Jan 2, 2019 · 34m · a16z
a16z Podcast | The IQ and EQ of Robots
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this a16z podcast episode, host Hannah speaks with Anki CEO Boris Sofman and CMU Professor Dave Touretzky about the evolution of domestic robotics, examining how combining smart hardware engineering with character-driven emotional design is transitioning companion robots from toys to practical home assistants.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 2.5 Guest teaching 4.2 Guest disagreement 0.6 The host pushing back 0.6
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 24:11 Flawed Compass Reframe
Boris firmly rejects the pursuit of humanoid consumer robots common in Asian tech development, labeling it a flawed compass that absorbs human limitations.
Hardest push from the host ▶ 2:51 Inquiring on Stalled TechHannah presses the guests on whether early consumer robotics like AIBO failed because underlying technical problems were fundamentally insurmountable at the time.
Biggest teaching moment ▶ 17:30 Spreadsheets Analogy for Consumer TechDave re-frames the conversation by explaining how everyday productivity tools like spreadsheets democratized complex computational concepts, applying that paradigm to consumer robotics.
The host holds their own ▶ 24:37 Patrol Robot Audio ObservationHannah demonstrates observational expertise by describing how urban security patrol robots play artificial mechanical whirring sounds to signal presence to pedestrians.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| The Bottom-Up Approach to Consumer Robotics | 2 | 5 | 1 | 1 | Hannah sets up the discussion by asking why Anki began with entertainment toys rather than complex manipulation. Boris and Dave explain the commercial history of consumer robotics including the Sony AIBO line and the distinction between Roomba utility and full humanoid aspirations. | |
| Evolving Companion Robots: Introducing Anki Vector | 2 | 4 | 1 | 1 | Hannah asks about current robot utility and what sets Vector apart from early generations. Boris outlines the architectural leap of placing onboard computation directly in the head rather than tethering to a mobile app. | |
| Open Robotics Platforms and Educational SDKs | 2 | 4 | 0 | 0 | Boris details how Anki exposes Python SDKs and APIs to allow users of all ages to program complex tasks without handling raw path planning. Dave highlights how CMU students use these platforms to build multi-story dollhouses with working elevators. | |
| Character Animation and Human-Robot Interaction | 3 | 5 | 0 | 0 | Hannah inquires about character animation and human-robot bonding. Boris explains how Anki hired animators from Pixar and DreamWorks to design probabilistic personality engines, noting how a minor adjustment in eye contact significantly boosted session duration. | |
| Customizing Robot Behaviors and Personality Dynamics | 3 | 5 | 1 | 1 | Hannah explores how personality dials adapt to user preferences. Dave introduces an extended analogy comparing consumer robotics literacy to the historical adoption of spreadsheets and PowerPoint. | |
| The Five Pillars of Robot Design and IQ/EQ Integration | 3 | 5 | 1 | 1 | Hannah prompts a discussion on ground-up physical design choices. Boris details the five design pillars and explains how small, cute form factors encourage users to forgive software or mechanical mistakes. | |
| Navigating the Uncanny Valley and Animal Influences | 4 | 4 | 1 | 1 | Hannah notes how Anki avoided the uncanny valley, sharing her personal observation of artificial mechanical sounds on security patrol robots. Boris criticizes humanoids as a flawed compass and explains using expressive tones instead of voice. | |
| Child Empathy and Hardware Drivers for Vector | 3 | 5 | 0 | 0 | Hannah and Boris discuss child empathy during gameplay. Boris describes how children intentionally throw games to prevent Cosmo from getting sad, which directly influenced Vector's capacitive touch and audio hardware. | |
| Voice Platforms and Character-Driven Smart Homes | 3 | 5 | 1 | 1 | Hannah asks how consumer robotics can scale to mass adoption. Boris and Dave discuss voice platform engagement numbers, dedicated AI vision hardware, and how software power compensates for mechanical manipulator deficits. | |
| Podcast Conclusion and Guest Farewell | 0 | 0 | 0 | 0 | Hannah closes the interview and thanks both guests for participating. Host-side scores are zero for this brief housekeeping sign-off. |
Statements from this episode (21)
Sofman: Robotic manipulation lags autonomous driving by five to six years
“Manipulation is today is probably where autonomous driving was five or six years ago, where you can already see the capability starting to take form, but the reliability and cost constraints are still pretty prohibitive, but you can kind of extrapolate where i…”
Sofman: Anki launched in entertainment robotics to build core capabilities
“For us, we started with entertainment because that was a way to create innovative applications and actually build the muscle early on that we can then carry over into these other spaces down the road.”
Dave Turetsky: Sony Sold Under 200,000 AIBO Units Before Exiting Robotics
“Sony sold fewer than 200,000 of these over the six year life of the product line. In January, 2006, they left the robotics business altogether.”
Boris Sofman: Anki Built Early Consumer Robots Using Smartphone Components
“We basically became for our early products scavengers of the smartphone industry. So you have computation, you have memory, you have cameras that now suddenly cost You know, 50 cents versus, you know, six dollars.”
Dave Turetsky: Roomba Succeeded by Solving a Task Robots Can Do Imperfectly
“The most popular robot ever in the history of robotics is the room of vacuum cleaner, because they found this one task that a robot can do poorly and still make you happy.”
Sofman: Anki Vector detethers from phones using onboard tablet-class processing
“From a technical standpoint, there was a huge, huge leap there because we now detethered away from a mobile device and put the equivalent of a tablet inside of this robot's head so that he can be always on and alive a hundred percent of the time, go back to hi…”
Sofman: Existing smart home products are isolated end devices in single rooms
“Right now, there's nothing that actually stitches the home together in that way. Everything is just an end device in a room.”
Sofman: Physical object manipulation is a mark of deep intelligence
“The moment you can intentionally manipulate something, it's a sign of really deep intelligence.”
Anki operates an in-house animation studio using Pixar talent and Maya
“We do, with folks from Pixar, DreamWorks, and these sorts of backgrounds. And so they literally, they're using Maya, a software suite that you would use to animate digital film and digital video games, but we've rigged up a version of Cosmo and Vector and all …”
Touretzky: In 20 years, all youth will grow up with computer vision
“20 years from now, No young person will have grown up in a world where computers could see, right? Computers could always see.”
Touretzky: Tabletop robot end-user control unlocks useful humanoid domestic robotics
“If you can get vector to push little things around on the tabletop, and you can make it easy enough to express your intent that the average person who's not a computer science major can do that, then you can scale up to the humanoid robot that's gonna cook you…”
Sofman: Cute robot designs make users forgiving of technical mistakes
“And people become forgiving of any limitations or mistakes that the robot then makes.”
Sofman: Depth camera costs dropped precipitously over past three to five years
“Like now in our future products, we can start putting in depth cameras, which give you a three D model of an environment at the cost, cost that would have been unimaginable five years ago or even three years ago.”
Sofman: Giving robots human voices sets unachievable expectations for intelligence
“Yeah, because the moment you have a voice, there's an expectation of intelligence that technology wasn't ready to meet yet.”
Sofman: Pursuing humanoid robots is a flawed approach to robotics design
“This idea of trying to pursue humanoid robots, which particularly is common in like Asia and Japan, it almost feels like a flawed compass because You end up absorbing all of the limitations of humans and their perceived kind of intent where you're stuck by wha…”
Soffman: Children intentionally threw games during play tests to avoid upsetting Cozmo
“What we were just shocked to see in some of our play tests early on is that you're playing this game that's kind of using the cubes as a buzzer called Quick Tap, where you're competing against Cosmo, and if you beat him, he'll kind of get upset, or maybe he'll…”
Soffman: Always-on operation is essential to maintain a robot's illusion of life
“If the moment you disconnect the phone, He dies. It kills all illusion of him being alive. But if he's always on and doesn't have that barrier, it completely changes the sort of emotional connection he can build.”
Sofman: Vector gets 10-12 queries daily versus 1-2 for smart speakers
“And so Suddenly, like, when you look at voice assistants that have an average of maybe, like, one to two queries a day, and we're at, like, 10 to 12 even before we have a voice assistants built in, it causes us to ask a lot of questions of how do you leverage …”
Sofman: Emotional interfaces drive higher utilization of functional tech features
“It's deceptive how important emotional interface is to make the functional elements get utilized in a better way.”
Turetsky: Advanced computing power compensates for physical hardware defects in robotics
“Computing power will make up for a lot of hardware defects. So if you have a underpowered Unreliable manipulator, but lots of good computing power. You can make that manipulator do amazing things, right?”
Sofman: Dedicated AI chips will become standard hardware within five years
“What you're going to see in the next wave of hardware is dedicated hardware that's not just for raw traditional computation, but for very specialized AI applications like deep learning and vision classification. These are the sort of things that in the next wa…”