I’m falling into a few old patterns of not being able to apply a good filter on things I want to be working on. This is especially problematic at a place where opportunities come really fast and constantly. How do you say no to awesome, shiny bobbles that are in high supply at MIT?
I used to really be disappointed that Media Lab people rarely showed up at Dorkbot-Boston events. I’m beginning to understand why. There is just too much to do here & it’s hard to break away!
One amazing thing MIT offers is an Art Loan Program where items from the MIT collection are available for students to take home or to their offices. There’s a lottery system, though, and a few weeks ago, Lauren and I picked out 3 great pieces and crossed our fingers. Sadly, we weren’t selected, so no art for us.
I also half-way participated in art hack day – boston this past weekend. The organizers were friendly and helpful; they arranged for great food and did their best to provide a comfortable hacking environment. They went out of their way to connect teams with resources needed to complete their projects, and the end result was a set of wonderful ideas. Read More
A couple days late with this week note. I wrote a good chunk of this over the weekend, but didn’t finish editing until I realized I hadn’t posted. I backdated the posting date, at any rate.
Another week flying by at MIT. This one was all about trying to settle into a schedule and a bit of structured routine. I find that I do my best work in the early morning; when I can get myself up, my day goes so much better if I hit the ground running and get 3-4 hours of work in before breakfast. This early morning tendency is both a blessing and a curse at MIT. It seems most people tend to wander in mid-morning/lunchtime. This means if I can get into the lab super early, I can get a lot of uninterrupted thinking / making time to myself. For me, the afternoons are great for meetings, meanderings, and brainstorming work when my focused energy is low. Might be a good time to visit the gym, too, for that matter. (I have rented a locker at the MIT gym, but I couldn’t actually find it after 20 minutes of searching.) Read More
SPOILER ALERT: I reveal plot points of the movie, Robot & Frank, in this post.
Paro robot seal with elderly patient. Photo: Reuters
For my Human-Robot interaction class, I was assigned to go watch the indie movie, Robot & Frank and do a warm-up writing exercise to critique what I saw as it relates to social robotics. In the movie set in the near future, a companion / healthcare robot interacts with a man suffering from Alzheimer’s disease. Hilarity and tearjerking ensues. A particularly enjoyed the “grumpy old man” moments as played by Frank Langella.
Robots are already used for soothing, theraputic purposes, incidentally. A significant segment of the social robotics field focuses on elder care, work with autistic children, and other healthcare applications.
Design of Robot
I found the physical and behavioral design of Robot provocative. It was approximately three quarters of the size of the humans around it and humanoid, mostly (unlike the library robot which appeared to be designed to be more functional in its environment with some limited social behaviors; it could talk and noticed when its human operators entered a room). Because Robot needed to operate in typical human environments (home, store, walking on street, etc) and also needed to evoke non-threatening social connections with its patients, a diminutive humanoid seemed an appropriate choice.
Robot’s physical design seemed to have a clear influence from robots like Honda’s Asimo with a shiny plastic exterior and dome “face.” This may be a poor choice; indeed, Frank, the patient, suffering from memory loss and confusion was initially scared by Robot’s appearance by his bed, “What are you?!” and even called the robot a “spaceman.” I would have expected the design of a robot companion to be a little more comforting like the Paro Therapeutic Robots used for Alzheimer’s patients. (Question: is this a matter of uncanny valley? If the robot was any more realistic, would it have been equally scary to Frank? Should the robot have a face?)
In the movie, Robot exhibited behaviors and actions that would be far beyond any realistic robot performance today. It moved around too fast over a wide variety of terrain (including an uneven forest floor), could climb up buildings and stairs, it was very light (one character was able to lift it easily from a car trunk), and its manual dexterity seemed exaggerated. It only had a thumb and forefinger with flipper third through fifth fingers, and yet it was able to accomplish a variety of specialized tasks like lock picking and making elaborate egg breakfasts. It seemed able to adjust to new environments without any prior planning or knowledge; for example, it could find its way around Frank’s kitchen when it first arrives at his home and was able to navigate its way through a chaotic cocktail party scene while finding Frank and responding to his commands from across the room. This unrealistic portrayal of robot capabilities was necessary for the storytelling, but it may also create unrealistic expectations of current robots for audiences.
Robot’s voice was soothing, reassuring, but firm. Its movement seemed deliberate and it used subtle, non-verbal cues (like tilts of the head to indicate attention or upturned palms to emphasize points while talking). I was amused by the moments when Robot “looked off into space” as it was calculating or “thinking.”
Frank’s interaction with Robot
Frank’s physical interaction with the robot is limited. He rarely touched the robot except for overt examples like arm wrestling or transferring bags of groceries. The robot’s walking pace matched Frank’s when they were on sojourns together (although in one scene, when Frank was a little more excited, he admonished Robot to “keep” up). Matching pace and body language of a counterpart is a technique for developing rapport and trust. (Question: what work has been done in robots building rapport through mimicry of human counterparts?)
Robot’s purpose was to provide both companionship and care to its patient and would alter its behavior to that end. He routinely refused some of Frank’s commands (like serving breakfast cereal), but was also willing to negotiate with Frank if it had a belief that it could ultimately improve Frank’s health through a sequence of actions (even if those actions were, on their own, of dubious merit). It also seemed able to tell Frank knowingly incorrect information to induce guilty feelings from Frank. This would indicate a very sophisticated model of other minds in Robot’s AI; it was able to reason about Frank’s intentions and motivations.
Frank seemed to bond with his robot. The entire metaphor of the movie (Frank the alzheimer’s patient with memory issues vs. Robot the machine with a volatile memory, itself) played out in the poignant, final scene between Frank and the Robot. In order to erase the robot’s memory, Frank needed to reach behind the robot’s head and depress the off switch. This resulted in an embracing interaction (another example of touching) between Frank and the Robot, “I knew you had an off switch”
Moments like these, implied intentions of Robot, and Robot’s communicative interaction with Frank helped the movie audience empathize with a non-human character.
Robot’s Social Place
The movie attempted to address an overarching question of what it means to be human and how a robot companion might fit into the social structure of a family. I’m not convinced the characters worked this out completely. For example, the human characters never gave Robot a name; instead, they used terms like, “slave,” “butler,” “babysitter,” “companion,” “thing,” “appliance,” and “friend” to describe the robot’s relationship to the family depending on the current circumstances. It seemed that the characters had difficulty and confusion assessing the robot’s place in their social structure. The daughter felt there were ethical concerns with the robot’s place in the household whereas Frank alternatively seemed to find the robot a “friend,” “co-conspirator,” and even a “tool” to accomplish some underhanded goals (i.e. burglary).
Robot did exhibit selfless behaviors in its final act to his companion — knowingly willing to erase its memory in an attempt to protect Frank. The robot “knew” that it was not alive and did its best to assure Frank that it could not die.
I’m reviving a practice of writing weekly summaries called Weeknotes, inspired by Berg’s weeknotes. A friend sends out daily messages he calls “wilts” (short for “what I learned today”); he uses those posts to share what’s on his mind or to present interesting problems or ideas. I’m hoping my weeknotes posts function like “wi*w”… what I wondered/tinkered/learned/built/shared every week.
I get asked every now and again about how to operate a services business offering my skills for hire. I’d like to share some lessons learned the hard (and expensive) way primarily earning income from clients for the past 15 yrs. I just got into the MIT Media Lab and am hibernating my business, so this post serves as a reference for others who might consider getting into the consulting game sometime in the future.
I am beginning a transition this week. It’s been a long time coming, actually, but there will be a flurry of activity in the upcoming weeks, and I wanted to be sure to capture these moments.
I’ve been taking daily, self portrait, narcissist photos for a long while now, but this video above documents a subset from about April of this year until yesterday. This represents the time between when I was accepted to the Media Lab at MIT until now
Software Artists or Plumbers
I really can’t complain about my career so far. I’ve made some really fun stuff with some amazing people. I’ve run a studio, have worked with “name-brand” clients, and have seen stuff I’ve built making people happy. I’ve been able to feed myself (too well, actually) all while avoiding the “9-5 grind.” (granted, it’s been more like a 24/7 grind, but at least it’s a grind of my own choosing).
With some notable exceptions, though, the past couple of years have been really difficult for me, professionally. I’ve found myself hitting a ceiling where the kind of work I’m typically hired to do is no longer creatively fulfilling. When people want apps made, they want to hire plumbers, they don’t want to hire software artists. I found the most conflict with projects where the clients just wanted me to shut up and build.
I realized I’m better than that and likely smarter than they or I have given myself credit for.
I used to talk about building avant garde robot performance art, robot ballerinas and robots that played with kids in classrooms… I used to talk about that stuff A LOT.
If you had asked me, even last year, whether or not my career has moved me closer to those visions, I’d sigh and say, “no, I’m ‘just’ building iPhone apps.”
Then, one day, I actually stopped and looked at my portfolio from just the past few years and then remembered every job and startup in which I’ve participated.
Learning. Robots. Technology. Art. Design.
It’s all there. It’s always been there.
… that’s when I got excited again. That’s also when opportunity slapped me in the face.
MIT Media Lab and Personal Robots
Since many of my projects involved building apps and robots and games for kids, a friend sent over an RFP from a professor at MIT who needed some help with Unity 3D apps for kids, Dr. Cynthia Breazeal.
I knew exactly who she was, given my long-lived interest in robots that can express emotion. She is a luminary in social robotics and human robot interaction; one of her most famous projects, Kismet, is recognized as revolutionary in the way it participates in human social interaction while simulating human emotion and appearance.
I met with her and her Ph.D student, we hit it off, and I have been working as a consultant with her group, Personal Robots, at the MIT Media Lab for about a year and half. We’re engaging in app, robot, and toy development to help in early literacy development among young people. For example, we worked with OLPC to send literacy apps to remote villages in Ethiopia.
Through all this, I still kept working with my other clients and plodding along, but I started getting hooked by the audacity of the work and the constant pushing of the envelope she and her students do every day. It’s MIT, for crying out loud.
Midway through the project, I had a long conversation with Cynthia about what it might mean for me to be a student in her group. She let me in on just how big her vision is and I realized I’d been thinking way too small.
At this point, I wanted to go all-in. I wanted this to be what I worked on 24/7, very badly.
It’s just too important and there’s no way I could do it on my own.
So, I tossed my application in last December, and earlier this year, I was humbled by my acceptance, among some of the world’s most talented thinkers and makers, into the Media Lab as a Research Assistant (i.e. graduate student) and will be beginning my work there, full-time in the next week.
I have a lot more to say about this transition. I’m slightly older than the average student (but I don’t think the oldest), I have a partner of almost 8 years who was impacted by this decision (she, thankfully, has a wonderful job two T-stops away from MIT and is rooting for me). It took a lot of number crunching to build financial ROI justification (all students have their tuition covered and get a small stipend, but it’s not pretty… I’ve decided it’s easily one of the best investments I could make right now). I’m scared to death of the intellectual challenges ahead. I hate the pace of academia. I’m not sure, yet, if this is actually a short term or long term play for me. Fraud complex is going to be a constant demon in my life. I’ll talk more about those in later blog posts.
For right now, though, I’m happy and excited.
This past summer, I’ve been practicing extensive quantified self experiments; I’ve been keeping minute-by-minute logs of how I’m spending my time, what I eat, movies I watch, etc. I have a good sense of what’s difficult to capture (ex. “who is in my proximity right now) and what’s rote (ex “what time did I get up?”).
My computers take my photo and capture what’s on my screen every few minutes or so. I’m wearing a fitbit, I’m logging everything I can think about, obsessively, and have written software and prototyped tools that make this easier and automtatic… frictionless.
I haven’t done much work, yet, on what to do with this data. I’m just excited to be collecting it, fairly robustly. I have faith I’ll develop interesting ways to process this data to give my insight sometime soon.
Ultimately, I’d like to answer questions about the conditions that lend to a more prolific creative output. Sure, I’d want to a/b test and optimize for ouput (ex. “At which coffee shop and at what time of day do I tend to write my best stories?”)… to be able to see trends that make me happier and more productive is powerful.
However, more importantly, I’m interested in how the next couple of years at MIT will change me. To do that, I need baseline measurements and tracking all along the way. I need to run experiments in an environment where eccentric experimentation is pretty normal. I will walk out of the Media Lab with terrabytes of data about my body, life, and the way I think. How do you put a price on that?
Making Things and My Blog
If I look at the times in my professional life when I was happiest, it was when I am actually making things that I 100% believed in. The more I make, the happier and healthier I am (literally — I recently discovered that my weight fluctuates with my professional satisfaction).
Indeed, this yearning for “being more creative” drove me to take the next couple of years to break out of my comfort zone and experiment wildly with intellectual ideas.
I’ve written this blog on and off since 2002. About a decade. In the early first few years I wrote some epic rants and short stories and writeups of my misadventures that still make me laugh and cry today.
In 2011, I wrote just one post. One. And it was just a stupid youtube link.
While I’ve made some really great things for clients recently, it’s hard for me to list more than a few things I made strictly for myself. Like a lot of people, I have a notebook stuffed with ideas and sketches. I have a a folder on my computer filled with half-baked code prototypes.
So here’s what I want, now:
I want to live a healthy, curious life.
I want to wonder about my world and ask questions.
I want to tinker and experiment to answer those questions.
I want to make things that demonstrate what I’ve learned.
I want to share them with you, probably on this blog.
Repeat every day, until I die. Indeed. memento mori
“For 10 sweltering days in mid-June, a small army of crawling, rolling, hopping and hovering robots invaded Ft. Benning, Georgia, a sprawling training post near the Alabama border. The occasion: the U.S. Army’s Robotics Rodeo, a competitive evaluation of the latest ground-combat robots.”
Near-future setting touches on themes of “robot companions for old people” w/ a dark comedy heist twist. (Frank Langella’s character on his robot): “I need him.” “What do you need him for?” “He’s my friend.”
Also something in there re: paper books and physical libraries vs. ebooks & internet.
“In experiments at six public universities, students assigned randomly to statistics courses that relied heavily on “machine-guided learning” software — with reduced face time with instructors — did just as well, in less time, as their counterparts in traditional, instructor-centric versions of the courses.” — they’re using “robots” in the headline to describe software teaching systems
“Researchers at the University of Rochester have uncovered a new finding about how babies understand and investigate the environment around them. Dubbed the “Goldilocks effect”, their research shows that infants in the study are not interested in things that are too simple or too complex, but just right.”
I learned about Quix today. It’s a bookmarklet-driven shortcut tool for web browsing and has LOTS of potential for shortcuts and helpful information processing workflow due to its built-in and custom command features.
Here’s an example of my custom quix.txt configuration file. I expect this will evolve rapidly over time (and I’ll eventually get this into github or something)(Done!)
I use Chrome as my primary browser. Unlike safari, it does not allow for keyboard shortcuts for bookmarks or bookmark (ex. if you press cmd-1 while in Safari, the browser launches the first bookmark in the bookmark bar.
To hack this into Chrome (without using one of the clunky extensions available for this purpose), you use the Mac OS X System Preferences > Keyboard > Keyboard Shortcuts.
Instructions to set keyboard shortcut for bookmarks in Chrome
1. Go to System Preferences > Keyboard > Keyboard Shortcuts. Select Application Shortcuts and click + to add a new shortcut. Pick Google Chrome exact name for the Quix bookmarklet.
2. I choose ctrl-1, which works pretty well since I use the caps lock key as a second control modifier. You should see the keyboard short right in Chrome’s Bookmarks menu
3. Just ctrl-1 your way to productivity. An added benefit over the extension route is that the shortcut works even if the omnibar has focus.