141: The Former Head of DJI LiDAR Developed an "Electric Wheelchair"? Talking Entrepreneurship with Strutt's Hong Xiaoping After Two and a Half Years: Not Doing Humanoid Robots Can Still Lead to Embodied Intelligence

1. Key Themes

The Intentional Path to Entrepreneurship: From Physics PhD to Robotics Founder

Hong Xiaoping's journey wasn't impulsive—it was a decade-long preparation. After earning his physics PhD from Berkeley, he joined DJI in 2016 specifically to learn product development at scale. "I've always wanted to start a business, but I didn't know how to begin," he explained [00:06:11]. His time at DJI (2016-2019) leading the LiDAR division taught him how to take products from concept to mass production, a critical skill missing from academic training. The key insight: entrepreneurship requires not just technical excellence but understanding the entire product lifecycle and building for manufacturability from day one.

Solving Aging Through Robotics: The Social Mission Behind Commercial Success

Hong identified eldercare as the defining challenge of our era—and a perfect application for robotics. "Aging is an irreversible trend worldwide, and there's no solution," he observed [00:23:40]. The insight goes deeper than demographics: in wealthy countries, care worker shortages mean even high-end nursing homes struggle to provide basic services. Singapore's Changi Airport can't find enough people to push wheelchairs. This isn't just a market gap—it's a crisis where technology can genuinely improve lives. Hong's vision: "If robotics can help people, you should start with those who need it most and solve the most pressing social problems" [00:24:45].

The Platform Strategy: Building the iPhone of Personal Mobility

The EV1 isn't a wheelchair—it's a robotic platform. Hong was explicit: "It's far beyond what an electric wheelchair can do. It's actually a personal mobility device" [00:29:32]. The product features expansion ports supporting 200-500W power output, high-speed networking, and mounting systems for mechanical arms or other accessories [00:53:59]. This echoes Apple's iOS strategy—build the foundational platform first, then enable an ecosystem. The parallel to Google's Everyday Robots project is intentional: Hong wants to create a "robotic helper" that can eventually do much more than mobility [00:52:07].

2. Contrarian Perspectives

Humanoid Robots Are a Dead End (For Now)

While the industry races toward humanoid robots, Hong argues the path isn't viable: "I don't see any scenario that can provide such data—large amounts of continuous data about human living scenarios" [00:16:22]. His reasoning is mathematical: physical world robotics requires far more dimensional data than language models. "Language is one-dimensional... but robotics is at least four-dimensional: time is one dimension, space is three dimensions, plus touch, sound, vision—the physical world is extremely multi-dimensional" [00:15:15]. Without sufficient training data, the chicken-and-egg problem is insurmountable: poor performance means no users, no users means no data to improve. His alternative: "We need to create products with a clear path forward, solving real problems today while collecting data for tomorrow's breakthroughs."

Product Excellence Matters More Than Market Size

Hong challenges the startup obsession with TAM (Total Addressable Market): "Many times, people focus more on the M in PMF, but I think if the P is good enough, the PMF can be excellent" [00:41:26]. He cited 3D printing—the market seemed nonexistent until MakerBot made the product accessible enough for regular people. "It's not that there was no M—it's that the P wasn't good enough" [00:42:26]. This flips conventional VC wisdom: instead of finding huge markets with mediocre products, create exceptional products that expand their own markets. The wheelchair market looks small, but personal mobility affects everyone eventually—the category itself needs redefinition.

Higher Hiring Standards in Early-Stage Startups

Most startups lower hiring bars early on. Hong does the opposite: "Although we're a startup, we've never relaxed our requirements. Our culture and values are themselves a filter" [01:15:18]. People who don't fit the culture leave quickly—or are asked to leave. This seems risky when resources are limited, but Hong believes compromising on talent creates technical debt that's harder to fix than code. "Many startups face a painful process of having to eliminate team members, even core ones. We think about things in advance to avoid unnecessary corrections" [01:22:42]. The DJI influence is clear: maintain excellence even when it seems unaffordable.

3. Companies Identified

DJI (大疆)

Description: Leading drone manufacturer and Hong's former employer where he built the LiDAR division
Why mentioned: Exemplar of engineering excellence and product-market fit through technical superiority
Quote: "DJI essentially is a robotics company. Autonomous driving and these applications are all manifestations of mature robotics technology" [00:11:45]. Hong credited DJI's culture: "The technical atmosphere is very strong, and if we're only talking about technical matters, there aren't too many interpersonal issues. It's results-oriented" [01:09:27].

Livox (览沃科技)

Description: DJI's LiDAR spinoff that Hong led before it became independent in 2019
Why mentioned: Successfully brought affordable LiDAR to market, enabling autonomous vehicle development
Quote: "When Livox launched its first LiDAR, it was essentially a breakthrough in volume production. Before that, people thought 'let's make prototypes first, then see if future costs can come down'... but we designed everything for mass production from the start" [00:07:37].

Google (via Everyday Robots project)

Description: Google X's robotics initiative (now discontinued) that explored helper robots
Why mentioned: Inspiration for Robootics' vision of robotic helpers integrated into daily life
Quote: "Google published a series of very interesting papers about robotic thinking and sensor fusion... PaLM-SayKan, RT, RT2—the now-famous VLA (Vision Language Action) model was proposed in those papers. I felt that robots were beginning to have higher-level cognition" [00:12:00].

4. People Identified

Wang Tao (Frank) - DJI Founder

Description: DJI's founder and CEO, known for extreme product standards
Why mentioned: Role model for engineering excellence and product obsession
Quote: "Frank is an excellent product manager. He has very strong technical insight and very demanding requirements for product engineering excellence. There are often situations where some people think something is good enough, but he thinks it needs to be redone" [01:07:38].

Zhang Fu (张富) - HKUST Professor

Description: Hong Kong University professor and former DJI colleague who developed FAST-LIO algorithm
Why mentioned: Made LiDAR SLAM practical through open-source algorithms, enabling the industry
Quote: "Zhang Fu did a lot of work, especially developing an algorithm called FAST-LIO series, which basically optimized the entire SLAM framework for LiDAR... After he open-sourced it, it became the de facto standard algorithm" [01:12:04].

Xiao Wenlong (NIC) - Robootics Co-founder

Description: Former DJI power systems engineer, joined as Robootics COO
Why mentioned: Perfect complementary co-founder with mechanical/electrical expertise
Quote: "NIC is quite complementary to me. He wanted to find a good team, and I wanted to do something worth dedicating a lifetime to. So we matched... He's very detail-oriented and can really deliver excellent results" [01:06:02].

Steve Jobs

Description: Apple founder
Why mentioned: Role model for vision, taste, and articulating future clearly before it exists
Quote: "Many people see his product excellence and user insight, but I admire more his taste, his foresight. For example, with iPhone, before making it they spent a long time building iOS... Being able to think these things through clearly and articulate these ideas to employees before they're visible—I think this is what I need to learn" [01:21:26].

5. Operating Insights

Design for Manufacturing from Day One

Hong learned at DJI that prototype thinking kills startups. "Many people in the industry would say, 'let's make prototypes first, see how far costs can drop, then see what opportunities exist.' But we thought all design must serve mass production" [00:07:37]. This meant using only purchasable components or proven manufacturing processes—no waiting for future technology. The discipline paid off: their first LiDAR was significantly cheaper than competitors precisely because it could be manufactured at scale. Lesson: constrain your design choices to what's producible today, not tomorrow's promises.

The "Hold Button" Interaction Pattern

The team discovered that mode-switching killed adoption. Users wanted obstacle avoidance help but toggle buttons created cognitive load. Solution: "We added a button on the joystick—real-time authorization. When I want help, I press and hold it. When I don't need it, I release and return to normal driving mode" [00:49:25]. The brilliance: pressing the button creates psychological ownership—"I'm authorizing it to help me change direction"—even though the user isn't controlling direction directly. Takeaway: the best interfaces make assistance feel like empowerment, not automation.

Reliability Testing at 10x Industry Standards

Most electric wheelchairs test to 200,000 cycles on durability machines. Robootics targets 2 million cycles [00:58:02]. This seems excessive until you consider the product category: "Users sitting in this vehicle feel it's very safe, it won't fall apart midway" [00:58:06]. For mobility devices serving vulnerable populations, the margin for error is zero. The insight: in certain categories, 10x better reliability isn't overkill—it's the minimum viable product. Cutting corners on safety destroys brand trust permanently.

Culture as Constitution, Not Aspiration

Hong and his co-founder wrote their culture values before hiring anyone: "This became our company constitution. If you comply with these principles, it's a green light. If you violate them, it's red—unless we collectively decide to adjust them" [01:08:41]. Unlike most startups that relax standards early on, Robootics uses culture as a hiring filter. People who don't fit leave quickly. The trade-off: slower hiring but zero cultural debt. Key principle: "Don't trust experts first, trust independent thinking first" [01:07:11]—borrowed directly from DJI's playbook.

The "Iterative Roadmap" Product Strategy

Rather than perfecting everything in v1, Hong cut features ruthlessly: "We had designed a transformable structure—the entire vehicle could change shape, adjust height, the seat could tilt forward... but for us it was too complex" [00:44:41]. The calculus: complex mechanisms require extensive reliability testing that delays launch. Better to nail core value (intelligence, comfort, terrain capability) and add features in v2, v3. This requires iron discipline—killing good ideas that aren't essential—but accelerates time-to-market while maintaining quality.

6. Overlooked Insights

The Data Moat Hiding in Plain Sight

Hong casually mentioned something profound: their vehicle will collect real-world human living scenario data that humanoid robot companies desperately need but can't access [00:25:03]. While everyone chases the humanoid robot "holy grail," Robootics is solving a chicken-and-egg problem: "When you have tens of thousands, hundreds of thousands of such machines in homes worldwide attempting various actions—this is the best platform for embodied intelligence" [00:56:49]. The overlooked insight: mobility devices that people use daily in diverse environments become massive data collection platforms. This isn't their marketing pitch—it's their long-term competitive moat that compounds over time.

The "Premiumization of Assistive Technology" Opportunity

Hong observed that high-end nursing homes in the US charge $500K-$1M entry fees partly because they can't find care workers [00:24:02]. Meanwhile, airports can't find wheelchair pushers [00:24:19]. But he didn't explicitly connect the dots: there's a massive premium market for "luxury" mobility that has nothing to do with medical necessity. Wealthy aging populations will pay top dollar for products that don't look medical, work beautifully, and preserve dignity. The $20K+ price point isn't a barrier—it's a positioning opportunity. The assistive technology market is ripe for the premiumization playbook that transformed coffee, fitness, and a dozen other categories.