This 22-Year-Old Built TikTok for Mobile Games, and It’s Growing Fast | E2276

This 22-Year-Old Built TikTok for Mobile Games, and It’s Growing Fast | E2276

This Week in Startup: Exploring Nanog

Introduction to Nanog

  • Lon Harris introduces the episode featuring Jason Calacanis, Albert Brotherton, and Boris Rattalov, co-creators of Nanog, a platform likened to TikTok but for mobile video games.
  • The hosts express excitement about the product's availability on both Google Play Store and iOS Store.

Demonstration of Features

  • Boris begins demonstrating the app, explaining its functionality where users can quickly switch between casual mobile games similar to scrolling through TikTok videos.
  • Users experience "choice paralysis" with numerous game options; Nanog simplifies this by allowing instant gameplay without extensive research.

Game Creation Process

  • The creation feature utilizes AI templates that drastically reduce game development time from days to mere minutes.
  • Users can create games like a 3D version of Flappy Bird in approximately 90 seconds while being able to play other user-generated games for inspiration.

Remixing and Collaboration

  • After creating a game, users have the option to remix it or modify others' creations easily, fostering collaboration within the community.
  • The technology behind Nanog includes a custom game engine utilizing Gemini for asset generation and coding tasks.

Business Model Insights

  • Albert discusses their business model focused on interactive content rather than just traditional gaming, highlighting potential revenue experiments in this new field.
  • He explains how brands like Domino's could leverage interactive ads through engaging mini-games tailored around their products.

Introduction to the Founders' Background

Personal and Educational Background

  • The conversation begins with a light-hearted exchange about the founders, Albert and Boris, who are both young professionals. Albert is 22 years old while Boris is 24.
  • Albert mentions that he skipped college, while Boris attended the University of Leeds to study computer science but did not attend classes for two out of three years due to COVID-19.
  • Boris reveals his student debt amounts to around £40,000 - £50,000 (approximately $50,000), which he considers manageable compared to U.S. student debts.

Gaming Influence and Connection

  • Albert shares his long-standing passion for gaming since age six, particularly enjoying World of Warcraft—a game often associated with millennials rather than Gen Z.
  • The founders met through a mutual childhood friend rather than in-game interactions; they became best friends after being introduced three years ago.

Startup Aspirations and Funding

Business Development Plans

  • The discussion shifts towards their startup ambitions. They have successfully raised funds from investors and are currently based in London and Sofia with plans to expand into New York.

Market Insights

  • The host humorously warns them about potential dangers in New York while emphasizing its vibrant marketing scene that could benefit their startup's growth.

Authenticity in Marketing

Case Study: Staples Baddy

  • A case study is presented regarding "Staples Baddy," a TikTok influencer who authentically connects with her audience by discussing office supplies like pens.
  • Despite not purchasing directly from Staples, the speaker acknowledges how this influencer has generated significant brand engagement through genuine content creation.

Hiring Challenges for Startups

  • The conversation touches on hiring challenges faced by startups. It emphasizes the importance of making every hire count when running a small company.
  • LinkedIn Hiring Pro is recommended as an effective tool for streamlining recruitment processes, highlighting its success in helping companies find quality candidates quickly.

Fast Food Marketing and Creator Economy Insights

The Impact of Viral Marketing in Fast Food

  • Discussion on a viral TikTok video by the McDonald's CEO, where he awkwardly eats a burger, leading to negative perceptions.
  • Other fast food CEOs felt compelled to create their own videos showcasing enthusiasm for their products in response to the McDonald's clip.
  • Critique of the McDonald's CEO's robotic demeanor and lack of authenticity while promoting his product.
  • Observations about how the portrayal of burgers as "products" rather than food affects consumer perception.
  • Speculation on the marketing strategies behind these videos, questioning if there was adequate guidance from social media teams.

Authenticity in Brand Representation

  • Commentary on how it’s evident that some fast food executives do not genuinely enjoy their products based on their presentation style.
  • Mention of a meticulously prepared salad that contrasts with the CEO's apparent disdain for his own burger, highlighting issues with brand authenticity.

Monetization Strategies in Gaming

  • Transition into discussing monetization strategies for creators within gaming platforms, emphasizing organic content creation without needing permission.
  • Exploration of potential monetization through microtransactions and creator currencies similar to Roblox's model.

Experimentation with Creator Currency

  • Plans to experiment with creator currency as an incentive for game developers who produce engaging short-form content.
  • Comparison made between traditional social media engagement and gaming dynamics, noting differences in user interaction post-publication.

User Engagement Metrics

  • Overview of user engagement statistics since launch: 100k users with 20% identified as power users playing multiple games per session.
  • Insight into high retention rates and session times compared to typical social media platforms, indicating strong user interest.

How to Build a Successful Gaming Community?

Initial User Engagement Strategies

  • The discussion begins with the excitement surrounding a new gaming product, indicating potential for it to become a breakout hit.
  • The founders started by creating a Discord community, growing it to 10K members, but only about 2K were actively engaged as test users during the beta phase prior to launch in January.
  • They emphasized the importance of game sharing and social interactions within the platform, highlighting how users can challenge each other and share scores.

Cultural Insights on Game Sharing

  • A key point raised is that sharing games isn't part of current culture; while people share social media content frequently, games are less commonly shared among friends.
  • The conversation notes that while gamers do share specific gameplay clips (e.g., Call of Duty), this remains niche compared to broader social media sharing habits.

User Behavior and Content Sharing

  • Two major questions were posed: Do people want to share games? And will gaming shift towards short-form content similar to trends seen in platforms like TikTok?
  • Current data shows promising engagement: for every 100 likes on their app, there are between 30 and 50 shares, suggesting strong interest in sharing gaming experiences.

Gaming Session Dynamics

  • Users are engaging with multiple games per session; approximately 20% of users play through 25 games in around 21 minutes. This indicates a trend toward shorter gaming sessions akin to "doom scrolling."
  • The founders reflect on whether casual gaming could mirror the rapid consumption patterns seen in short-form video content.

Future Prospects and Support Systems

  • The hosts express enthusiasm about future developments and plan to revisit progress in six months.
  • Discussion shifts towards operational support for startups, emphasizing the need for efficient management of back-office tasks such as incorporation and payroll.

Humanoid Robots: A New Era in Home Assistance?

Introduction to Humanoid Robots for Home Use

  • The speaker expresses a desire for more help with parenting tasks, specifically changing diapers, and introduces the concept of humanoid robots designed to assist in home environments.
  • Burnt Bernick from OneX is introduced as a guest who will discuss their robot, Neo, which aims to make life easier for parents.

Evolution of OneX and Its Robotics Journey

  • Bernick shares that OneX initially focused on industrial robots before pivoting towards home applications, highlighting their early model called Eve Industrial on Wheels.
  • The foundational goals of OneX include creating safe robots capable of learning and interacting with humans while being scalable and affordable.

Insights on Humanoid Robotics Development

  • Bernick compares the evolution of robotics to computers, noting that just as PCs became widespread after starting with mainframes, humanoid robots are moving towards general labor applications.
  • He emphasizes the importance of creating machines that can enhance human productivity by breaking free from traditional limitations.

Transition from Eve to Neo

  • The transition from Eve to Neo involved refining technology to create a safer and more versatile robot suitable for home use.
  • Bernick stresses the need for real-world impact in robotics; he believes that successful robots must operate outside labs and be integrated into everyday life.

Features and Capabilities of Neo

  • Neo represents an advancement over Eve by incorporating modern AI techniques, making it lighter, softer, and more intelligent—designed to interact with the world like a human.
  • The focus on mimicking human interaction extends down to minute details such as skin stiffness, aiming for seamless integration into daily activities.

Understanding General Intelligence in Robotics

The Importance of Human-like Interaction

  • The integration of human-like sensing and interaction is crucial for developing general intelligence in robots, allowing them to leverage humanity's vast knowledge encoded in various formats.
  • Current methods like teleyop are useful but insufficient for creating a robust dataset necessary for training models; innovative bootstrapping techniques are essential.
  • Recent advancements indicate that the approach being taken is promising, with ongoing projects demonstrating potential pathways toward achieving true general intelligence in robotics.

Task Flexibility and Learning Efficiency

  • Robots today often perform fine-tuned tasks based on specific data sets; however, the goal is to develop robots capable of handling a wide range of tasks effectively.
  • Neo, a home robot designed with human-like physical attributes, can learn from human interactions more efficiently due to its design mimicking human capabilities.

Enhancing Learning Through Human Mimicry

  • A robot's ability to learn quickly improves when it closely resembles human anatomy and behavior, facilitating better understanding and execution of tasks.
  • Transferring skills from humans to robots can be challenging if the robotic design does not align with how humans naturally interact with objects.

Emotional Connection and Companionship

  • The design philosophy behind Neo emphasizes not just functionality but also emotional companionship, enhancing user interaction beyond mere task completion.
  • This companionship aspect addresses loneliness among the elderly by providing assistance while fostering emotional connections through technology.

Technology as a Solution for Aging Populations

  • As society ages, technology like Neo can help alleviate loneliness among seniors by offering both practical assistance and companionship without replacing existing relationships.
  • Emphasizing dignity in aging, technology should enhance life quality rather than serve as a replacement for family or pets; it adds value similar to how pets enrich lives.

Understanding the Context of Robotics

The Importance of Context in Robotics

  • The discussion emphasizes the significance of context when understanding robotics, particularly referencing Bill Wat's work as foundational.
  • Acknowledgment that without proper context, one may struggle to grasp the nuances of robotic applications and their development.

Current State of Neo Robotics

Capabilities and Limitations

  • Neo is currently available for pre-order, with shipping expected later this year; it aims to perform general household tasks.
  • There is curiosity about how quickly Neo can improve its performance in household tasks compared to existing demos.

The Role of Teleoperation and Learning

Enhancing Robot Learning

  • The speaker expresses interest in teleoperation as a means for experts to assist Neo in learning new tasks effectively.
  • Emphasizes that robots must learn within diverse environments, starting from home settings to understand social contexts better.

Early Adoption Challenges

Expectations for Early Users

  • The early adopter program is acknowledged as a rough start for users, marking a historical first in home robotics.
  • Current capabilities include performing laundry and basic companionship tasks like opening doors for guests.

Modes of Operation

Autonomy Levels

  • Two operational modes are introduced: best effort autonomy mode allows voice commands for task execution.
  • An example highlights Neo's ability to interact with objects (e.g., reading a post-it note), showcasing its potential for general intelligence.

Probabilistic AI and Task Performance

Understanding AI Limitations

  • While Neo demonstrates impressive capabilities, it also faces challenges with consistency due to its probabilistic nature.
  • The world model enables Neo to approach various tasks sensibly based on voice commands but requires iterative learning through user interaction.

Safety Considerations in Task Execution

Risk Management Strategies

  • For high-risk tasks (e.g., handling fragile items), reliance on fleet scaling ensures higher success rates while identifying potential failures early.
  • The world model mimics human thought processes by simulating outcomes before executing actions, enhancing safety during task performance.

Safety in Robotics and AI Development

Importance of Safety Measures

  • The development team is focused on ensuring that robots take the safest possible paths and avoid tasks with undue risk, emphasizing ongoing work in this critical area.

Aspects of Robot Safety

  • Safety has two main aspects: physical safety (robots being light, low-energy, and soft) and AI safety (limiting robot actions to prevent dangerous tasks).

Limitations of Current AI Capabilities

  • Currently, robots are restricted from performing certain tasks like cooking or handling hot liquids due to safety concerns; however, there are long-term goals for improved capabilities with a strong safety profile.

World Models vs. Language Models

Understanding World Models

  • A world model is essential for general intelligence in robotics; it involves creating a representation of the environment that includes spatial and temporal dynamics.

Limitations of Language Models

  • Traditional language models provide a 2D snapshot without capturing time or 3D dynamics, which limits their effectiveness for physical AI applications.

The Role of Predictive Learning

Learning Through Interaction

  • Robots learn by predicting outcomes based on interactions with objects (e.g., dropping an item), which helps them understand physics and improve their operational capabilities.

Social Navigation Challenges

  • For effective navigation in social situations, robots must simulate human behavior within their world models to interact appropriately with people.

Future Directions in AI Development

Embodiment in AI Training

  • There is excitement about training models based on real-world data from robotic interactions rather than solely relying on web data; this embodiment approach could enhance future AI capabilities.

Expanding Real-world Testing

  • Increasing the deployment of early adopter Neos will expand their real-world footprint, allowing for better testing against reality and enhancing learning through practical experience.

Understanding the Role of NEOs in AI Development

Importance of Early Adoption

  • The speaker emphasizes the significance of early adoption for learning and improving AI intelligence, indicating that rough initial iterations are acceptable to expedite data collection.

Data Requirements for Learning

  • The discussion revolves around how many NEO robots are needed to gather sufficient data for effective learning. The honest answer is uncertain due to the unprecedented nature of this endeavor.

Comparison with YouTube Data

  • Pre-training on a dataset comparable to YouTube can significantly advance general intelligence, but lacks agentic behavior and physical interaction data which are crucial for development.

Richness of Robot Data

  • Unlike video data, robot data encompasses internal states, goals, actions taken, and results achieved. This richer dataset allows for potentially less quantity needed compared to video datasets.

Scaling Insights from YouTube

  • Approximately 10,000 robots could provide a similar influx of data as YouTube's upload rate. This number is significant yet manageable in terms of scaling up operations.

The Vision for Intelligent Robots

Sustainable Progress through Robotics

  • The ultimate goal is to create intelligent robots that can assist in building sustainable infrastructure without compromising on cost or labor efficiency.

Role of AI in Scientific Research

  • Current AI models excel at generating suggestions but struggle with executing experiments and validating results. A humanoid robot could bridge this gap by mimicking human laboratory functions.

Training Robots at Home vs. Industrial Use

  • While training NEO robots occurs in domestic settings, their eventual application may shift towards industrial environments where they perform tasks autonomously.

Manufacturing and Development Strategies

Complexity of Home Environments

  • Home environments present complex challenges that make them ideal starting points for developing capable robots before deploying them into more controlled settings like labs or factories.

New Manufacturing Facility Overview

  • A new facility has been established that integrates design and manufacturing processes under one roof, enhancing efficiency from raw materials to final product development.

Unique Approach to Production

  • The company prides itself on its comprehensive approach—from material development to production—allowing rapid advancements in technology essential for creating innovative robotic solutions.

The Future of Robotics: Insights on Production and Design

The Importance of Iteration in Robotics Development

  • The speed of iteration across the entire stack is crucial for advancements in robotics, especially when relying on suppliers.
  • Collaboration between hardware and software teams can lead to innovative solutions, such as using neural networks to improve assembly tolerances and reduce costs.

Manufacturing Capabilities and Demand

  • The factory in Hayward is fully operational and capable of producing tens of thousands of home robots annually, with plans for future scalability to hundreds of thousands.
  • There is cautious optimism regarding early commercial demand; the focus remains on product quality before ramping up production significantly.

Testing and Product Stability

  • Current testing involves deploying robots into homes under NDA agreements to gather data for design improvements.
  • The iterative process includes rapid retooling based on market feedback until a stable product is achieved, emphasizing the uniqueness of this manufacturing approach.

Economic Considerations in Robot Pricing

  • Questions arise about the $20,000 price point concerning profitability; understanding cost structures from day one is essential for scaling production effectively.
  • A parallel is drawn between electric and gasoline cars to illustrate the importance of making foundational decisions that align with long-term goals.

Material Resources and Design Efficiency

  • Consideration must be given to raw material availability; Neo's lighter weight reduces resource demands compared to competitors.
  • As production scales towards billions, concerns about sourcing sufficient materials like aluminum become critical for sustainable manufacturing practices.

Industrial Applications and the Future of NEO

Evolution of Manufacturing Technology

  • The NEO represents an advanced generation of manufacturing technology, allowing for fewer parts and looser tolerances, which simplifies production.
  • Automation is a key focus, with robots building robots in the factory, contributing to cost reduction despite the complexity of the system.

Cost Considerations and Market Viability

  • A target price point of $20,000 is seen as feasible for consumer purchase without extensive financial calculations.
  • The speaker expresses optimism about achieving significant production volume at this price point while addressing supply chain risks associated with components sourced from China.

Supply Chain Challenges

  • Key components like magnets pose challenges; however, innovative processes are being developed in collaboration with Chinese companies to enhance magnet production.
  • Efforts are underway to ensure that magnet production can also occur in the U.S., although it remains a work in progress.

Capital Requirements for Launch

  • Initial funding has been secured from Norwegian investors, but there is curiosity about future capital needs for launching NEO and its next generation.
  • While sufficient funds exist to ship NEO, additional fundraising may be pursued to accelerate development due to its potential impact on humanity.

Competitive Landscape and Future Aspirations

  • The transition towards general intelligence in robotics could revolutionize daily life by reducing physical labor demands.
  • There’s recognition that scaling operations will require substantial investment in computing power and manufacturing capabilities.
  • The company aims to leverage better data for training models compared to competitors, highlighting a strategic advantage in their approach.
Video description

This Week In Startups is made possible by: LinkedIn Jobs - LinkedIn.com/twist Every.io - every.io Render - render.com/twist Today’s show: Imagine scrolling through mobile games the way your flip through your TikTok feed. What if you could build your own game in 90 seconds and share it with the world? That’s what over 100,000 users are enjoying right now on Nanogram. On TWiST, we meet the co-founders: 22-year-old CEO Albert Brotherton and 24-year-old CTO Boris Radilov. They demo the app for us on the air, and Jason immediately spies the potential. Power users are playing 25+ games per session already. We’re digging into the co-founders future plans, including the potential for ads in your game feed, in this TWiST exclusive. PLUS domestic helper bots are here. NEO from 1X Technologies can do the laundry, open the door for guests, and even read and understand Post-It Note messages. Alex talks with CEO Bernt Børnich about building the safety-first 66-lb humanoid, why world models are so crucial for training robots in particular, and why homes are even tougher places for robots to navigate than factory floors. Timestamps: 0:20 Nanogram co-creators Albert and Boris join the show 3:03 Nanogram is TikTok for casual AI-generated games 3:48 Building Nanogram with Google Gemini 5:10 Draper and Associates: https://www.draper.vc 5:32 Integrating ads into the game feed 6:57 Roblox: https://www.roblox.com World of Warcraft: https://worldofwarcraft.blizzard.com 7:46 Building in New York 8:28 Jason is obsessed with the Staples Baddie https://www.tiktok.com/@blivxx 9:33 LinkedIn Jobs - Hire right, the first time. Post your first job and get $100 off towards your job post at https://LinkedIn.com/twist 19:57 Every.io - For all of your incorporation, banking, payroll, benefits, accounting, taxes or other back-office administration needs, visit https://every.io 21:01 1X CEO and founder Bernt Børnich joins the show https://x.com/BerntBornich 21:52 Designing robots to actually live around people 25:51 Teaching NEO about movement in the physical world 28:21 TechCrunch coverage: https://techcrunch.com/2026/01/13/neo-humanoid-maker-1x-releases-world-model-to-help-bots-learn-what-they-see/ 29:19 Render: Find out why 5 million developers are already using the all-in-one cloud platform, Render. Go to https://render.com/twist and apply for the Render Startup Program to get $500-$100,000 in free credits, depending on your stage and backers. 30:25 Moving beyond training data to true general intelligence 32:38 Ensuring NEO pursues the safest possible path 33:40 Why are world models so important? 41:12 1X World Model Challenge on GitHub: https://github.com/1x-technologies/1xgpt 44:35 Robots need to train on a lot of data… Where does it all come from? 49:35 Manufacturing NEO from raw materials in San Carlos 55:53 Robots building robots Subscribe to the TWiST500 newsletter: https://ticker.thisweekinstartups.com Check out the TWIST500: https://www.twist500.com Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp Follow Lon: X: https://x.com/lons Follow Alex: X: https://x.com/alex LinkedIn: ⁠https://www.linkedin.com/in/alexwilhelm Follow Jason: X: https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis Check out all our partner offers: https://partners.launch.co/ Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland Check out Jason’s suite of newsletters: https://substack.com/@calacanis Follow TWiST: Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Instagram: https://www.instagram.com/thisweekinstartups TikTok: https://www.tiktok.com/@thisweekinstartups Substack: https://twistartups.substack.com