Is AI Overhyped Like NanoTech?
Nanotechnology was once so hyped and we cannot help compare it with what is happening with AI now. There are so many things that nano can do, but renaming projects to nano just to get funding was what happened among companies in 2000–2005.
Eg: Nano Face wash, Nano *Insert a title*
There is an explanation for this. It can be understood using the hype curve. It works according to Amara’s Law, which is a computer saying ,stating:
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
The General Curve is as shown below:
According to the Gartner Hype cycle for Artificial Intelligence 2017, AI is at the peak of Inflated Expectations so we can now expect negative publicity marking the stage of Trough of Disillusionment.
Artificial Intelligence first began in 1950 when the English Mathematician Alan Turing published a paper entitled “Computing Machinery and Intelligence”. But, the Technology trigger happened only over the last decade. We are in the stage where the mass media hype where “Data Science is the way to go”. The expectations are tremendous and we are talking about Robots being given citizenship which is good in some sense and scary as well. Andrew Ngvery recently gave a talk on how we have to move more talent to where it is most needed by training them. This shows how we are moving forward with this AI technology.
- We need Data Scientists with skills. Data Science is no more a skill, it is a way.
- Data science came long long ago when we first started to generate relations between different things. Now, it has been recognized as a separate entity because computer science boils down to applied mathematics which boils down to functions.
- Data Science is indeed very promising and a lot of funding is given to those who do it. (The pay at Goldman Sachs should say it all. It is approximately $104,578-$114,768)
- But, for something to become successful, one needs to wait for experiments to happen and results to come out. This is not the case today. We talk data almost everyday that we are busy doing shoddy work to get results out. This is not good and the prime reason why we are entering the phase of disillusionment.
Comparison with Nanotechnology
With Nanotechnology, the hype index shot very high and peaked mid 2002. It was the Data Science equivalent back then. You’d want to work there. The news was so full of Carbon Nano tubes and how the future was going to change. The news articles at that time went on and on about the miraculous properties of nano materials. But do we talk about it today? We read about it sometimes in the newspapers. That is it.
Nanotech in the mid 2002 was promising and the career prospects were great. But, an analysis showed that it could not live up to its hype because of time. It all comes down to this, doesn’t it?
In the 2005, we had talks of whether Nano is a boon or bane.
As early as in 2008, we got books on the hype of Nano tech “Nano-hype: The Truth Behind the Nanotechnology Buzz”
In 2017, we hardly hear about it but some real work is going on. Nano tech is now in the plateau of productivity. Lithium Ion batteries and startups focus (MIT’s 30 under 30 has so many people working on Nano tech and not just Data Science) on this now better but ironically they lack funding because the hype now is data science and investors run towards the hype. Nobody can help this.
Comparison with CFC Discovery
When CFC was first invented and its refrigeration properties were identified in discovered in 1928 by Thomas Midgley, he was in search of non toxic alternatives to the existing refrigerants during that time namely Ammonia and Sulfur dioxide. It caught the media and every single refrigerator used it until they found out that it destroyed the ozone layer in 1970. For thirty years no one knew the detrimental effects it had on the environment. And Funny enough it has appeared in the 30 worst discoveries by the leading TIME magazine. Now, it has been banned and we are trying to solve the problem created by the previous solution.
From the above analysis, few points are to be noted:
- We tend to provide solutions to solve problems which end up producing further problems and we end up cleaning the mess. We seem to be caught in this cycle.
- In every single problem, whether the hype led to productive output, it brought money. From the above, one can infer that “Research goes where money flows” and it is not the other way round. That’s life.
- Data Science has been carried out since the beginning of time. Just that it was named Physics, Chemistry, Maths, Biology and so on. It was interpretation of data and the science behind it. So, they named it appropriately.
- In today’s exciting world, we want to do anything with data which was not thought of before. Hence, Data Science.
- Data Science is a way and not a skill. Mechanical Engineering is a skill. People who understand this will win.
The Prominent people like Balaji Viswanathan CEO of Invento who does ML for his bots uses it, Andrew Ng Sees the need to teach it, Adam D’Angelo believes in it. The other CS giants know it. And I, a mechanical Engineering student, am contemplating about this and making sense of it.
The future looks good but this shall also pass. We are going to create solutions, create a mess, clean it up, create a mess, and the cycle will repeat.