AiThority Interview Series with Ronnie Vuine, CEO and Co-Founder, Micropsi Industries

Neither is Machine Learning a silver bullet solving everything, nor is AI a magical black box that will make everything somehow more intelligent

Know My Company

Tell us about your interaction with AI and other intelligent technologies that you work with, in your daily life.

Using Machine Learning, I make robots do complex tasks — sometimes literally, using our product. Beyond my professional dealings with Machine Learning, ML-enabled products play the same role in my life as in everyone else’s — which means I, too, shout commands at Alexa and get angry with her because she won’t shut up when I’m on the phone talking to a real person.

How did you start in this space? What galvanized you to start Micropsi?

Micropsi Industries comes out of a group that got interested in cognitive architectures in the early 2000s at Humboldt University in Berlin. We started to work on a framework for deploying software agents in dynamic environments called “Micropsi” back then. When Deep Learning enabled us to move out of toy environments with that framework in the mid-2010s, I said, “Let’s do it properly; let’s create a company.”

Read More: AiThority Interview Series with Dr. Alex Yang, CEO at V SYSTEMS

How do you differentiate Microspi from other Robotics Providers?

We’re a robotics software company that enables third-party industrial robots with the hand-eye coordination capabilities. For example, robotic arms from Universal Robots and other companies now have the ability do decide how to move, based on what the robot sees, in real-time, and to learn these movements from humans — without having to program. That’s pretty unique, both in terms of what it allows you to do with an industrial robot and in how we approach it.

How do you see the raging trend of including ‘AI in everything’ impacting businesses?

It’s a solid trend indeed. Yet, neither is Machine Learning a silver bullet solving everything, nor is AI a magical black box that will make everything somehow more intelligent. We now know how to do pattern matching much better and much faster than 10 years ago — that’s a big deal, pattern matching is or can be everywhere. So using it everywhere makes sense.

What are the biggest challenges and opportunities for AI companies in dealing with inflating technology prices?

Are technology prices inflating? I see awesome technology — GPUs, storage, bandwidth for instance — get cheaper and cheaper. Which is what enabled the current AI boom.

Read More: AiThority Interview Series With Scot Marcotte, Chief Technology Officer at Buck

How should young technology professionals train themselves to work better with AI, especially Robotics?

Training yourself for a more robotic future will be tricky as most robotics applications are still highly specific today, specific to the factories and to the things they make. If you want to work in the field of automation with robots, start by going to a good school.

Machine Learning is a different matter. I have long maintained that everybody who interacts with technology should learn to code. And to that I now add —  everyone should learn the basics of Machine Learning. Invest a few hours watching YouTube tutorials. And don’t trust anyone claiming “neural networks work like the human brain.”

How do you consume information on AI/ML and related topics?

I am very privileged to be working with a team that has a deep ongoing research interest in the field. So, we have access to some exciting research papers. VC’s Nathan Benaich’s newsletter is also great for industry news.

What makes deploying AI so hard?

In many cases, what I see is a misunderstanding of what it actually is. Management treats AI as a software technology trend that they need to connect to somehow. But it’s just statistics in a fancy gown. You still need a statistician who knows the math.

Some of the large corporations have gotten quite good at selling “AI cloud services” to companies who don’t have trained mathematicians to do useful things with them.

The way out, as the market matures, will be more specific, applied Machine Learning applications. We do that with our products: We package something very complicated into a very simple product that allows pretty much anyone to solve a hard computer vision and motion generation problem without having to code or understand how the algorithms work.

Read More: AiThority Interview Series with Roberta Antunes, CEO at Hack

Which is harder — choosing AI or working with it?

In a build vs. buy sense? Most Machine Learning-enabled applications promise to make things easier in some form, but finding those that actually deliver on that promise is, indeed, hard.

How potent is the human-machine intelligence for businesses and society? Who owns Machine Learning results?

Those who have the data and know how to use it. Today, we assume it’s going to be the internet giants with their centralized databases, but if data ownership moves back to the data producers, which I believe eventually will happen once we have good distributed data access technology, Machine Learning results will be traded in exchange for data access.

Where do you see AI/Machine Learning and other smart technologies heading beyond 2020?

At some point, someone is going to figure out how to actually go beyond pattern recognition and move closer to actual cognitive systems. These will be even more exciting times.

The Good, Bad and Ugly about AI that you have heard or predict?

Do you know GlaDoS (a fictional AI computer system) from (video games series) Portal? It’s all there: the good, the bad and the ugly of AI.

Read More: AiThority Interview Series with Bhaskar Roy, Head of Growth at Workato

What is your opinion on “Weaponization of AI”? How do you deal with the challenge here?

It will, of course, happen, and is, already happening. AAs with all dangerous weapons, it matters who has them. It’s important that the good guys have them. And it’s as important as ever to be able to tell who the good guys are (hint: it’s complicated). So, we deal with this challenge like we dealt with others, or should have dealt with others, before: stay ahead, stay on top of things, and never use the results to do anything you wouldn’t want done to your children.

What AI start-ups and labs are you keenly following?

Like most people in the field, we watch what comes out of the labs at Berkeley and MIT. And then recently, someone introduced me to the work of Madeline Gannon, who did a truly wonderful demo with the IRB 1200, a very good robot by ABB.

What technologies within AI and computing are you interested in?

I’d like to see someone figure out unsupervised hierarchical representation learning, and see the tops of these hierarchies be useful for symbolic reasoning. It’s really the most interesting problem.

Read More: AiThority Interview Series with Dr. Michael Green, Chief AI Officer at Blackwood Seven

What’s your smartest work-related shortcut or productivity hack?

I’m a no-shortcuts type of guy. But I urge people to go visit a tree every now and then, listen to the wind in the leaves.

Thank you, Ronnie! That was fun and hope to see you back on AiThority soon.