The Big Mac effect in technology

We can say a lot of things about internet, but can’t say that is not responsible for making regular civilians a little geeks. After internet, no matter your age, sex, or profession, people is involved in technology. If you work in marketing, you know about technology, if you work in finance you know about technology, if you work in logistics, you know about technology.

The problem is that, as you can imagine, not all of us are experts in everything, and that’s alright. The problem is that at some point it seems that we don’t know exactly what we don’t know and the result of the lack of knowledge in technology is “believing that computers can do anything”.

The GAP between what people believe technology can do and what technology can really do is getting to a pretty bizarre situation.

The level of over promise is so ridiculous and disconnected from reality that we can call it the Big Mac effect.

 

We can can classify the “overpromise” in two different types:

1. Things that can’t be done: With the information we have today, Artificial Intelligence is a mere set of algorithms, or conditional format processes that can help us generate outputs after processing an input (data sets). Some Wikipedia collaborator added the following phrase to the definition that I found pretty clear “Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”. Mimic is a keyword here. Is not really generating a “learning or problem solving” solving, but emulating it. You can put a cat in front of a book, and even when it seems that he is reading, he is not. He is not, because he can’t. 

Computers can’t replicate Natural (human) Intelligence, at least they can’t with the information we have today. The key is the “self consciousness”. We have no idea what it is and how it works, and then, we can’t replicate it.

However a simple way of identifying what can and what can’t be done by computers in scientific terms is, we can mimic “intelligence” in those cases where we have a limited set of values each variable of the model can take, we can’t mimic intelligence in those cases the values of each analyzed variable tend to infinite. Text recognition has a limited set of values each variable can take, then is possible. Text interpretation and “understanding” tends to have infinite set of values each variable can take, then is not possible. Al least, as I said before, with the level of knowledge we have today. So it’s importante to separate between:

1. 1. R&D: There are thousands of very interesting projects working on AI, even trying to mimic “consciousness” (artificial consciousness) in labs. Most of them can’t be used in production. The main problems are:
      1. The level of certainty is low.
      2. Unable to even determine a level of error.
      3. The processing level we have today is not enough.
   2. Production: This are the ones that after being developed, tested and evaluated, where release to the market. There are no projects today that really replicate human intelligence in production stage.

As the famous phrase says, “in theory, theory and practice are the same, in practice they are not”.

2. Things that can be done, but not by the company making the offer: Business presentation are getting so extremely sophisticated that at some you don’t know if people buy the project because they really trust that is what they need or because they don’t want to say they have no idea what is it about. It’s kind of crazy when some companies present the rocket science project and nobody ask questions, it’s so obvious that nobody understands what’s all is about that at some point is like a formal deal…”if you don’t want me to lie you, don’t ask me questions”.

How can we avoid “hyperbolic proposals”? Simple, every time someone presents you a proposal just ask how are they going to make it. The more you ask, the better. You don’t need to be an expert, just ask them the questions and be sure you add that into the contract. Also, ask different players from the vendor side, how are they going to make it. Go into details, all you can.

You can say, what’s the problem with being a “little” exaggerated in a business presentation? The answer can be ethical, because is not truth. Or can be practical, there’s no need to. In any case the output is always negative. When you under delivery you generate a negative impact in your brand, ergo the purchase intention towards your company product or services and finally in your company value. Also, every failed project will negatively impact in the intention of people towards adopting new solutions, developing a “data driven culture”.

I’m not saying that we have to change the Big Mac. More than 550 million Big Macs are sold each year in the U.S. alone. But at least we can work a little on reducing the gap between the “commercial” and the “reality”.