Pleo Creators Form New Venture

Many readers were sad to see Ugobe – maker of the Pleo dinosaur robot – go, but we have good news today. John Sosoka, former CTO of Ugobe, has started a new company. GetRobo got to talk with John on the phone about his new venture. Also he shared with us some very important experience and insights about why Ugobe failed and the future of the consumer robotics market.

 

(Photo:John with Pleo)

 

Q. What have you been working on lately?

 

A. I have co-founded 2 new companies with Tyler Wilson whom I worked with at my previous ventures including Ugobe. Tyler was the head of software development at Ugobe and the architect and driving force behind the Pleo Development Kit (PDK). By the way, the PDK has just been released as an alpha by Innvo Labs, a part of Jetta (that bought Ugobe’s intellectual property). Tyler and I are very happy that all the effort that we put into the PDK will now open up the creative possibilities for Pleo fans.

 

 So, one company we started is called Vita Robotica, which was formed when I got stopped being paid by Ugobe. While I was working without getting paid to try to deal with all the issues of Ugobe, I started working on a completely new implementation of a brain for robots that interact with humans. It’s something that I really had hoped to do at Ugobe, but never got to.

 

 I am fascinated in this life-in-robotics thing. And what I’m currently most interested in is in developing robots for special needs such as medical applications and companionship in institutions. For example, there are so many interesting things that have come up in the use of therapy dogs, and I think there are a lot of areas where medications don’t work well and where you can make use of robotics. Automated diagnostics at home is another area. Companion robots could help people keep track of their health everyday so that they can identify problems before they become real problems. There are not enough products that address those areas, and I think the market is pretty wide open.

 

 So Vita Robotica is a research company that builds the platform and tools that could be used to develop robots in this area.

Q. What is the other company?

 

A. The other company we started is named Pulse-Robotics. Pulse-Robotics is a commercial company which will be creating and selling a solution by using Vita’s tools. We’re trying to keep the two companies separate because we think it’s better not to have the robot and the robot OS tied together.

 

 One of the problems with Ugobe was that LifeOS (the core of Pleo) got kind of caught up inside the company. Initially, my hope was that people in various industries could take Pleo at wholesale and then put special programs and features in them so that it could be sold to each of their markets. Of course, that never happened. Although our plan was to be very open and we tried very hard to get the community tools work, we got hung up because the business people didn’t understand why this was going to be valuable. But I still believe that this is something that could work.

 

Q. What exactly is Pulse-Robotics working on?

 

A. Pulse-Robotics’ mission is two-fold. First, Pulse-Robotics will provide concept-to-manufacturing development of social robots for companies in medical, assistive, educational, and consumer markets. The marketing, distribution and support of those robots will be handled by the industry-specific company that is contracting with us. In this role we function as a design/manufacturing firm specializing in social robots. We will be receiving an NRE (non-recurring engineering) development cost plus a per unit license fee for the commercial version of the new “brain” which is called PULSE-CORE. This business model will allow us to partner with established players in these markets to introduce revolutionary new products.

 

 The second part of Pulse-Robotics will be commercializing the PULSE-CORE tools themselves. This is our path to giving creative people in any domain the ability to develop social abilities on existing or novel robot platforms. Currently we're looking at hosting these tools online to allow us to offer developers a single, simple development space in place of the large collection of tools required to do this today.

 

 As you know we are also very active in the art community and hope to release some of our tools tailored for artists as time permits.

 

Q. Since Jetta bought all of Ugobe’s intellectual property, will you have to start from scratch in developing new tools?

 

A. Yes. LifeOS was pretty well together by beginning of 2008 and then I couldn’t get any more development done on it after that because we (Ugobe management) kept switching. Oh we’re going to be in the cell phone industry, oh in this other industry, etc. So I had lots of time to think about what was important and what needed to be done differently and then no place to put it. So that’s kind of where I am in now. I’ve identified the things that I didn’t like about the system and I’m working on new ways to do those things that will work much better.

 

 One of the things Tyler and I are working on is to build tools for robotics so that a great story-teller, a great character designer can develop a character without being able to program. It’s the same thing as in animation. If animation had to go through the programmer it will never look good. Programmers are already over tasked and they often don’t understand all the challenges of animation and character design.

 

 Lifelike robots need character and they will not become true if people work the way they do now - which is programmers trying to integrate the stuff from character designers (non-programmers). The character designers need tools of their own. It’s really hard to get robotics people and programmers to think that way, because the problems that you’re trying to solve are really different from the way robotics people tend to work. That’s why you don’t see many lifelike robots.

 

 I believe it’s vital for researchers to have expressive physical platforms in order to unlock the power of situated conversational interaction. Mechanisms for integrating character and “expert” knowledge into robotic control systems are hard to create but I believe that this is the only way to bring autonomous characters to life at this point.

 

Q. So what happened at Ugobe?

 

A. The problem was Ugobe as a company was never able to settle on who the customer was. Is Pleo a super Furby, a high-end toy, a companion robot for adults, or is it just a cool robot that you can program yourself which can also be used as a platform for research? We couldn’t figure out as a company what to do, and as a result, even if we had gotten past the funding challenge, we would very likely have hit the next block.

 

 Outpouring of excitement by the early adopters was interpreted as support from mainstream customers by Ugobe. They were thinking, OK our job is to put out more product and it will be fine. Given the large amount of venture money invested in Ugobe and a hot IPO market in Hong Kong, the push was on to scale up quickly. Rapid growth drove many decisions and it focused resources in area that didn’t address the core marketing problem. The core problem was what were we going to produce and whom are we going to produce it for.

 The sad thing for me is we released Pleo in 2007 and other than some downloads, we never provided any more customer value for the rest of the time. That was very frustrating.

 

Q. How do you feel about the consumer robotics market?

 

A. Consumer robotics is a very important area and it’s going to be a huge market but how it’s going to develop is tricky. Consumer markets are tricky to begin with because you have to tell such an involved story right off the bat and go out in such quantity and reach so many people quickly.

 That’s part of the reason I’m interested in doing a venture that grows more slowly and is directed at solving more specific problems in medical and rehabilitation. Those attributes we were pushing for companionship can be used to address other markets that don’t have the barriers to entry and don’t require the extensive education that the general consumer market needs.

 So for example, there are people who want to develop kitchen robots. Try building a commercial robot that works in a small business first, get all the kinks worked out and THEN try to deal with the more fickle consumer market. That might be a better place to start.

 

Q. Do you think companion robots need to be useful in some way other than being a companion?

 

A. I believe it can be just a companion. I just think it takes time for people to get it. The reason I think it can be successful is that spending so much time around people with Pleo, I found that the people that did buy Pleo were looking for that kind of companionship - which is something between a Beta fish and a dog. People get it and when they do get it they become very attached to it. Pleo does serve that role.

 

 Realistically, many of the advances in technology in the consumer space come from entertainment - VCRs, DVDs, etc. It’s a popular misconception that all consumer products need a utility function. Historically, entertainment has proved to be a much better mechanism in introducing new technologies than utility has been.

 

(This story originally appeared on the Sept. 2009 issue of ROBOCON MAGAZINE )

Comments

[driver robot]

these robots are certiainly gonna rule us one day! =))