Hi Andra, and thanks to Rob for joining these dots. I'm definitely keen to participate in events like this in Sydney, and was planning to hold a mini-hackathon with some of my students and colleagues to prototype parts of this robot anyway. There's a large number of students at the Australian Centre for Field Robotics (ACFR) where I work who would be keen to be involved too, and I'd be happy to be a point of contact there for you.
I take a bit of an issue with the need finding approach as described in Pantofaru et al.'s paper. Not all products fulfil needs. I don't need an iPhone - I could use a normal phone, plus an MP3 player, and only check my email when I'm home or work. But the combination of technologies available in that one device make it a compelling purchase. The app store model is brilliant because it lets the iDevice become a platform which gets better over time without Apple having anything to do with it except taking their 30% cut. I don't know that I could put a finger on what part of the iPhone was the killer app (the ipod bit?) because to me it was always about the overall platform. I'm not suggesting that the approach or the paper is bad (I actually thought the paper was quite good), just that its likely not always appropriate. I think the fact that I'm struggling to come up with pain points I could solve for the average consumer indicates that it might be a non need fulfilling product. (Or it could just be a bad product idea... That is unfortunately the other possible take-away from this! ;-) And that's what a lot of the advice of people like Geoff in this thread has been about; that it's easier to start a company by solving a particular problem, rather than dreaming big about the platform that people could use to solve their problems. That leaves me with two options; get better at convincing people that my platform approach is viable, or find a suitable problem to solve with the platform, and do that first. There's some other interesting points raised by the paper, such as how to deal with catastrophic failure, and using humans as supervisory controllers. MIT has some good work in this area too, in particular what they call "human-ON-the-loop". They use this to imply that an autonomous system should be capable of operating without human interaction, but can be reconfigured or redirected by a human. Cheers, Tom On 8 April 2012 07:44, Andra Keay <andrak...@gmail.com> wrote: Thanks for sharing the discussion with us, Rob! By way of introduction: I'm a human-robot culture researcher who is exploring the evolution of robot startups while I'm in Silicon Valley. My partner Michael Harries is running a global technology accelerator, Citrix Startup Accelerator. I've founded a community "The Robot Launch Pad", with a few other roboticists and startup geeks, like Erin Rapacki and Ahmed Siddiqui, which looks like becoming the sort of useful resource that Silicon Beach is for Aust startups, but it's still early days and we're bootstrapping. We kicked off with a global cloud robotics hackathon, we're about run the Silicon Valley Robot Block Party at Stanford (huge!) and we go straight into running Robots as the vertical at MegaStartupWeekend at Microsoft. .... (skip the rest of my intro if robot cheerleading not req'd) We think that startup methodology, particularly the lean startup and customer development ideas of Eric Ries and Steve Blank, offer robotics the methods to launch a raft of agile pushes in to existing markets. We need the increasing affordability of actuators and sensors and consumer robotics to continue in conjunction with the cloud and outsourcing computation to smart phones and web services and other existing products. We will see an explosion in modular robotics platform approaches. But we won't see a robot slave in every home until these building blocks exist. We will see a robot car on the streets... actually we already do. We will see robot cars in the shops in the next year or two. The app and extension market for robots that we will use for ordinary things could be the way the next 'new' robots are developed. Or not. ... back to point. Tom made some interesting proposals for a robot that generated some thoughtful responses. Some have already pointed out that a robot is not yet a robot business and suggested that the way forward is to select a market segment and observe, then pick a problem that you can solve. MVP, test, iterate, repeat. Leave the building. Willow Garage are involved in a lot of observations to find the problem that consumers want solved (ie. urban white collar families who buy appliances). Caroline Pantofaru's recent paper is an excellent description of how to solve the question Tom posed. [1] There are a lot of companies selecting a segment of the market and developing for it, even if the robots haven't been showcased publicly. Tandy Trower's company Hoaloha Robotics is another good example. Tandy started Microsoft's first usability lab, then founded their Robotics Developer Studio and has recently left to build a socially assistive home robotics platform using affordable and available technology. But there's a lot of space for fast moving startups to piggy back on phone and cloud, to provide interfaces and extensions to existing technologies. etc. Maybe to augment the senses as Rob suggested. (and something's in the pipe for Aust for sure, Rob - let's start a new thread soon) Great to meet you all, Andra [1] Exploring the Role of Robots in Home Organization Pantofaru, Caroline., Takayama, Leila., Foote, Tully., and Soto, Bianca Proc. of Human-Robot Interaction (HRI), Boston, MA, p.327-334, (2012) On Friday, April 6, 2012 11:59:41 PM UTC-7, Tom Allen wrote: Geoff, Thanks for your comments - seriously. Despite your negative tone, this is actually the kind of feedback I'm after, because it forces me to better articulate my rebuttals. I agree that none of the items on that list solve any particularly acute pain points for the average person. I disagree that this means the concept is intrinsically a bad one. Collectively it provides utility in a number of areas - possibly any one of these areas could be solved by a piece of fixed infrastructure, perhaps even better solved, but I argue that nothing out there can do all this and more. If you want (for example) to track people's movement throughout your home, would you rather purchase one if not more sensors per room and corridor, and have the hassle of wiring them up and calibrating them, or have one robot that does it all? (Ok, one robot per floor of the home - unless you are prepared to add infrastructure to your home, in which case I'd still prefer to install one robot chair lift on the stairs, than sensors throughout the home. A chair lift is going to be faster and safer than a legged or hybrid platform, and doesn't require complex algorithms, increased payload weight and power requirements, etc.) You're forcing me to get to the underlying assumptions behind this product, which is good, because I don't think it's too hard to imagine what the MVP that validates them would look like. Assumption one is "it should cost less to do a house-wide task with a robot than to solve the same task with fixed infrastructure." Assumption two is "even if the costs are similar, people would prefer a robot to the hassle of installing infrastructure." Assumption three is "a robot can provide a platform technology which can be reconfigured to solve multiple problems, whereas fixed infrastructure is less amenable to reconfiguration." I've spent quite a lot of time researching existing home automation systems, and the history of the many, many companies that went bust trying to create that market. The general opinion seems to be that customers didn't want to install sensors and actuators throughout the home, and that doing so in bulk was expensive. In most cases, the successful (comparatively) companies focussed on automating homes being built from scratch. I believe a robot can solve both these points - no infrastructure, and lower cost. That it can also do a bunch of gimmicky, somewhat useful, or just generally entertaining things is a bonus. That its applications can be extended over time is a double bonus. Finally, you said; "What you're saying is that when you ask people what they want from a robot, they ask for something useful which can't be delivered and won't settle for a strange gimmicky platform that doesn't do _anything_ particularly compelling." Yes and no. They ask for something that can't be delivered, so I offer an alternative, and as yet I've not been very good at stating its value proposition effectively. This whole thread is me working on that... :-) Thanks for your feedback, Tom On 7 April 2012 13:45, Geoff Langdale <geoff.langd...@gmail.com> wrote What you're saying is that when you ask people what they want from a robot, they ask for something useful which can't be delivered and won't settle for a strange gimmicky platform that doesn't do _anything_ particularly compelling. Not one item on your list of 10 (or perhaps 9, assuming the Death Star plans wasn't entirely serious) (a) solves an urgent need of anyone's and/or (b) is well solved by a single mobile robotic platform rather than a wired-up house. Frankly, you need a manipulator of some kind or the mobility of trundling around the house (minus stairs - you wouldn't be able to cover more than 30% of our single-story dwelling without the ability to get up and down at least 1-2 steps) is pointless and easily replaced by fixed infrastructure. Geoff. -- You received this message because you are subscribed to the Silicon Beach Australia mailing list. Vist http://siliconbeachaustralia.org for more Forum rules 1) No lurkers! It is expected that you introduce yourself. 2) No jobs postings. You can use http://siliconbeachaustralia.org/jobs To post to this group, send email to silicon-beach-australia@googlegroups.com To unsubscribe from this group, send email to silicon-beach-australia+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/silicon-beach-australia?hl=en?hl=en -- You received this message because you are subscribed to the Silicon Beach Australia mailing list. Vist http://siliconbeachaustralia.org for more Forum rules 1) No lurkers! It is expected that you introduce yourself. 2) No jobs postings. You can use http://siliconbeachaustralia.org/jobs To post to this group, send email to silicon-beach-australia@googlegroups.com To unsubscribe from this group, send email to silicon-beach-australia+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/silicon-beach-australia?hl=en?hl=en -- You received this message because you are subscribed to the Silicon Beach Australia mailing list. Vist http://siliconbeachaustralia.org for more Forum rules 1) No lurkers! It is expected that you introduce yourself. 2) No jobs postings. You can use http://siliconbeachaustralia.org/jobs To post to this group, send email to silicon-beach-australia@googlegroups.com To unsubscribe from this group, send email to silicon-beach-australia+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/silicon-beach-australia?hl=en?hl=en
