Moonshots with Peter Diamandis

Episode #156: A Humanoid Robot in Every Home? It's Closer Than You Think w/ Brett Adcock (at A360 2025)
Date: March 17, 2025

Overview

In this forward-thinking episode of Moonshots, Peter Diamandis sits down with Brett Adcock, founder and CEO of Figure, to explore the rapid advancements in humanoid robotics and artificial intelligence. Adcock details his vision of integrating AGI (Artificial General Intelligence) with robotics, discusses the technological moonshots, and shares the breakthroughs and challenges in bringing humanoid robots into homes and the workforce. The episode dives deep into technical, ethical, and societal implications, combining engineering insights with real-world business impacts.

Key Discussion Points & Insights

1. The Figure Vision: Humanoid Robots as AGI's "Deployment Vector"

[00:08–02:22]

• Brett Adcock shares his motivation for founding Figure after successful ventures like Archer Aviation. He envisions humanoid robots as the practical embodiment for AGI, solving the challenge of bridging digital intelligence and the physical world.

• Quote:

  "The humanoid robot is like the ultimate deployment vector for AGI. You need something that with no hardware changes can do everything a human can." – Brett Adcock [02:22]

• Adcock sees a risk in AGI existing only in servers, relying on humans for action, potentially leading to dystopian futures. A mechanical human, as a platform, allows AGI to act autonomously in the world.

2. Blazing Speed in Hardware Iteration & Vertical Integration

[03:33–05:18]

• Figure achieved a working robot in under 12 months from its founding, and ships new hardware platforms every 12–18 months, driven by rapid iteration and lessons learned from prior hardware.

• Quote:

  "My rule of thumb is the first or second generation hardware is always going to suck. ...We are designing a new hardware platform every 12 to 18 months." – Brett Adcock [04:18]

• Complete vertical integration was a necessity due to the absence of a supply chain for humanoid robots—everything from actuators to operating systems is built in-house.

3. Design for End-Use: Home and Workforce Applications

[05:48–09:30]

• Figure’s approach starts with clean-sheet design to maximize general-purpose capability, aiming to replace mundane household tasks—walking the dog, making coffee, laundry—and penetrate commercial labor markets, which represent half of global GDP.

• Robots are already operational with major customers like BMW, performing high-precision, repetitive tasks in manufacturing.

• Quote:

  "Our robots are doing that fully autonomously at the speeds we need to. ... No days off." – Brett Adcock [07:18–08:02]

• The target retail price is projected at $20,000–$30,000, making personal ownership feasible and attractive.

4. Building the "Best Team in the World" & Startup Culture

[09:51–11:50]

• Brett attributes Figure’s velocity to assembling a world-class engineering team, cultivated through a culture of clarity, accountability, and strong work ethic.
• The team's unity is reinforced by a clear mission and the “dopamine hit” of shipping cutting-edge products.
• Quote:

  "In order to build one of the world's greatest products, you need one of the world's greatest teams. ... Everybody knows what they should be doing." – Brett Adcock [10:46–11:50]

5. Tackling the Three "Impossible" Robotics Problems

[14:21–16:12]

• Recruiting technical talent hinged on Adcock’s upfront funding and the transparency that success was far from assured—three core challenges had to be solved in five years:

  • Humanoid hardware with human-like speed, range, and reliability
  • End-to-end neural network control rather than classic rule-based robots
  • True generalization: performing unseen tasks on command via natural language

• Quote:

  "You have to build incredible hardware… you need to solve this as a neural net problem, not a control problem ... and then you have to generalize: do something you’ve never seen before through speech." – Brett Adcock [14:21]

6. Emergence of Helix: Proprietary AI Tailored for Robotics

[16:12–18:33]

• After experimenting with OpenAI models, Figure developed "Helix," an in-house vision-language-action model that allows robots to generalize and execute new tasks given only English commands.

• In an onstage demo, two robots who’d never seen the test groceries were able to interpret the instruction “put these groceries away,” inferring and executing all steps using Helix.

• Quote:

  "This is probably the most important AI update for robotics in human history." – Brett Adcock [18:21]

• Helix's emergent behaviors include robots looking at each other to coordinate object handoffs, paralleling critical nonverbal human communication.

7. Next-Generation Robots: Figure 3

[19:45–21:49]

• Figure 3’s design is described as a quantum leap over previous models: 90% cheaper, more compact, with superior sensing and hardware explicitly designed for AI learning. Rollout is scheduled for 2025, with mass production ambitions.
• Quote:

  "Figure 3 is just next level design. ... For me, like, the most proud moment I’ve had in engineering in my career." – Brett Adcock [21:21]

8. Commercial & Home Deployment: Timeline and Scale

[21:49–26:06]

• Workforce robots will hit mass deployment first; demand is so high that major companies would absorb all available robots instantly. The home market is technically far tougher, as domestic environments are unstructured ("the Wild West").
• Still, Adcock predicts in-home alpha tests will begin this year (2025), with significant progress expected across the decade.
• Data—not algorithms—is now the last bottleneck; as datasets grow, home applications will rapidly approach viability.
• Quote:

  "I think you’re going to see it in the coming years being put into homes just through speech, be able to do very long horizon hours of work without any prompt." – Brett Adcock [26:06]

Notable Quotes & Memorable Moments

• On AGI in the physical world:

  "You can’t solve this with anything else besides a human, like a mechanical human." – Brett Adcock [02:22]

• On culture and team:

  "We're in Silicon Valley, but almost like the anti Silicon Valley. You have to work every day in the office. ... We've assembled now a couple like hundreds of like the best engineers in AI robotics in the world." – Brett Adcock [11:19]

• On emergent AI behaviors:

  "Part of this was like emergent from training. So when the robots are doing handovers, they actually look at each other... as the clearing way signal for like we should be releasing the item into each other's hands, which is like really interesting." – Brett Adcock [18:42]

Timestamps for Key Segments

• [00:08] Motivation for founding Figure; AGI in the real world
• [03:33] Fast iteration and hardware lessons
• [05:48] Design philosophies for product and market
• [07:18] Case study: BMW factory robots
• [08:19] Pricing, business model, home utility
• [09:51] Figure team and recruitment philosophy
• [14:21] The "three impossible" technical challenges
• [16:12] Building proprietary AI: Helix
• [18:33] Emergent robot-robot behaviors, communication
• [19:45] Figure 3 announcement and features
• [21:54] Mass production plans, workforce vs. home rollout
• [24:32] Home alpha tests in 2025
• [26:06] Long-term vision for home robots

Takeaways

• Figure is executing a staggering pace of innovation, with humanoid robots transitioning from factory floors to homes possibly within the next few years.
• The current frontier is not just hardware or legacy AI, but emergent, fully end-to-end neural networks capable of generalizing across unstructured, unpredictable environments.
• Adcock’s candidness about the technical and cultural risks, as well as his leadership style and the “moonshot” mentality, offers a transparent look into the next decade of exponential technology.