The major credit I think Jim and I deserve … is for selecting the right problem and sticking to it. It’s true that by blundering about we stumbled on gold, but the fact remains that we were looking for gold. – Francis Crick
No scientist is the same, and people become involved in science for different reasons. I am in science because I’m addicted to the feeling of being on the verge of solving a scientific mystery, discovering scientific “gold”. But as it often happens, a scientist, like any treasure hunter, hits nothing but rocks, and the months and years of explorations remain futile.
I never aspired to become a scientist—it simply happened. In my first year of university studies, I was reading in the library, having stumbled half-accidentally upon the issue of consciousness. How can the whizzing of brain cells create consciousness? How are neural processes linked to conscious perception? I was captivated by the puzzle, and before I realised it—after reading books and journal articles and carrying out my first experiments—I had become a scientist.
It all seems so simple at first. The less you know about a problem, the easier it seems to find a solution. Why couldn’t I be the one to solve the problem of consciousness? So, I grabbed a pickaxe and started on my quest to find “gold”. I ploughed vigorously in various riverbeds, occasionally coming across some yellow nuggets.
Nine years later, after submitting my doctoral thesis, frustration hit—in the nine years, I had not arrived any closer to solving the problem of consciousness. My most important contribution to research was demonstrating, with my colleagues, that the problem of consciousness cannot be solved using the existing experimental approaches to study it. The “gold” was not buried where I had been digging.
In science, it is a common practice to specialise in a discipline and delve deeper into a subject. This makes sense because after having investigated a single topic for years, it would be foolish to write off the investment and start over elsewhere. Therefore, you have only yourself to blame for the poor choice you made early on in your academic career to take up studying the enormous subject of consciousness. You made your own bed—now lie in it for the rest of your life!
Many researchers would do the right thing in this situation and go deeper. But I am hunting for “gold”: I need the feeling that I’m solving a puzzle; I need to believe that I’m closer to finding “gold”. At the very moment that I realised that I was unable to solve the problem of consciousness, it was time to move on. Where else is “gold” to be found?
You have probably heard about the feats of artificial intelligence and deep learning. Has the artificial mind already caught up with the human mind? Will it happen in the near future? In which aspects is the human mind better, and in which is it worse? How should AI be implemented to tackle practical problems? These questions have intrigued my friends and me at the Institute of Computer Science of the University of Tartu for more than six years.
Obviously, there is a lot of “gold” to be found here: AI is the basis for algorithms and technologies that will improve our lives and has now become a natural part of our daily life. But where there are vast deposits of gold, there are also many treasure hunters.
Corporate giants like Google, Microsoft, Facebook, and others have their own impressive groups studying AI, and they are not mining with pickaxes but with excavators. What can we do with our resources in a little country like Estonia? Also, foreign colleagues have often insinuated that doing research in Estonia is scientific suicide: here, in this wind-sculpted land, we can only resort to watching how others, major players, dig for “gold”. I have never agreed with their view because you need more than big machinery to get the “gold”—what matters is also where and how you dig.
In that case, what could be the best method for finding “gold” in science? To find out, I had to experience first-hand how successful “gold miners” in science operate. For that and also for family reasons, I moved to Berlin to the lab of Matthew Larkum, a treasure hunter who had repeatedly come upon scientific gold and had learned his lessons under a Nobel laureate.
To learn about “gold mining”, I once again had to change my research topic and area. From consciousness and artificial intelligence I had come to study single nerve cells and their biophysical processes. My research investigated what happens in neurons when an organism is learning something new.
At the same time, I myself learned something new about “gold mining”. What would you say is the most important thing for discovering “gold” in science? Is it order, system, thoroughly planned projects, or strict deadlines? My colleagues at the Berlin laboratory would snigger at this list because, to my astonishment and shock, there was often complete chaos in our lab. But after recovering from the initial shock, I could see the beauty in this chaos—the lab head did not guide or limit or impose any obligations, but instead enabled and enhanced. I wouldn’t go as far as to say that chaos is useful for discovering “gold”, but our lab in Berlin proved that strict order is not the most important thing.
What mattered to my supervisor was that anyone who wished had the chance of discovering “gold”. He wouldn’t force you to mine but said that not all that glitters is gold—when you come upon some yellow nuggets don’t waste your time mining those and instead move on. And when you are convinced that there is definitely “gold” to be found at site X, you can keep on mining there for years, if needed.
A friend of mine, for instance, started at the lab seven years ago, wrestled with various projects, erred and stumbled, and perhaps would have been sacked at some other lab, but in his fifth year, he had a major breakthrough and discovered a treasure. He could only do that because the lab head gracefully allowed him to leisurely plough around for seven years, three of which were spent on fruitless detours.
Another friend in Berlin decided to start cutting back on the time he spent sleeping so that at certain periods he was sleeping for a mere 3 hours a night. Even now, he tucks his children into bed every night, sleeps for 4–5 hours and at 4 in the morning drives to the lab. As our joint research project came to a halt upon my leaving Berlin, he assured me: “Don’t worry, Jaan, I will make sure that it gets finished because if I don’t, I have failed”. And failure is not an option for him. He is addicted to finding “gold”. And there is some hope that we will find “gold” together, as we have already explored this “vein of gold” that he discovered and scratched out some yellow nuggets.
Now I’m back in Estonia. I’m taking a walk at the seaside, the morning sun shining on my smooth head. I’m not resting but trying to figure out the most prospective site where a researcher like me could start digging. I have hunted “gold” for nearly 16 years, and although I have discovered little, I have learned a great deal.
There is no point of looking for “gold” in a place where it is buried too deep (such as the problem of consciousness). Also, there is no hope of discovering “gold” where major players are already mining with their big machinery (such as typical AI studies). In my travels, I have come to know the “gold” deposits of psychology, brain research, and AI—what, why, and where has something been found. What if all these “veins of gold” will eventually merge in some uncharted territory?
Jaan Aru is Senior Research Fellow in Computational Neuroscience and AI at the University of Tartu. This article was originally published in Estonian.