Olisakwe… Homeboy on a mission to create human-sized AI machine
Kassy Timilehin Olisakwe is a 20-year-old Network security engineer (CCNA, CCNP, Comptia S+) with certification in multiple areas in IT, including Machine Learning, Deep Learning and Python programming language as well as Neuroscience. He has created three virtual Personal Assistants – Viki, Enki and T- Bot. The young IT guru processes information quite faster than usual and understands concepts without putting much or any effort into it. Using the techniques in his just completed first book titled The Guide to Artificial General Intelligence, he’s currently working on creating a human-sized physical Artificial Intelligent assistant (Robot), as well as pioneering other groundbreaking work in the field of Artificial Intelligence (AI). In this interview with MARIA DIAMOND, he spoke on Artificial Intelligence and how the technology would shape the future.
How did the passion for ICT and AI begin?
This all started after I completed my secondary school education; I was 16 at the time waiting for my JAMB exams after my WAEC. I got enrolled into the National Institute of Information Technology (NIIT). That period of my life was the best I had ever experienced as it dawned on me that my excitement around computers was not ordinary; it was a gift. I swiftly passed through all their programmes without completing most classes before going through my exams as my teachers felt I needed something more than what they could offer.
It dawned on me after I completed all their courses in the space of months, that the usual Nigerian educational system could wait; I needed to take my IT skills professionally. So, I went on to a more advanced and globally recognised I.T institute, called New Horizon. I focused on Cisco Networking certifications because they were the best in the industry.
While studying at New Horizon, I used my spare time to work as a part time network engineering intern at my former school NIIT so as to garner job experience when I eventually complete my courses. It was during this period I learnt about network security and python coding. In fact, I literally wanted to learn everything I saw or heard about.
In 2018, I took my first set of professional certifications and added them to my collection. While I was doing all these, I was also applying for full time jobs until the reality of how hard it was to get a job hit me. I actually gave up searching at a point and decided to wait out a few months before trying out Obafemi Awolowo University (OAU) Ife Post Unified Tertiary Matriculation Examination (UTME) to study Architecture since I already had a great JAMB score from the previous year; little did I know that this period of waiting would bear good tidings.
I was 18-year old at the time, and due to my eagerness to learn, I stumbled upon the topic, Artificial Intelligence and it consumed me. I read everything I could about it. Even after getting a job as a network engineer at IPxpress, an Internet service provider, I still took about seven online courses in total, and completed every single one of them. Due to the similarities between Artificial Intelligence and Neuro-science, I registered for the Neuroscience course by Duke University on Coursera, which I completed in about one month instead of the initial one-year span. By now, you would notice a trend of me starting and completing courses in the space of weeks or months, that’s because I process information a bit faster and I seem to understand concepts without putting much or any effort into it.
After consuming all I could on the cutting edge of Artificial Intelligence, I noticed some major problems in the current algorithms in use today; everything was just so wrong and we were not making any progress. There was nobody looking in the other direction, so I decided to be that person. I started jotting down, based on my intuition, what I thought an artificially intelligent system would look like. I stayed up every night for about three months, testing algorithms and building little AI assistants on my old laptop till I stumbled upon some major breakthroughs, which led to the writing of my book; of course, without my main algorithm included in it.
Right now, I am working on raising about 3,000 U.S dollars in order to kick-start my Artificially Intelligent Project. My aim is to be the first mass producer of artificial generally intelligent assistants for public use; the same way we have the Samsung and the iPhone in the Smartphone World is the same way we would have KASI (Kinetic Artificial Supreme Intelligence) my new start-up as the pioneer of this wide-field, which is projected to be worth about 325 billion dollars by 2025 and over five Trillion by 2030. The Forbes list isn’t ready for me yet; this project would mark a new beginning.
Tell us about the three virtual personal assistants you created? What are they for?
The first Artificially Intelligence (AI) assistant I created was a speech recognition assistant and I built it to replace my Google Assistant. Not because Google assistant wasn’t doing the job, but because I wanted maximum privacy and due to Google’s policies, their voice assistant sends recordings back to their headquarters, which is what I don’t fancy as an inventor, because some of my ideas are not yet patented. This assistant’s task is to type down my ideas when I say them out loud; it also assisted me in writing my book by typing out what I read to it from my personal jotter.
The second AI assistant is called T-bot, and was built as a forex trading AI. As a matter of fact, because of my interest in forex, I attended a number of forex classes here and there. However, because of my busy schedules, I never really had time to be fully involved, which is why I decided to build an AI assistant to replace me and trade in the background whenever I wanted it to, so that I could focus on other things.
The last AI assistant I built, VIKI, is a virtually integrated kinetic intelligent assistant, she helps me to do basically anything on my gadgets. She also chips in ideas from time to time; this is my closest assistant to an Algorithm (AGI), but she still has a long way to go.
What are the similarities between Python programming language and neuroscience?
The python programming language, like any other programming language, is a language we use to communicate with the computer and instruct it on how to build software or an application, and also run such an application. The python programming language itself is not related to neuroscience, only a branch of it is; that branch is machine learning.
Machine learning deals with the use and development of computer systems that are able to learn and adapt without following explicit instructions through the use of algorithms and statistical models to analyse and draw inferences from patterns in data. So, in order for a machine to learn, it has to mimic human thinking, and that is where neuroscience comes in.
Human beings think with neurons in the brain, and so in machine learning, we try to create artificial neurons and those neurons go on to learn whatever task we give them and that is when we say it has become intelligent, hence the term artificial intelligence.
What do you think is responsible for your expertise in IT and the extremely fast information processing ability? Is it raw talent or acquired skills?
I think it is a combination of both. Though I would lean more towards the raw talent because that feels like a fair advantage to begin with. Though, without acquired skills through studies and research I wouldn’t have discovered how much I didn’t know with my raw talent, and how to go about understanding them. So, I would say talent and hard work goes hand-in-hand for me.
Let’s talk about your book, The Guide to Artificial Intelligence, what informed your decision to put it together?
My book is about the creation of super-intelligent thinking machines. The first section of my book presents the overall case that intelligent thinking machines are not only possible, but also inevitable. Then, I present a model of capabilities that a system needs in order to appear intelligent. Also, the behaviours we can expect from a system built following that model.
The final section extrapolates the behaviours that result from a system created along the lines of the model of Section II, so we can reach conclusions about what such machines will be like and what we might do to coexist with them. The focus of this book is to show that it is possible for computers to be more intelligent than humans. Also, to explain why such computers are inevitable; to argue that machine intelligence will be created sooner than most people think. I also wanted to demonstrate that subsequently, vastly more powerful intelligences would be created only a few decades later. And in conclusion, that such ‘genius’ machines will lead to options and opportunities for how humans will coexist with, and prepare for them.
Tell us about the major breakthrough you stumbled on that led to the writing of your book?
In the space of AI, it is a common belief that it is impossible to achieve artificial general intelligence with the current computer hardware, and that we would only start to make serious progress in 2025. But with my algorithm, I figured out that rather than it being the hardware, it was the software that needed tweaking, and with the right software, we can have artificial general intelligence today, as opposed to the year 2025.
What is your position on Artificial Intelligence in the nearest future – dispensable or indispensable?
Artificial Intelligence is totally indispensable. AI machines are already here, and they will remain here long after human extinction. As early as the next five years, artificially intelligent machines will take over 50 per cent of the job market.
The way computers came into the picture and slowly took over our lives is the exact same way intelligent machines would. Tesla has already brought self-driving cars to the limelight, and in as early as three years we wouldn’t need human Uber drivers. In China, no physical store has a human cashier as we speak. Robots are already being trained in Law, Journalism, Surgery and many other fields in which they outperform humans by a mile. AI is here to stay and artificial general intelligence is just the next step because when AI machines get generally intelligent, they will be a thousand times smarter than humans. The advancement of AI to Artificial General Intelligence is like going from landlines to smartphones.
You are working on a human-sized general AI project, how far have you gone?
I already have the algorithm in place for my human-sized Artificially General Intelligent assistant, but this is just the software. I still have to build the hardware and I am only limited at the moment by finances. In order to portray my point clearly, I could say the software is just like the human mind, your memory, emotions and personality. While the hardware consists of your limb, flesh, bones, face and the rest of what makes you physically you. So, I can’t give a particular timeframe for when I will be done, all I can say is, as soon as I am done, the world will not remain the same.
What are the challenges you noticed in the current algorithms in use today and what’s the way forward?
AI machines today have lots of bits of intelligence, but none has any underlying ‘understanding’. AI programmes have mostly been developed to solve specific problems; they have no contact with the real world. They are always built into systems like your computer with no physical interaction. Then, after they are run, we wonder why they don’t have any real-world understanding.
Artificial General Intelligent machines will necessarily emerge in the context of robotics, as robots are the only technology based on real world interaction. Consider the self-driving car which is just a big autonomous mobile robot currently being created as narrow AI. Abstract concepts like obstacle, destination, and pedestrian will eventually need real world meanings which would be impossible within the controlled verbal-only environment of your laptop or your self-driving car. For example, once this real-world understanding emerges in various robotic areas, it will be transferred to permeate most other areas of computation.
How often do you take professional decisions based on your intuition?
I always do; intuition rules the world. It helps you breath creativity and bring your ideas into fruition. Every aspect of life deals with intuition. From the pastor on the altar receiving insights, to all the ideas scientists came up with that led us to where we are today technological and socially. One of the most important gifts from God to humanity is our imagination, creativity and intuition. That is the only thing that separates us from other species of animals. In essence, a dolphin is smarter than most human beings, but it lacks the ability to imagine and carry out specific tasks.
How do you intend to raise 3,000 US Dollars to complete your AI project?
Well, I have my book on Amazon, so I am actively promoting that. And I am reaching out to private investors as well.
What’s next after the project?
Well, after my project, I intend to use my already fine-tuned algorithm to start my company where I would kick-start the race to mass produce artificially general intelligent machines that would look like human beings with nano-coated materials that resemble human flesh. Meaning, they will be indistinguishable from you and I. They would carry out any task you lay before them with profound efficiency. Machines like this are in our future, whether I do anything about it or not, but I just seem to be at the forefront, and will lead humanity into the next era of civilisation.
Any plan to go back to the university?
Yes, after the completion of my project I would return to school to study Mechanical Engineering. I want to focus on Mechatronics at the Massachusetts Institute of Technology in the U.S in order to learn advanced robotic hardware to add to my software expertise.