Neurons, evolution, 3D and artificial intelligence

Tobias Günther is a geek, a professor of computer science at DHBW Mannheim, and founder of Elaspix, a fairly successful online 3D-configurator. At the same time Tobias is quite good at table tennis, has printed the Mannheim water tower in 3D, and can drive a tank. We met up to talk about Star Wars, Mannheim, evolution and artificial intelligence.


Hi Tobias! Your company Elaspix is seated in the Mafinex business incubator. Is this a coincidence, or very much intended?

I grew up in Burg near Magdeburg and was really into playing soccer. At that time I also became quite the computer nerd. My dad was a Major in the National People’s Army of the German Democratic Republic (DDR), and this was probably the reason I decided to enlist in the military after finishing school. Due to my obedient mentality I didn’t even consider an alternative, such as civilian service.



Where were you stationed at?

Near Bremen. Which was good, as I was quickly posted as a tank driver. I was put in charge of a Marder armored infantry fighting vehicle, and was responsible for repair and maintenance of it. That went well with my interests in technical things.

And afterwards you went back to Magdeburg?

Exactly. I began studying information technology, with a focus on the fields of robotics and artificial intelligence. As a Star Trek fan that was right up my alley. We were developing robots back then that could play soccer and so forth. And after a while the public became more and more interested in these fields, which I found very pleasing. I spent my required internship at Fraunhofer Institute in Bonn in 2002, where I would also write my thesis in 2004. This was an ideal preparation for everything that would follow.



For your PhD for instance? I’ll just go ahead and ask … what was the subject of your doctorate?

Data mining in diagnostic charts.

Well … OK then!

I wrote my doctorate at the Medical Faculty in Magdeburg. It was focused on neuroscience. We studied the way our brain processes visual stimulation – how it can see. I had already done the same thing when we taught robots to play soccer, as they had to see the playing field. We exposed the test subjects to visual stimuli and measured their reactions. My job was to interpret the data and give therapy recommendations. All this was also a matter of artificial intelligence.

Which is currently being implemented into many devices …

Yes. And this can sometimes feel scary, as we think about things like Terminator, with a scenario of the human race being conquered by the machines.

Which has already happened in a way … wouldn’t you say?

Well, we can still turn the machines off at this point. Though as humans we strongly depend on systems. Our biological evolution stalled 100.000 years ago. We are stone age people in a hyper-complex world. And we live under the illusion that we are in charge of things. 200.000 years ago we thought we could tame a volcano by dancing, and today we think we can turn off these computers! However, this is not strictly true. As humans we can adapt in two main ways. Firstly, we can change our software, our thought processes. Here we remain quite adaptable and are still in a process of development. Our hardware also bears some development potential. The evolution never stopped, and our brain has remained constant in volume and connection density for the last 20.000 years.There are certain tribes in Africa that have developed a resistance to malaria. The pressure of evolution in certain regions can cause such things. Due to the narrowness of the birth canal though, our brain can not grow indefinitely.



So should we move over to Petri dishes?

It would be a possibility. But even now the energy balance of humans is extreme. A more potent, larger brain would require us to never cease feeding ourselves.

So into the Matrix then? Something we are afraid of as well.

Yes. The fear comes from the idea of being completely copied. Which we don’t like, as we like to feel unique! Aside from this our physical bodies create powerful emotions through sensory impressions. For example eating our favorite food. This kind of drive is very massive in our human psyche, and much much stronger than any will power we can muster ourselves.



And should we simulate such input?

Well, if you had an electronic body to deliver such input, then you would not have lost your physical self.

True. But that body could jump 40 feet into the air!

It is very thinkable that we will use implants and the like to improve ourselves. But I honestly think we are already there, with our many smart devices. The question of whether these devices are inside or outside of our bodies is not that important. We already function as units with them. What I find much more interesting is the fact that we know how a single neuron functions. What we don’t know though, is how billions of them work together. Once we know this we can construct an AI that is just as smart as we are.

At which point things get interesting.

Indeed. If we had reached that point, we could begin to improve the AI and would produce a mind more advanced than our own. In turn we could solve problems that were unsolvable until then. Or we might create problems that were unthinkable … If one understands how these neurons interact, and is capable of reproducing a singular neuron and measuring its activity behavior, then one can replace a neuron in ones biological mind with an artificial one. Were one to then replace the rest of them, one would become immortal, while remaining the same person.

We should get back on track here. How did you get from Magdeburg to Mannheim after living in Bonn?

My partner was working in Weinheim and I later applied for a job in Bonn. As a result we moved to Mannheim in 2008, especially for the direct train connections. In the end we really liked the city.



What do you like about Mannheim?

To me Mannheim is an Eldorado. The universities are very interesting and offer great research and teaching. When we arrived here we got great support from the city and could easily begin to create networks. In the beginning I gave lectures at the university and DHBW to make ends meet. And eventually I was giving so many lectures that I was offered the professorship.

And at the same time you founded your start-up Elaspix?

Yes. On January 1st, 2009. At first we were experimenting with cropping people from pictures and animating them. That didn’t really work that well though. In 2012 we tried ourselves in animated films, which also didn’t quite work out. And after these two pivots we started developing 3D-models for mass configuration.



For instance for product configurators on an automobile manufacturers website?

Exactly. With all the options offered today, one will quickly face tens of thousands of model options. By now we have gotten really good at this and have many clients, as it has shown that these product configurators create higher revenues. People want to know what they are ordering, before making a decision.

Aside from virtual product pictures you also dabble in 3d printing?

Very much so! We just recently printed the Mannheim water tower. Though 3D printing is not just fun and games to me. At the moment the profit is made by selling the 3D printers. But what about the applications? That area is still almost untouched. We are currently working on medals that can be designed and then ordered.

Tobias, this is all very interesting! We shall continue this conversation soon.



Interview: Paul Heesch / LA.MAG Content. Corporate. Communication.

Photos: Ricardo Wiesinger


Beitrag teilen: Share on FacebookShare on Google+Tweet about this on TwitterShare on LinkedIn

Recent Posts