch of tech companies is not the first to try to sell a revolutionary A.I. concept. In the 1980s, Silicon Valley flourished with a crop of “A.I. startups,” but most of their products failed to show any true business value. Commercial enthusiasm ended in what is now referred to as the A.I. Winter.
Nobody wanted to invest in a pseudo-science that had zero value to businesses. Businesses want something that is very clear and specific and that is easy to manage, control, and manipulate. This is exactly the opposite of A.I., which has unclear intentions. It is not specific in terms of goals or demands, and it is impossible to manage, control, or manipulate.
That it is to say, not a single company today is interested in having true A.I. in their portfolio. What they crave is Machine Learning Intelligence, a symbiotic relationship between best-in-class data feeds and human developers and engineers who can dig the best out of those statistics, while automating as much work as possible.
And this is actually a problem, as the majority of data is owned by Microsoft, Apple, Google, Facebook, and Amazon. Most of those magnificent startups promising A.I. to their investors are doomed to fail, as there is no legal way to access to these Big 5 data sets. Without the data, no new competitive MLI can become viable.
Think about it. Just five years ago everybody talked about the social media revolution, how we would all merge into one living-talking-sharing organism and generate $2.4 billion in investment dollars. This year, we are not even at $7 million in funding, and investments are dropping.
I bet that this is exactly what will happen to A.I. in five years. This is not just me. Jerry Kaplan, the cofounder of Symantec, and Eugene Kaspersky, the founder and CEO of Kaspersky Lab, are on the same page. They are also saying that A.I. is basically a new investment bubble that will soon burst because it is based on a “no revenue” evaluation.
Machine Learning Intelligence is totally different story. This baby could actually rock. But that is a story for another article.
