Last week, I learned about the manufacturing of the cutest robot toy ever, the first eye-surgery performed by a robot, the fact that robocars are not likely to happen BUT driverless buses just MIGHT, and many different opinions about robotics & jobs (see you in a hundred years to cut it short).
Robotics and AI related
No code neural network: AETROS, which would be useful if you’re not an expert in deep neural net, but have to use it anyway.
- Why is Cozmo so tiny? For a robot to be fun, it needs to move fast. It’s hard to make a robot moving slowly, exciting. And making a big robot move fast equals dangerous!
- Tiny doesn’t mean Cozmo isn’t complex. It’s made of about 300 parts…
- No Cozmo demo because “they are too many people in the room”. Oh, come on, cut the crap.
- OMG the animating team did even animate Cozmo sleeping on his charger. It’s SO DAMN CUTE.
- Cozmo can play games – and simpler the game, the better it is to show the robot’s personality.
- It was important for the founders that the robot is not seen as repetitive, whatever he’s doing. They implemented a “personality engine” allowing the robot to change its mood, with parameters like confidence, happiness, courage… Great idea, it allows for smal comportemental changes (as it’s done for us, humans). The movie “Inside Out” is a pretty good methaphor of what’s happening inside Cozmo!
- Congrats on doing a lot of testing to make Cozmo appeal on both boys and girls! No gender marketing for Cozmo, YEAH.
- “The moment you turn the robot on, and he moves and behaves, nobody’s thinking about colors“. “It doesn’t matter what color your puppy has” – tht’s true, but it’s the very first time I’m hearing something like that. I actually wanted a grey kitty, and picked up another one because of its personality… thank you for the lesson.
- Hiring animators is the KEY (it has not fallen on deaf ears, thank you). They use Maya to animate Cozmo first, and it can play live on a physical Cozmo directly (trough a magical pipeline 🙂 ). The animators learned that eye-contact is VERY important for human-robot-interaction (obvious for roboticians, but not tha much). A kid (toddler) makes eye-contact every 8 seconds, even if (s)he doesn’t need anything – Cozmo does the same 🙂
- Manufacturing: there is a lot more testing when you manufacture a robotics product than any other product. It takes one year from a fully functionnal prototype to your product on shelves.
- SDK: it’s important that the SDK can be used by non-robotics persons. Only one line of code to do a face detection! 🙂
- “Kickstarter is not funding. This is a loan on your future revenue”
Robotics might be the salvation for banks, by raising the customer satisfaction (sense of novelty, innovation…).
“Fully Autonomous Cars Are Unlikely, Says America’s Top Transportation Safety Official“. An MIT Technology Review article is always worth reading – so, click and read it please.
Altough Andrew Rosenblum raises some legit concerns, the fact that I hear more and more often people say “I don’t want to have a robocar, I genuinely LOVE to drive my own car” reminds me of the time where horses were powering wooden cars. People at this era should have said the same: “BUT I LOVE MY HORSE leave me alone please”. The economy around horses has shifted: horses are for leisure, competition, but not the primarly source of power. In my opinion, we will see the same happening with cars: driving its own car will be one day only a sport, or considered as leisure (“I’m taking my vintage car to drive to the coast on week-ends”).
SoftBank’s Robot Buses to Take Grandparents Home on Country Roads: the japanese ageing population relies on buses and public transportations to get around. But as only 1/3 of the bus lines are profitable, the government is looking for a profitable solutions – driverless buses might be the One, as they would cut the operating costs by half. The vehicles will be tested on open roads (!!!) next year, and if everything goes well, by 2019 they will be fully working. How come it can be so fast? Japan invested many years ago in a nationwide network to monitor roads and traffic. Hence, it’s easier for smart vehicles to navigate…
Robots, Economy & Ethics
Jobs and skills in the digital economy: “To many workers, the words “digital technologies” may evoke one simple, dismaying image: a human-like robot sitting at their desk, doing the work that they used to do! This anxiety is not different from the fear of coachmen witnessing the diffusion of cars in the 1920s. In a sense, coachmen were right: cars did replace horse coaches. However, their children and grand-children found new and often better paid jobs in the wealth of new activities made necessary or possible by cars: automobile manufacturing, car repair, travelling sales, home delivery, mass tourism, road building, the petrol business, and so on.”
The point: it was incredibly hard in the 20s to imagine what the future of jobs would be like. It’s still hard today – so please, stop whining about the future of jobs and embrace this technological future. By preparing us for the future, we might create a better future….
On the other hand, an interesting article to read, about the dilemma work vs. play created by technology:
We are rapidly advancing towards a future where we have virtually no work and we sit in a chair being entertained by films, music or stories delivered to us by robot.
Robot and Ethics in France: (in french, obviously): Audition publique sur les Robots et la Loi, avec des personnalités connues du monde de la robotique mais étonnamment peu de représentation des grosses entreprises françaises, qui ont à mon sens aussi leur mot à dire (exception faite de Renaud Champion, d’Awabot, mais Awabot est un distributeur et non un constructeur).
La transcription peut être lue ici.
Innovation et organisation de la Recherche à Singapour: interesting to understand how institutions and research labs are cooperating in S’pore, considering the government is devoting 15M $SGD to the robotics industry.
Personas, you don’t need them: I had to use Persona while working in various companies, and EACH time my opinion was the same: they were biased, and the manifestation of the hopes and wishes of the company, rather than a real concept of interaction with the product. “The Jobs to Be Done (JTBD) framework has offered an alternative to the traditional persona. Instead of focusing on the customer for a product, a designer instead focuses on the job that customers need to get done with a product or service.”