Key Moments
4 wild AI predictions from a $39B tech founder
My First MillionWant to know something specific about what's covered?
We've already dissected every moment. Ask and we will deliver (with timestamps).
Key Moments
Humanoid robots are nearing a point where they can perform unsupervised tasks in homes, but the key challenge remains training them for novel environments and situations.
Key Insights
AI advancements are predicted to be 100 times larger than the internet, fundamentally reshaping everything.
Hark's AI system can now perform almost any task given to it on the internet, a breakthrough achieved through novel post-training and reinforcement learning processes.
Humanoid robots from Figure can perform multiple shifts a day, seven days a week, achieving human speeds in tasks like package sorting and already providing ROI for customers facing labor shortages.
Brett Adcock believes that truly intelligent AI models are developed by only about 20-30 people in California, with Meta actively acquiring this talent for astronomical sums.
The development of custom AI chips for Cover, designed to detect weapons from a 20-foot standoff, has faced significant fabrication lead times of up to a year, with current chips taking six months for further production.
Brett Adcock's philosophy is that tackling hard problems, like developing humanoid robots or advanced AI, offers a significantly higher payoff for a manageable increase in difficulty compared to easier ventures.
The AI revolution will dwarf the internet's impact
Brett Adcock, founder of Figure.ai and Hark, predicts that the AI revolution will be approximately 100 times larger than the internet in terms of its impact. He emphasizes that current AI systems are demonstrating remarkable capabilities, with deep learning proving highly effective and progress accelerating rapidly. This growth is not just incremental; it's a fundamental shift that will permeate every aspect of life, both physically and digitally, effectively consuming and transforming the existing internet landscape.
Hark cracks autonomous computer use for AI
Hark, Adcock's AI lab, has developed a system that can perform almost any task on the internet autonomously. Unlike previous attempts that relied heavily on APIs or struggled with basic navigation, Hark's AI interacts with computers by looking at screens, moving cursors, and using keyboards. This was achieved by spinning up virtual computers for each agent and, crucially, through a proprietary reinforcement learning process in post-training. This breakthrough means AI can now autonomously handle tasks like booking flights, ordering food, or managing financial models, bridging the gap towards a truly general-purpose AI assistant.
Humanoid robots are ready for commercial and home deployment
Figure.ai is making significant strides in deploying humanoid robots. These robots are designed for 24/7 operation and can already perform tasks at human speeds, such as sorting packages in logistics and manufacturing. Customers are turning to Figure due to severe labor shortages and high turnover rates, finding that the robots offer a compelling ROI. While initial applications focus on commercial environments, Adcock believes the home environment is also becoming increasingly solvable. The core challenge is enabling robots to perform unsupervised, multi-task operations in novel environments without prior specific training, a problem Adcock dedicates several hours daily to solving.
The future interface: Beyond the smartphone
Adcock posits that current hardware, like smartphones and laptops, is outdated for AI interactions. He is developing new AI-native devices designed to replace phones and computers, featuring real-time, always-on AI that understands context and memory. These new devices will abstract away most apps, potentially leading to an AI operating system instead of an app store. He envisions devices with personalized AI agents possessing an individual's memories and data, akin to Jarvis from Iron Man, offering a profoundly personalized and superhuman assistant experience. While glasses and wearables might play a supporting role, the primary focus is on a new 'mega device' category.
The intense battle for AI talent
The AI talent pool is extremely scarce, with Adcock estimating only 20-30 individuals in California possess the true expertise to build cutting-edge AI models. Companies like Meta are aggressively acquiring this talent, offering compensation packages in the tens of millions of dollars over four years. Adcock respects this strategy but personally prefers to hire individuals deeply committed to the mission rather than 'mercenaries' motivated solely by money. He argues that a smaller, dedicated team can achieve more than a large, hired workforce driven by financial incentives alone, though Meta's aggressive acquisition is undeniably shaping the talent market.
Cover's mission: Detecting weapons to prevent school shootings
Adcock is also leading Cover, a company focused on preventing school shootings by developing a system capable of detecting weapons from a 20-foot standoff. This project is a deep tech, engineering-intensive challenge involving both hardware and AI. While they aim to beta-test a full-scale system by the end of the year, fabrication lead times for custom AI chips have caused delays. These specialized chips are not off-the-shelf and require unique manufacturing processes, pushing the timeline for market entry back by several months.
The philosophy of tackling hard problems
Adcock champions the strategy of pursuing difficult challenges, arguing they often have less competition and a significantly higher potential payoff. He believes that hard problems are typically only two to five times more difficult than easier ones, but can offer a hundred or even a million times better outcomes. This principle guides his focus on developing advanced humanoid robots and AI, areas with immense potential but also considerable technical hurdles, rather than pursuing niche markets like quadrupedal robots, which he sees as having limited economic value.
Navigating 'rock bottom' and founder conviction
Adcock shares his experience of hitting 'rock bottom' financially multiple times during his startup journey, often facing near-bankruptcy and personal debt. His method for dealing with these lows is to focus on day-by-day survival, breaking down overwhelming tasks into manageable 'punch lists.' He draws parallels to ultra-marathon running, advising to focus on the next small milestone rather than the entire daunting distance. This resilience is fueled by a deep conviction in his work, stemming from thorough understanding and hands-on experience, enabling him to push forward despite overwhelming odds and skepticism from others.
Mentioned in This Episode
●Products
●Software & Apps
●Companies
●Organizations
●People Referenced
Common Questions
Brett Adcock has founded several companies including Vetteri (recruitment, sold for $100M+), Archer (unmanned planes), Figure (humanoid robots), Cover (school shooting detection), and Hark (AI lab focused on general computer use and AI-human symbiosis).
Topics
Mentioned in this video
Mentioned as a source for Brett Adcock's net worth.
Jet Propulsion Laboratory, mentioned as the origin of the chief engineer for Cover.
A startup incubator where Adcock's company Vetteri participated; he notes the high failure rate of companies within it.
Mentioned as the origin of the chief engineer leading Adcock's Cover project.
Adcock's first company, sold for over $100 million.
A company founded by Adcock aimed at preventing school shootings.
Adcock's new AI lab focused on developing AI for general computer use and digital-human symbiosis.
Mentioned as an example of a service that requires general computer use by AI, lacking a consumer API.
Mentioned in relation to its CEO Jensen Huang's management style.
Mentioned in the context of engineers who previously worked on projects like the Tesla X or Cybertruck.
Discussed for its aggressive strategy of hiring AI talent with high salaries and RSUs.
Mentioned for resurfacing Adcock's old database on successful people.
Founder of Vetteri, Archer, Figure, Cover, and Hark. Discussed his career, philosophy, and predictions about AI and robotics.
Mentioned in relation to AI computer use capabilities ('MacroHard').
CEO of NVIDIA, described as an unbelievable operator with a unique management style.
Mentioned as an example of someone fully dedicated to their craft, admired for his pursuit of being the best.
An admired figure for his product-led approach and operational skills, mentioned in the context of iPhone development.
An investor in Figure, who has provided Adcock with advice on scaling the company.
Mentioned as a system that has attempted computer use capabilities, though Adcock implies Hark has surpassed it.
From Iron Man, used as an analogy for the ideal superhuman AI assistant Adcock envisions.
An AI tool that Adcock found meaningful for communication, standing out in the last six months.
Mentioned as a current AI model, with a discussion about its voice capabilities and limitations.
More from My First Million
View all 60 summaries
46 min3 killer businesses hiding in plain sight
56 minWe accidentally dropped 9 genius business ideas while drunk
49 minWe found the internet's best sales advice
47 minWhat happened when I accidentally sat next to...
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free