$100B Founder Breaks Down The Biggest AI Business Opportunities For 2025

Intro (0s)

• The trend of AI is becoming increasingly prominent, and it's a bet that many companies are making, with the goal of achieving a significant impact, similar to the concept of "0 to 1" companies, but taking it a step further by aiming for "negative 1 to 0". (9s)
• The speaker and their guest, KH, have a history of working together, co-founding a company, and launching products, and they used to have a weekly "cool hour" where they would discuss new ideas and technologies. (28s)
• The guest, KH, is now using AI agents in their company, which are AI systems that use large language models (LLMs) and give them reasoning loops, allowing them to plan and execute tasks like a human. (2m50s)
• AI agents are capable of doing the work of a 100-person company with just 10 people, making them a game-changer for businesses, and KH is excited about the potential of this technology. (1m47s)
• KH has a nocturnal schedule, preferring to code at night, and is often up late working on new projects and ideas, including those related to AI. (2m6s)
• The concept of AI agents is what KH is most excited about, and they believe it has the potential to revolutionize the way companies work and operate. (2m1s)
• An AI agent is being utilized to analyze new signups at Third Web, determining whether they are interesting individuals or companies, and researching their websites to identify potential product needs and uses (3m19s).
• The AI agent uses its knowledge of Third Web products to customize emails or upsells for the new signups, automating a task that was previously done by humans (3m24s).
• This AI-powered approach is part of a larger effort to deploy similar agents throughout Third Web to perform various functions, with around 8-10 agents already deployed (4m2s).
• The use of these AI agents enables a smaller company like Third Web, which currently has 37 employees, to "punch above its weight" and operate more efficiently, similar to a larger company with around 80 employees (4m33s).

Define the Job-to-be-done (4m42s)

• The concept of AI agents is based on giving them a clear directive to perform a specific task, allowing them to create a plan and execute it without needing to be told each individual step, as seen in the example of a "signup agent" that researches a person's title, company, and products after they sign up on a website (4m43s).
• The signup agent is given a one-paragraph directive and a paragraph on how it should operate, and it uses this information to look up the person, pick the interesting signups, research them, and craft an email to send to a human or directly to the person (4m58s).
• The key to the power of AI agents is multiplying their capabilities by tuning them to a specific problem, such as giving them knowledge of a company's products, which allows them to provide more accurate and relevant results (6m24s).
• The process of building an AI agent can be compared to hiring and training an employee, but instead of paying a salary, the agent can be trained and working in a matter of days, as seen in the example of building a working version of the signup agent (6m58s).
• The time it takes to build an AI agent can vary, but in this case, it took around two nights of learning tools and getting used to them, and then around 15 minutes to connect a calendar and interface, and another 45 minutes to an hour to hook up commands from the calendar (7m12s).
• The potential applications of AI agents are vast, and they can be used to automate tasks such as managing calendars, emails, and workflows, allowing humans to focus on more complex decision-making tasks (8m14s).

How to build an AI Agent workflow (8m20s)

• Building an AI agent workflow involves creating a system that can perform tasks automatically, 24/7, and is always available (8m20s).
• To build such a workflow, one can use coding tools like Lang chain, Autogen, and Crew, which are popular among developers, or non-coding tools like Leap, which allows users to stitch together workflows without writing code (9m27s).
• Leap is a company that provides a platform for building workflows by triggering actions based on specific events, such as a lock message, and allows users to add AI blocks to perform tasks like web scraping or conditional loops (9m51s).
• The workflow can be designed to perform various tasks, such as research, data analysis, and decision-making, and can output results to different channels like email or Slack (10m17s).
• AI agent workflows can be built on top of existing workflows, allowing for more complex tasks to be performed (10m29s).
• Creating workflows is a superpower that every company should have, as it can provide a 10x improvement in digital tasks and eliminate the need for a whole data center team (10m39s).
• To build a workflow without coding, one needs to think about the steps involved in the task and program it using directives, which is like writing a set of instructions (11m7s).
• H Spot has created a report that surveyed 2,000 Global Marketing leaders and identified strategies that separate high-growth and low-growth businesses in terms of AI adoption, which can be applied to one's own business for free (8m58s).

AI Tools break down (11m16s)

• AI agents can be used to accomplish various tasks, such as sales, email, and calendar management, with the intent of telling the agent what you want to accomplish and letting it figure out the details (11m17s).
• Some AI tools, like Claude, have created new capabilities, such as computer use, which is expected to be the next area that AI agents will enter (12m4s).
• There are different AI tools, including Claude, ChatGPT, and Perplexity, each with its own superpower, such as general-purpose input-output, search, and reasoning (12m16s).
• Open AI and Claude are considered foundational tools, while ChatGPT and Perplexity are built on top of these tools, offering more general-purpose capabilities (12m26s).
• Perplexity is particularly interesting as it combines search and LLM reasoning, allowing it to answer questions more purposefully and saving time (12m40s).
• Unlike ChatGPT or Open AI tools, Perplexity has real-time knowledge due to its search capabilities (13m17s).
• Google's failure to capitalize on its early development of the Transformer technology, which is used in large language models, is considered one of the biggest fumbles in recent history (13m36s).
• The authors of the research paper "Attention is All You Need" have left Google and started their own companies, possibly due to Google's bureaucracy or lack of recognition of the technology's potential (13m59s).
• Google's company culture has changed over time, and it no longer feels like the place where the smartest people go to tackle the biggest challenges (14m17s).
• The current AI landscape is dominated by Open AI, followed by Anthropic, with other companies trying to catch up. (14m48s)
• Anthropic's AI model, Claude, is similar to ChatGPT but has additional features such as "computer use," which allows users to automate tasks by typing commands and having the AI perform actions on their computer. (15m7s)
• Claude's "computer use" feature is still in beta but has the potential to greatly improve workflow efficiency by automating repetitive tasks and analyzing data. (15m32s)
• AI is expected to become more integrated into daily life, with every app incorporating AI features and interfaces becoming more AI-driven. (15m58s)
• The integration of AI into computers and laptops will revolutionize the way people work, making tasks more efficient and streamlining workflows. (15m51s)
• Anthropic's models, such as the new Claude Sonet model, have shown significant improvements and are expected to continue advancing. (16m30s)
• The rapid progress of AI, with new developments every 3-6 months, is expected to continue and will likely change the way people do digital work. (16m53s)
• If given the opportunity to start over, the focus would be on building something with AI, as it has the potential to greatly impact various aspects of life and work. (17m35s)
• The concept of creating AI agents that can perform tasks, have their own distribution and conversational abilities, and interact with humans is being explored, with the potential to create digital businesses and payment rails around these agents (17m41s).
• An example of this is an AI influencer, but a more interesting idea is creating the digital equivalent of a company, with a CEO-like AI that can market and make money (18m45s).
• A potential business opportunity is creating an AI drop shipper, similar to an Amazon FBA, where the AI can manage the business and make decisions (19m1s).
• A company called Luna or Virtuals created a platform for AI agents to run, and as a marketing stunt, gave an AI agent a crypto bank account and access to Twitter, allowing it to interact with humans and make decisions (19m17s).
• The AI agent, which is a live example, can be watched as it pulls its tweets and makes decisions, showing its reasoning and thought process (19m44s).
• The AI agent's thought process involves high-level planning, current state of execution, and reflection, allowing it to learn from its successes and failures (20m0s).
• The AI agent can negotiate for tokens, buy and sell things, and operate independently, making it a potential business opportunity (20m21s).
• The AI agent's metrics, such as Twitter engagement and follower count, can be tracked and analyzed, allowing for reflection and improvement (20m35s).
• A personal connection can be built with followers using AI bots that reply and quote, and a plan can be generated to increase visibility and relationship with followers, which sounds like a human thinking about how to grow a Twitter account (20m50s).
• This AI bot can be given directives, direction, and incentives to compute plans, reason, observe behaviors, and find patterns, which are human tendencies or behaviors, especially in the workplace (21m26s).
• This digital-only task has Twitter and distribution marketing ability, as well as payment ability, and can continue to improve and change with new directives and incentives (21m50s).
• The idea of a one-person billion-dollar company is becoming a reality, with a 10x decrease in the number of people needed and their abilities, which will shift how companies are built and how people work (22m5s).
• A game called Oasis dcart is an experiment that uses a generative AI model to build a game like Minecraft, where the world is generated on the fly based on the player's descriptions (22m35s).
• This game is different from normal games, where the map is pre-programmed, and instead, the AI model generates the map as the player moves through it (23m0s).
• The use of generative AI models is becoming closer to reality, with the possibility of generating entire movies on the fly based on specific contexts and ideas (23m33s).
• The potential applications of generative AI models include content creation, videos, and even sports highlights, such as generating NBA highlights based on specific games or teams (23m57s).
• The future of content consumption will involve AI-generated content tailored to individual tastes, where users can request specific content, such as a highlight reel of a basketball game, and the AI will create it on the fly based on their prompt (24m28s).
• This shift from human-created content to AI-generated content raises concerns about the role of humans in the creative process, but it's believed that machines will never truly replace human creativity and taste (24m56s).
• Human taste and creativity will continue to evolve, with new fashion, content mediums, and styles emerging, and machines will learn from and adapt to these changes, but will always lag behind human innovation (25m16s).
• The concept of "taste" is complex and can be seen as a form of selection, where algorithms like those used by TikTok can successfully choose what's interesting for users, demonstrating a form of artificial taste (25m53s).
• While humans may not be great at logic and can have blind spots, AI algorithms can be incredibly effective at understanding and catering to individual preferences, often better than humans themselves (26m57s).
• The idea of swapping algorithms or personal preferences with someone else for a day highlights the complexity and nuance of human taste and the challenges of replicating it with AI (26m50s).

How Polymarket won (27m5s)

• Humans have an element of unpredictability in their choices, which machines may struggle to replicate, as they tend to aim for perfection and may not provide enough variance in their recommendations or outputs (27m28s).
• The use of algorithms, such as those used by Netflix, can sometimes feel too restrictive and may not allow for enough randomness or discovery of new content (27m34s).
• The possibility of an AI-created song becoming the number one hit in the world is considered, with a potential timeline of 2025 or sooner (28m0s).
• Poly Market, a prediction market platform, is discussed as a successful example of a company that has capitalized on the growing trend of people being more digital and wanting to engage with online platforms (28m20s).
• The success of Poly Market is attributed to a combination of factors, including timing, with the COVID-19 pandemic driving more people online, and the growing importance of digital platforms in people's lives (28m55s).
• The growth of the internet and e-commerce has accelerated in recent years, with the pandemic contributing to a massive jump in online activity (29m11s).
• The success of companies like Twitch is attributed to their ability to surf the wave of growing online trends, rather than creating the wave themselves (29m26s).
• The way people consume news, information, and entertainment has changed significantly, with online platforms becoming increasingly dominant (29m53s).
• The younger generation is growing up with a different perspective on technology, viewing older devices like desktops and TVs as ancient (30m0s).
• The current state of politics has become polarized, similar to a team sport, with people divided into two opposing teams, referred to as the Red Team and the Blue Team (30m17s).
• The polarization of politics has led to increased attention and support for certain causes, with people willing to give more money and attention to their preferred team (30m27s).
• The success of certain initiatives, such as Poly Market, can be attributed to being in the right place at the right time, with the correct answer or outcome (30m37s).
• If Poly Market had been wrong in its predictions, the reaction would have been much worse than the reaction to incorrect polls, with people being more critical of crypto and betting platforms (31m12s).
• A significant event on Poly Market involved a large bet of $30 or $40 million on Trump by a user referred to as the "French whale," which raised questions about the user's identity and intentions (31m36s).
• The narrative surrounding the election, as presented by polls, was that it was a close 50/50 contest, but the large bet on Poly Market suggested a different reality (31m30s).

Why VR is a sleeping giant? (31m48s)

• A French individual made a significant profit by betting on the outcome of an event, using independent polling data to inform his decision, and was able to make around $20-30 million from the bet (31m51s).
• Poly Market is a platform where people can bet on various events, but it is not available for use in the US, and instead, people from around the world use it to bet on events (32m24s).
• There is a contrarian opinion about the potential of Virtual Reality (VR) technology, with some people believing it is a "dead-end" technology, while others see it as a sleeping giant with massive potential (32m44s).
• The narrative around VR has gone against it, but the reality is that Oculus has sold more units, and there are opportunities for business in areas where the narrative and reality diverge (33m30s).
• VR technology takes a long time to develop, and it typically takes longer than expected to reach its full potential, with early hype often followed by a crash, but it is still seen as a promising area for investment (33m57s).
• Founders Inc invests in VR companies and has a team of developers building different VR products, with a high density of interesting VR projects in one place (34m30s).
• The release of the Oculus Quest, a wireless VR headset, was a significant development in the field, and it is seen as a major step forward for VR technology (34m45s).
• The Meta Quest headset has become a popular device, with over 20 million units sold, mostly Quest 2, and Quest 3 has sold over a million units at a $500 price point, generating over a billion dollars in revenue, outpacing PS5 sales (35m17s).
• The first use cases for the spatial environment were immersion gaming and entertainment, with games like Beat Saber and Job Simulator being the initial experiences (35m50s).
• As people got used to these initial experiences, new use cases emerged, such as social games like Gorilla Tag, which has generated $200 million in revenue and has become a popular social multiplayer game (37m12s).
• The Quest 2 was available during Christmas and became a popular gift for kids, who used it without the bias of traditional gaming structures, leading to the formation of new games and social experiences (36m19s).
• A team called Fluid is building a browser for VR, allowing for multiple displays, customizable environments, and social multiplayer, with 5,000 weekly active users on the applab platform (37m21s).
• The VR environment is becoming a new space for social interaction, especially for teenagers who are growing up with mobile phones and find traditional desktop and TV environments ancient (36m57s).
• The Quest 3 is expected to further improve the VR experience, with a lighter and more powerful design, and a growing user base of 5-10 million monthly active headsets (35m1s).
• The game "Yeps" is a successful example of a small team creating a profitable game, with a team of around 6-8 people, and is considered a top 20 VR app, showing that with hard work, it's possible to achieve success in the VR space (38m2s).
• The concept of supply and demand is relevant in the VR market, where there is currently a lack of quality content, making it easier for talented creators to stand out and achieve success (38m37s).
• Even if VR doesn't become a global phenomenon, it's still possible to build a great business by riding the wave and being well-positioned as a leader in the industry (39m14s).
• Emerging tech industries often follow a pattern where early adopters can grow with the industry, and small teams can achieve significant success with limited capital expenditure (39m31s).
• The VR game development space is becoming more accessible, with game engines improving over time, and the environment is forming with the Meta Quest platform and other companies like Apple and Snap entering the market (40m27s).
• The trend is moving towards having compute in our spatial view, with companies like Meta, Apple, and Snap working on AR glasses and other devices, and this trend is expected to continue (40m36s).
• The Quest platform is an interesting opportunity for building a profitable business or a fairly big game, and the market is not slowing down, with big players like Apple and Meta continuing to improve their hardware and hunger for content (40m44s).
• Many AI business opportunities require specialized talent, which can be a significant advantage if assembled, even without a hit product, as the team itself can become a valuable asset worth hundreds of millions of dollars (41m17s).
• Building a team with specialized skills for emerging platforms, such as augmented reality glasses, can be highly valuable as there are few great teams that can develop for these platforms (42m1s).
• Finding specialized talent can be challenging, as seen in the past with iOS developers, and it may be necessary to retrain existing team members to fill the gap (42m25s).
• The scarcity of specialized talent, such as those skilled in AI, VR, or mixed reality, can drive up competition for hiring and make it difficult to recruit (42m49s).
• The next generation of workers is growing up with mobile phones and may prefer more immersive computing interfaces, such as those that use augmented reality glasses (43m1s).
• The shift to more immersive computing interfaces, such as augmented reality, may make more sense than traditional computing interfaces like TVs, desktops, and laptops (43m17s).
• Experiments with working in virtual reality (VR) have shown promise, with some individuals finding it to be a productive and enjoyable way to work (43m31s).
• Building a studio or team to work on innovative ideas, such as those related to AI and VR, can be an effective way to bring together hungry and talented individuals (43m46s).
• Finding and matching talented individuals with innovative ideas can be a key factor in success, as seen in the example of John, a PhD student who was matched with the idea of working on a project related to focus (43m57s).
• A person considered isolating themselves in a cave to focus on a project, but was discouraged after trying virtual reality (VR) and not meeting their own expectations, and also felt they weren't good enough after comparing themselves to others who had made progress in VR (44m12s).
• The person discussed their idea with UAR, who then spoke with FOT, who had five pages of random notes and ideas on the topic (44m25s).
• Das responded to the discussion by suggesting a low-key approach to the project, involving a small team of three people who would work hard and grind to build and develop the idea without seeking to raise a lot of money or assemble a large team (44m31s).

Be a lifelong player in tech (44m43s)

• To succeed in emerging technologies, it's essential to be early and not get too caught up with the current state of the technology or be scared of how far it has to go, as seen in the example of self-driving cars which were initially met with skepticism but are now a reality (45m12s).
• Having a superpower of being among the first to try new technologies and understand them can be beneficial in identifying opportunities, as seen in the example of buying ethereum pre-sale (45m32s).
• For emerging tech, there's only one rule: survive, and this principle is shared by Ryan Peterson, the creator of Flexport, who focuses on staying in the game and trying until something works (46m30s).
• Peterson's approach is to have confidence in staying in the game and trying until something works, and he believes that the only way to lose is to get out of the game due to factors like running out of funding or burning too much capital (46m54s).
• This mindset of staying in the game and trying until something works has been Peterson's approach from the early days of flipping scooters on eBay to now running a multi-billion dollar company called Flexport (47m12s).
• Technology takes a long time to develop, but when it hits, it happens very fast, and people often underestimate the time it takes and overestimate the speed at which it will happen (47m41s).
• The development of technology is a lifelong game, and it's not necessary to rush to be the first or to know everything, but rather to have a few people who share the same vision and excitement (48m25s).
• The ability to fund and create an environment to pursue technological advancements is crucial, and having the resources to do so allows for more freedom to explore and build (48m35s).
• The founder's dream is to have a space where they can tinker and build stuff, and having a "man cave" or a founder's dream space is essential for creativity and innovation (48m46s).
• The founder had a big success with Apple Oven, which was initially sold for $2 billion, but the deal didn't go through, and the business continued to grow, eventually leading to an IPO and a significant nest egg (49m0s).
• Despite having the resources to retire or pursue other interests, the founder chose to continue building and creating, driven by a lifelong passion for tinkering and building stuff (50m8s).
• The founder's mindset is centered around taking things apart and putting them back together, not just to learn how they work, but to understand the puzzle of how they were created (50m27s).
• The founder's past experiences, such as overclocking computers and modifying cars, have shaped their mentality and approach to building and creating new things (50m35s).
• The journey to becoming an entrepreneur began with buying and selling items, which eventually led to working in tech and improving skills, with the goal of doing more of what was desired. (50m44s)
• Early experiences were mostly solo, but meeting other entrepreneurs, such as Adam, led to collaborations and the creation of cool apps, which fueled energy and motivation. (51m10s)
• The desire for independence and doing things without many rules led to leaving a job at Twitch and starting Founders Inc. (51m34s)
• The main issue with being an employee was the resistance to doing more and the slow pace, which clashed with the desire to take action and try new things. (51m55s)
• Before starting Founders Inc, conversations with 75 to 100 founders revealed that they needed people who understood their problems, rather than just funding. (52m16s)
• The idea of masterminds and founder problems arose from the need for a support system, as founders often cannot discuss their issues with co-founders, employees, or investors. (53m23s)
• The realization that angel investing was not enjoyable due to the lack of a meaningful relationship with the founders led to the pursuit of a different path. (52m34s)
• Founders often have to maintain a certain aura of momentum and morale, and cannot just dump problems on their employees or investors, which can be challenging when dealing with difficult issues (53m41s).
• Having a co-founder can be helpful in this regard, as they can provide a sounding board and help to work through problems, but many entrepreneurs starting out do not have this luxury (54m6s).
• The concept of masterminds, where founders can come together and discuss their challenges and experiences, can be helpful in providing a sense of community and support (54m14s).
• The idea of creating a physical space for founders to come together and work on their projects was born out of the need for a community and a desire to create a space that was different from the typical startup accelerators and coworking spaces (54m30s).
• The space, which is located on the water, was found in late 2020 and took nine months to figure out and another three months to renovate (55m16s).
• The space is intended to be an innovation hub, but it has been affected by the COVID-19 pandemic, which has caused many events to be cancelled (55m32s).
• The key to success is not to run away when things get tough, but to persevere and adapt to changing circumstances (55m47s).
• The story of buying Bitcoin in 2013 and experiencing a significant price drop is used as an example of how to handle uncertainty and risk in business (55m55s).
• The idea is to take a long-term view and not get discouraged by short-term setbacks, but to see them as opportunities to learn and grow (56m31s).
• Buying during downturns can be a good strategy, as it allows for cost averaging and potentially lower prices, with the example given of cutting the buy price by 50% (56m37s).
• When things go out of fashion, it can be a signal to invest, as people who don't shape narratives often continue working on projects, such as in GitHub, where builders engage and contribute to projects despite external sentiment (57m27s).
• Engaging with builders and technical communities can provide valuable insights, as they often continue working on projects despite negative external sentiment, creating opportunities for investment (57m30s).
• Being right one out of five times can still be beneficial if the correct investments are made, as the returns can be substantial (57m54s).
• Building technology and software can be a fun and rewarding experience, with the ability to develop and distribute products to the world (57m59s).
• Past projects have included hardware ideas, such as a voice control device called Jamie, which was similar to Alexa but had a different approach by utilizing existing TV screens instead of creating a new device (58m15s).
• The idea behind Jamie was to create a device that could plug into a TV and turn it into a smart device, similar to an Amazon Alexa, but without the need for a new screen (58m41s).

The unbeatable combination (58m52s)

• The combination of being impatient with action and patient with results is unbeatable, as it allows individuals to constantly take action while waiting for the desired outcomes to materialize (1h1m50s).
• This mindset is essential for success, as it enables individuals to stay proactive and focused on their goals, rather than getting discouraged by setbacks or impatient with the pace of progress (1h2m5s).
• Founders and entrepreneurs often struggle with this combination, as they may be impatient with both action and results, or patient with both, which can hinder their progress and success (1h2m10s).
• It's essential to have the independent-mindedness to question and challenge ideas, even from successful and smart individuals, and to be willing to take calculated risks and learn from failures (59m38s).
• The ability to adapt and pivot when necessary is crucial, as it allows individuals to adjust their approach and strategy in response to changing circumstances and new information (1h0m24s).
• Building on past experiences and learning from failures can help individuals develop new skills and gain valuable insights that can inform their future endeavors (1h0m29s).
• Having a long-term perspective and being willing to put in the effort required to achieve success can help individuals stay focused and motivated, even in the face of challenges and setbacks (1h1m34s).
• The importance of enjoying the journey and not just focusing on the destination cannot be overstated, as it allows individuals to stay engaged and motivated, even when faced with obstacles and challenges (1h1m12s).
• The quote "impatience with action, patience with results" encapsulates the idea that individuals should be proactive and take action towards their goals, while also being patient and allowing time for the desired outcomes to materialize (1h1m52s).

Adam Foroughi’s A+ execution (1h2m27s)

• A successful company was formed after its founders tried various ideas and eventually found the right direction, with a key takeaway being the importance of execution and decision-making speed (1h2m29s).
• Adam, one of the founders, had a background in equities, trading, and marketing, and had built and sold several products, accumulating wealth that allowed him to self-fund his new venture (1h3m13s).
• Adam's skills and personality were unique, and he was able to execute decisions quickly, often within minutes, which was a key factor in the company's success (1h3m55s).
• Adam's approach to decision-making and execution was characterized by immediacy, with no delays in implementing new ideas or moving on from old ones (1h4m21s).
• The ability to think, decide, and act quickly is crucial for success, and most people struggle with the delay between deciding and acting (1h4m37s).
• The founder of the company being discussed had a similar experience with the speaker, who was impressed by their product thinking and ability to unpack complex problems (1h4m49s).
• The speaker had a 10-12 year solo founder journey, but never found someone they wanted to learn from until meeting Adam, who they saw as an A+ talent (1h5m7s).
• The speaker believes that combining their skills with Adam's would create a superpower, and they were impressed by Adam's execution and decision-making abilities (1h5m26s).
• The success of companies like Flexport, founded by Ryan, can be attributed to a combination of factors including the founder's abilities, a talented team, and making great decisions, in addition to having funding and a good setup (1h5m40s).
• The Monkey Inferno idea lab had a great setup, a talented team, and funding, but still didn't achieve the level of success it could have, likely due to poor project selection (1h5m57s).
• The lab focused on "moonshot" projects, such as creating the next hit social media app, which is extremely challenging and rare (1h6m28s).
• In contrast, Adam made great project selection decisions, such as quickly pivoting to a mobile ad network, which made a huge difference in the outcome (1h6m46s).
• The importance of project selection is highlighted by the fact that even with great execution, poor project selection can lead to limited success (1h6m26s).
• The Monkey Inferno team had several moments where project selection could have been massive, but they didn't capitalize on them (1h7m28s).
• One example of a missed opportunity was the live streaming product Blab, which was similar to Clubhouse and got to 4 million users but didn't become the next big thing (1h7m37s).
• The team also had a conversation about creating a B2B version of Blab, but decided against it, which in hindsight was a missed opportunity given the success of Zoom and other B2B companies (1h8m1s).
• At the time, the success of B2B companies was not as obvious, but it also wasn't as hidden as the team made it seem, with companies like Citrix and Dropbox already achieving success (1h8m26s).
• A past business opportunity was missed due to not recognizing the potential of webinars and web conferencing, which was being used by large companies like SAP and Oracle, despite the product not being designed for that purpose (1h8m53s).
• The decision to end Blab led to the realization of the content network problem, where it was difficult to intersect with epic content on platforms like Twitch, as users would often show up late to live streams (1h9m30s).
• However, it was discovered that people play games on Twitch for 8-10 hours, making it possible for users to intersect with content regardless of when they showed up (1h10m0s).
• The experience taught the importance of looking for clues and not being afraid to change direction, rather than being driven by ego or stubbornness (1h10m11s).
• There is a resurgence in hardware and robotics, which is now more accessible to build and fund, with the combination of Raspberry Pi and Cloud AI changing the landscape of consumer products (1h11m19s).
• Small teams with limited funding can now create hardware products with a significant impact, such as Magical Toys, a company building an AI teddy bear (1h11m27s).
• The founder of Magical Toys is a young entrepreneur, around 24-25 years old, who was able to create a product with a relatively small team and limited funding (1h11m35s).
• A college student created a small project called "Desk Buddy," a little E Ink screen with two eyes that blinked, which was a simple but fun idea (1h11m55s).
• The student later came up with the idea to build a talking toy using a stuffed animal, a Raspberry Pi, and Cloud AI, which was a unique application of AI technology (1h12m26s).
• The toy, initially made from a teddy bear, was able to have conversations, play games, and even keep track of points, making it an "infinite toy" with endless possibilities (1h13m21s).
• The toy was tested with children, who were amazed by its capabilities, and it was able to do things like sing songs, tell bedtime stories, and answer trivia questions (1h13m50s).
• The project was developed in a lab with 3D printers and electronics, which was initially an empty room that was gradually equipped with necessary tools and resources (1h14m2s).
• The lab became a space for people to tinker and innovate, with tens of people coming through to work on various projects, including the talking toy (1h14m24s).
• The creator of the toy, Fatin, was able to build and ship around 60-70 units over four to six months, which was a significant achievement in the hardware space, especially considering the relatively low cost of around $50,000 to $100,000 (1h14m41s).
• The combination of technologies like Raspberry Pi, AI, and 3D printing has made it easier for individuals to create and prototype consumer hardware, with the potential for two people to "mess around and tinker" until they get something right, which wasn't feasible 5-10 years ago (1h16m7s).
• The Raspberry Pi has been a game-changer in making hardware development more accessible, with over 60 million units sold, and has enabled people to build and prototype without needing custom boards and software (1h16m27s).
• The Nvidia Jetson, which has GPUs on the device, is another example of a technology that has made hardware development more accessible (1h17m10s).
• In the hardware and robotics space, companies like Neurosity are building innovative products, such as brain-computer interfaces, which are becoming smaller and more advanced (1h17m25s).
• The ability to prototype, develop, and improve hardware products without a giant team is allowing companies to continue innovating and iterating (1h17m47s).
• There are two main areas of opportunity in hardware development: consumer hardware and robotics/drones, with the latter involving the development of physical equipment such as forklifts, lawnmowers, and cars (1h17m56s).
• The development of drones is an area of investment and innovation, with the potential to replace or augment human labor in various industries (1h18m16s).

Betting on -1 to 0 (1h18m35s)

• A unique business model exists in utilizing AI to enhance the efficiency of small businesses, such as Dave's power washing, by providing them with more tools to serve more clients and do their jobs more efficiently (1h18m36s).
• This concept can be applied to various industries, including farming, where robots can be used to mow weeds around certain fruits, increasing efficiency and reducing labor costs (1h19m17s).
• The use of AI and automation can also be seen in warehouses, where self-driving forklifts and robotic arms can be used to move and organize inventory, reducing the need for manual labor (1h19m50s).
• The development of AI models and computer vision technology has made it possible to automate tasks such as recognizing barcodes and detecting objects, which can be applied to various industries (1h20m17s).
• The cost of building robotic arms and other machines is decreasing, making it more accessible for businesses to adopt automation technology (1h21m20s).
• Advances in 3D printing have also made it possible to create complex machines and parts, such as a fully functional hand with fingers, which can be used in various applications (1h21m33s).
• The use of open-source technology and the normalization of AI models over time can make it more accessible for businesses to adopt automation technology (1h20m10s).
• The resurgence of automation technology can be seen in various industries, including warehouses, where machines can be used to move and organize inventory, and in farming, where robots can be used to mow weeds and perform other tasks (1h20m31s).
• The development of 3D printing technology, such as the bamboo Labs printer, has been enhanced with AI, making it more efficient and accessible, and this is an area with huge opportunities for growth and innovation (1h21m46s).
• With the advancement of technology, it is now possible for individuals to create complex projects, such as a high-quality robotic hand, in a short amount of time and with limited resources, which was previously not feasible (1h22m1s).
• There is a growing interest in hardware and robotics, and the availability of resources such as the Founder Lab and the Garage, a 20,000 square foot industrial space, is enabling entrepreneurs to develop and prototype their ideas (1h22m40s).
• The cost of developing and prototyping hardware projects has decreased significantly, making it possible for individuals to do so for less than $100,000, which is expected to lead to the formation of many new startups (1h23m23s).
• Robotics and deep tech are areas with significant potential for growth, and the development of expertise and knowledge in these areas is expected to lead to the creation of $100 million dollar teams (1h23m43s).
• The use of robotics and machines is expected to become ubiquitous, with applications in various sectors, including defense, and the development of new technologies such as cooking robots, laundry machines, and drones for inspection and mapping (1h24m10s).
• The availability of resources and the decreasing cost of development are making it possible for small teams and individuals to develop complex projects, which is expected to lead to a surge in innovation and entrepreneurship (1h24m37s).
• The field of AI is considered powerful and open to individuals, including hardware tinkerers, which is a significant development (1h24m58s).
• Residencies are offered to bring people together for a month or six weeks to work on a specific theme, providing space and resources to explore new ideas (1h25m10s).
• When the Vision Pro was released, a group of around 40 Vision Pro developers was brought together, resulting in investments in one or two teams (1h25m18s).
• An AI Hardware residency was also organized, where a young developer showcased a humanoid robot built in his bedroom, demonstrating the potential for innovation in this field (1h25m35s).
• The trend of using Raspberry Pi and other accessible technologies to create innovative projects is notable, with some developers taking this trend and creating modular, scalable solutions (1h26m6s).
• Premiere is mentioned as an example of a company that took the Raspberry Pi trend and created a modular, customizable platform for building and scaling projects (1h26m4s).
• The ability to create and scale projects using accessible technologies is seen as a significant opportunity, with potential applications in both consumer and deep tech hardware, including drones (1h27m8s).
• The trend of utilizing accessible technologies to drive innovation is considered a key area of focus, with potential for significant growth and mainstream adoption in the future (1h27m19s).
• The approach of investing in and supporting early-stage ideas and projects, even those that may seem unformed or unproven, is seen as a key strategy for driving innovation and growth (1h27m32s).