I am a student. What do I need to learn for AI?
It is a question I asked myself for some time now. Once you get interested in technology, and that is now over 30 years, you know that it never stops evolving. It is a cliché but AI and Machine Learning are again there to confront us with this reality. AI and all things around are nothing new, but thanks to the calculating power of computers, access to masses of data and new languages we suddenly come in the enrolment of what for long was only available in labs of universities and big companies.
The big companies have definitely a huge advantage as they can spend a lot of money and most of all, have a lot of data. Google is pushing toward an AI-first world. Amazon will for sure surprise everyone, and IBM, Apple, Facebook, Tesla, Intel, Salesforce, Microsoft and more of this size have the resources and the vision. The big ones are buying starting companies who showed they understand how to use AI to solve real problems or real opportunities. This gives of course a huge advantage to the deep pocket companies and a huge challenge for startups.
According to the specialists, AI is in mainstream adoption in 2 to 5 years. That is only 4 new iPhone announcements far. In Gartner’s hype cycle, which ranks technologies based on how the market perceives them, machine learning has reached the top of the hype. By the way, VR has passed acceptation and is growing towards a new sector. The AI hype is also visible in the job demands from big companies who are specifically demanding for AI specialists. For me, somebody knowing anything on AI is already a specialist as there is the query, a program language, analytics, name it. The lingo is quite new for me and I guess for most.
What leaves that to students? I still have the feeling they are at Visual Basic level in schools while the industry has totally different demands. This is nothing new, but my guess is that the gap between offer and demand has never been bigger. Besides that, it is still too much a tool without a purpose. I remember CRM, IoT, Big Data. Many buzz word compliant guru’s out there but not too much sustainable solutions for accurate problems. Always reminds me of the football supporters at the bar. When you listen to them, they should all be top trainers but the reality is that they go home after the game, most of the time with a not so steady walk.
Doing some research, I saw mathematical knowledge is foundational, so you need to get a solid grasp of probability, statistics, linear algebra, mathematical optimisation. You are developing algorithms after all.
Programs popular with AI developers include Python, Lisp, Smalltalk, Prolog, Haskell and Scala, C and C++ and Java as the known ones. There are still certainties. These are only the most mentioned, please Google for more and don’t forget that each language most of the time is there for a purpose, so first understand what you want to solve/create. You cannot paint a wall red with blue paint.
Another thing I would like to touch, and this is important for students, is that only top universities can follow the rapid paste of the evolving technology. The danger is, and already present, that only the happy few will be able to succeed. The students who are not at Harvard or MIT will need to tap into their own motivation and drive, and learn online and by practice. As a student, don’t be afraid to knock on the doors of the big companies and offer your services for any price so you can learn at the front row. You need to work after your studies for approximate 40 plus years, time enough to be ambitious.
As we are at the top of the hype, it is great to organise a Startup Weekend AI but at the same time it holds a risk. When we organise or facilitate a classic SW we see many teams with many crazy ideas. It is fun and the main thing is to validate the 9 fields of the business canvas. The focus is here always the validation and the MVP. When we go vertical we see more focus on the MVP and less on the validation. The reason is quite simple, most people who are into verticals are more specialists so before you know, you have a hybrid event between a hackathon and a SW. Not bad a priori, but personally I see less learning in how to start a startup and validate on paper first as that is cheaper than try, fail, learn, pivot in realty. I know what I am talking about.
My only thought I want to share is, do not be blinded by technology. Focus on problem/idea relation and if AI is the solution then you are good to go. And yes, I kind of pity students today because of more and more skill sets are demanded. Or maybe I am just getting old as the buzzwords exceed my old knowledge. Oh, and did I mention ethics? Next blog.
Guest blog for Startup Weekend AI Global 2017 by Patrick Bosteels, co-founder Stage-Co.
Istanbul on Sept 20-Oct 1: https://swaiistanbul17.eventbrite.com
More info on SWAI Global:https://startupweekend.org/interests/AI with 8 cities participating