The Humanoid Robot Trap: What Automation Really Needs | Clara Vu
Automated Podcast 46:10
2,245 views · 25 likes Watch on YouTube ↗
Humanoid robots are attracting enormous investment.
Clara Vu thinks the form factor is a trap.
In this episode of Automated, Brian Heater sits down with the veteran roboticist and Veo Robotics co-founder at MassRobotics to discuss what it actually takes to turn an impressive robotics demo into a reliable product.
Clara says a prototype that works 80% of the time may represent only 2% of the work required to ship. The remaining challenge is buried in edge cases, environmental variability, exception handling, and all the failure modes a polished demonstration never shows.
She also explains why building robots in the human form creates unnecessary engineering constraints, how anthropomorphism causes people to overestimate what a machine can do, and why venture capital's demand for enormous markets has helped fuel the general-purpose humanoid boom.
Brian and Clara trace those lessons through her career, from early iRobot oil-well robots and Hasbro's My Real Baby doll to Roomba, Harvest Automation, Rethink Robotics, and Veo. Along the way, Clara reveals why the original Roombas were made in a toy factory and why reliable robotic products must begin with deceptively simple problems.
The conversation ends with a more practical vision for automation: combine existing robots, sensors, AI, computer vision, and software with application-specific engineering. Instead of waiting for one machine that can do everything, Clara argues that the industry should build easier systems that can create value now.
If a pallet needs to move, make the pallet jack intelligent. Do not build a humanoid just to operate it.
KEY MOMENTS
00:00 Why an 80% robot demo is only 2% finished
03:08 Discovering engineering by accident
04:30 The iRobot internship that changed her life
05:50 Learning to code from a LISP book
06:41 Joining iRobot before she was legally old enough to drink
07:00 Oil-well robots, Hasbro, and interactive toys
08:48 How toy manufacturing shaped the original Roomba
11:04 How Roomba and PackBot transformed iRobot
13:27 Why Roomba's internal software was called TL
15:30 Why home robotics is so difficult
17:06 Where robotics can create the most value
18:08 Building Harvest Automation's greenhouse robots
19:24 The 3-week test for a 3-year robotics product
21:51 How Rethink Robotics led to Veo
23:12 What the self-driving boom missed
24:08 Why a working prototype can still be nowhere close
24:57 Describe the problem before trying to automate it
27:35 Could people teleoperate any robot form factor?
28:13 Why humans do not actually see in 3D
29:38 The danger of anthropomorphizing machines
30:33 Why the humanoid form factor is a trap
32:46 The mismatch between venture capital and robotics
35:31 Why custom automation needs a different funding model
39:20 Combining what humans and robots do best
40:12 The wall between industrial robotics and the AI world
42:27 Can specialized robots deliver faster ROI?
44:20 Why robotics does not need another fundamental breakthrough
45:13 Make the pallet jack intelligent
Connect with Clara Vu
https://www.linkedin.com/in/clara-vu-244941/
Learn more about Veo Robotics and its acquisition by Symbotic
https://www.symbotic.com/news/symbotic-acquires-veo-robotics-to-enhance-efficiency-and-safety-innovation/
We'd love to hear from you. Have thoughts or guest suggestions?
Reach us at podcast@automate.org
You can find the transcript and more episodes of Automated at automated.fm
Unlock full access to Automated and explore everything automation.
Subscribe today and leave a review on YouTube, Apple Podcasts, Spotify, and the Automated Newsletter.
https://www.youtube.com/@automatedpodcast
https://podcasts.apple.com/us/podcast/automated-with-brian-heater/id1837762221
https://open.spotify.com/show/60olq6brlBEIJWggx2fMR6
https://www.automate.org/automation/automated-newsletter
You can also find us on:
LinkedIn https://www.linkedin.com/showcase/automated-podcast-by-a3/
Instagram https://www.instagram.com/automatedpod/
Clara Vu thinks the form factor is a trap.
In this episode of Automated, Brian Heater sits down with the veteran roboticist and Veo Robotics co-founder at MassRobotics to discuss what it actually takes to turn an impressive robotics demo into a reliable product.
Clara says a prototype that works 80% of the time may represent only 2% of the work required to ship. The remaining challenge is buried in edge cases, environmental variability, exception handling, and all the failure modes a polished demonstration never shows.
She also explains why building robots in the human form creates unnecessary engineering constraints, how anthropomorphism causes people to overestimate what a machine can do, and why venture capital's demand for enormous markets has helped fuel the general-purpose humanoid boom.
Brian and Clara trace those lessons through her career, from early iRobot oil-well robots and Hasbro's My Real Baby doll to Roomba, Harvest Automation, Rethink Robotics, and Veo. Along the way, Clara reveals why the original Roombas were made in a toy factory and why reliable robotic products must begin with deceptively simple problems.
The conversation ends with a more practical vision for automation: combine existing robots, sensors, AI, computer vision, and software with application-specific engineering. Instead of waiting for one machine that can do everything, Clara argues that the industry should build easier systems that can create value now.
If a pallet needs to move, make the pallet jack intelligent. Do not build a humanoid just to operate it.
KEY MOMENTS
00:00 Why an 80% robot demo is only 2% finished
03:08 Discovering engineering by accident
04:30 The iRobot internship that changed her life
05:50 Learning to code from a LISP book
06:41 Joining iRobot before she was legally old enough to drink
07:00 Oil-well robots, Hasbro, and interactive toys
08:48 How toy manufacturing shaped the original Roomba
11:04 How Roomba and PackBot transformed iRobot
13:27 Why Roomba's internal software was called TL
15:30 Why home robotics is so difficult
17:06 Where robotics can create the most value
18:08 Building Harvest Automation's greenhouse robots
19:24 The 3-week test for a 3-year robotics product
21:51 How Rethink Robotics led to Veo
23:12 What the self-driving boom missed
24:08 Why a working prototype can still be nowhere close
24:57 Describe the problem before trying to automate it
27:35 Could people teleoperate any robot form factor?
28:13 Why humans do not actually see in 3D
29:38 The danger of anthropomorphizing machines
30:33 Why the humanoid form factor is a trap
32:46 The mismatch between venture capital and robotics
35:31 Why custom automation needs a different funding model
39:20 Combining what humans and robots do best
40:12 The wall between industrial robotics and the AI world
42:27 Can specialized robots deliver faster ROI?
44:20 Why robotics does not need another fundamental breakthrough
45:13 Make the pallet jack intelligent
Connect with Clara Vu
https://www.linkedin.com/in/clara-vu-244941/
Learn more about Veo Robotics and its acquisition by Symbotic
https://www.symbotic.com/news/symbotic-acquires-veo-robotics-to-enhance-efficiency-and-safety-innovation/
We'd love to hear from you. Have thoughts or guest suggestions?
Reach us at podcast@automate.org
You can find the transcript and more episodes of Automated at automated.fm
Unlock full access to Automated and explore everything automation.
Subscribe today and leave a review on YouTube, Apple Podcasts, Spotify, and the Automated Newsletter.
https://www.youtube.com/@automatedpodcast
https://podcasts.apple.com/us/podcast/automated-with-brian-heater/id1837762221
https://open.spotify.com/show/60olq6brlBEIJWggx2fMR6
https://www.automate.org/automation/automated-newsletter
You can also find us on:
LinkedIn https://www.linkedin.com/showcase/automated-podcast-by-a3/
Instagram https://www.instagram.com/automatedpod/
Category (YouTube): Science & Technology
Playback is via YouTube's official embedded player. Data from YouTube; Exumo is not affiliated with YouTube.