AI has arrived at a miraculous time for the economy. For the last 50 years, technology changed slowly while population growth declined. Without AI, there would be a panic about the future. Instead, robots and AI are appearing exactly when they are needed most. This means the remaining human workers will be at a premium rather than a discount.

We're going to have AI and robots precisely when we actually need them. The remaining human workers are going to be at a premium, not at a discount.

Marc views AI as a modern version of the philosopher's stone. It takes silicon, which is essentially sand, and turns it into human thought. This shift is changing how companies are built. Some leading founders are even exploring how to run entire companies alone. While the public focuses on job loss, it is more helpful to think about task loss. The job often survives even when specific tasks are taken over by machines.

The roles of engineer, designer, and product manager are merging into what Marc calls a Mexican standoff. Because of AI, everyone in these roles now believes they can do the work of the others. This intuition is largely correct. The value of a person who masters two or three of these domains is not just doubled or tripled. The impact is much higher because they become a super relevant specialist. To stay ahead, individuals should spend their time using AI to train themselves and improve their careers.

Now we have a technology that transfers the most common thing in the world, which is sand, converted into the most rare thing in the world, which is thought.

The current era is a historic turning point. Marc observes that trust in legacy institutions is in full scale collapse. These structures have proven they are not up to modern challenges. At the same time, the global conversation has become liberated. There is an incredible revolution in freedom of speech and thought. People can now openly discuss topics that were off limits just a few years ago. Marc expects this trend toward broader discourse to continue.

This is a very historic time. I think 2025 was maybe the most interesting year in my entire career and probably life. I would expect 2026 to exceed that. We have this incredible revolution in freedom of speech, freedom of thought, and the ability for people to openly discuss things that maybe they couldn't discuss even a few years ago.

Massive geopolitical shifts are also happening across the US, Europe, China, and Latin America. Long held assumptions are being pulled into the daylight and re examined. These shifts are colliding with the rapid rise of AI and a new level of citizen participation in public debate. Marc compares the magnitude of these changes to the fall of the Berlin Wall or the end of World War II. He believes we are only at the beginning of these three major forces colliding.

There is a common belief that we are living through a period of extreme technological change. However, statistical evidence tells a different story. Economists measure the impact of technology through productivity growth. In the United States, this growth has been running at about half the pace seen between 1940 and 1970. It is only about a third of the pace seen from 1870 to 1940. Technology impact in the real economy has actually slowed down significantly over the last several decades.

Statistical evidence shows that technology impact in the economy has actually slowed way down. AI is going to hit, but it is hitting an environment in which we have actually had almost no technological progress in the actual economy for a very long time.

This slowing progress is happening alongside a demographic collapse. The rate of human reproduction is in rapid decline across the West and in countries like China. Many nations are expected to depopulate over the next century. This means AI is entering a world with low productivity growth and fewer people. Marc believes we actually need AI to work to get economic growth back up. We will need machines to perform the roles that we simply will not have enough people to do as the planet depopulates.

We are going to need machines to do all the jobs that we are not going to have people to do because we are literally going to depopulate the planet over the next hundred years.

When considering the impact of AI on individuals, specifically children, there are two distinct paths. AI can take someone who is already proficient and make them very good by raising the average performance across the board. However, the more exciting transformation happens for the truly talented. These individuals can use AI to become spectacularly great and super productive. This creates a super empowered individual who can harness these tools to reach levels of output that were previously impossible.

AI is going to take people who are good at doing things and it's going to make them very good at doing things. But then there's this other thing where the really great people are becoming spectacularly great. If you're very good at it and you can really harness AI, you can become spectacularly great and super productive.

A central goal in raising children today should be fostering agency. While the modern education system often prioritizes following rules, the real premium in life is on taking responsibility, leading projects, and creating new things. Marc notes that while structure is necessary, AI acts as the ultimate lever for a child with agency. It allows them to be primary contributors in the world, whether in physics, coding, or art, rather than just passive participants.

The concept of AI is compared to the historical pursuit of alchemy. Early scientists like Isaac Newton spent decades obsessed with the philosopher's stone, a theoretical process to turn common lead into rare gold. AI represents the realization of this dream, transmuting sand into thought. It takes silicon, the most common material, and converts it into the most valuable resource: intelligence. For parents, the priority is ensuring children know how to leverage this modern philosopher's stone to its full potential.

We literally with AI, have a technology that transfers sand into thought. The most common thing in the world, which is sand, converted into the most rare thing in the world, which is thought. AI, it is the philosopher's stone. It actually is that.

Education is often viewed through two different lenses. At a national level, we focus on how to educate every child through large systems like the K12 school system. At the individual level, however, the goal is to maximize the potential of a single student. For centuries, the ideal method for teaching an individual has been one on one tutoring. This is the approach royal families and aristocrats have used throughout history. Aristotle tutored Alexander the Great, who eventually took over the world.

The ideal way to teach a kid by far is with one on one tutoring. If you have an individual kid and the goal is to maximize an individual kid, by far you get the best results with one on one tutoring. It has been known for centuries.

There is statistical evidence to support this called the Bloom two sigma effect. This research shows that one on one tutoring can raise student outcomes by two standard deviations, moving a child from the 50th percentile to the 99th. This success happens because of a tight feedback loop where the student stays on the leading edge of their capability and receives corrections in real time. Until now, this was only economically possible for the very wealthy.

Marc believes AI now makes this elite level of education available to everyone. A child can use an AI to explore interests, ask endless questions, and get instant feedback. You can even ask the AI to simplify a topic or quiz you on what you have learned. Marc sees a future where a hybrid approach becomes common, combining traditional schooling with AI tutoring to provide high quality education to many more families.

A kid that is super interested in something can talk to an LLM about it and they can ask an infinite number of questions and they can get instantaneous feedback. You can even tell an LLM, 'Teach me how to do the following,' and you can say, 'I do not quite understand what you are saying. Dumb it down for me a little bit. Now quiz me.'

Many people worry that AI will replace jobs and leave young people with fewer opportunities. Marc argues that this view is too simplistic. He points out that for the last fifty years, we have lived through a period of very slow technological change. Even if AI triples productivity, it would only return the economy to the growth rates seen between 1870 and 1930. During that historical period, people did not see technology as a threat but felt the world was full of new opportunities and career paths.

Even if AI triples productivity growth in the economy, which would be a massively big deal, it would take us back to the same level of job churn that was happening between 1870 and 1930. People just thought the world was awash with opportunity at that rate of technological transformation.

The timing of AI development is actually fortunate because of global population trends. Many countries are facing shrinking populations and less immigration. Without AI to boost productivity, these economies would likely shrink and stagnate. Instead of human workers being replaced, they will likely be at a premium because there will be fewer of them available. AI and robots are arriving exactly when they are needed to keep the economy from declining.

If AI leads to massive productivity growth, the basic rules of economics suggest a very positive outcome. When technology allows us to create more output with less input, the result is a surplus of goods and services. This causes prices to collapse. When the cost of things like housing, healthcare, and education drops, everyone effectively gets a huge raise because their money goes much further.

In this scenario, there is no version where everybody is just poor. In fact, it is quite the opposite. Everybody gets a lot richer because prices collapse. And then it is actually much easier to pay for the social safety net for the people who for some reason cannot find a job.

Marc reflects on his past predictions and acknowledges that he has grown to appreciate a different perspective on technological progress. While software and digital technology have advanced rapidly, the physical world has remained largely stagnant. This is the distinction between progress in bits and progress in atoms. Between 1870 and 1930, the built environment underwent a total transformation. In contrast, the world today looks remarkably similar to the world of fifty years ago. Many of our cities, bridges, and dams are decades or even a century old.

In the last fifty years, there has just been very little technological innovation in most of the economy. Anything involving atoms has seen very little real world technological change. The built world is just not that different today than it was fifty years ago.

This stagnation is often the result of structural barriers rather than a lack of invention. Red tape, bureaucracy, and political restrictions prevent new ideas from taking root. Many industries are controlled by cartels or monopolies that resist rapid change to protect their status. For example, AI could likely serve as a highly effective doctor, but it cannot legally practice medicine or prescribe drugs. These regulatory and economic structures act as a brake on progress. There is hope that the arrival of powerful new technology like AI might force us to rethink these old systems and finally choose to build for the future again.

The traditional roles of product manager, engineer, and designer are entering a phase Marc calls a three way Mexican standoff, similar to a scene from a John Woo movie. In this scenario, individuals in each role are pointing their figurative guns at one another. Every coder believes they can now be a designer and a product manager. Every product manager thinks they can code and design. Marc observes that they are all actually correct. AI is now a capable coder, designer, and manager.

Every coder now believes they can also be a product manager and a designer because they have AI. Every product manager thinks they can be a coder and a designer. And then every designer knows they can be a product manager and a coder. They believe that with AI, they do not need the other two roles anymore.

This shift marks the rise of the super empowered individual. Instead of remaining in specialized roles, talented people will use AI to become proficient in all three areas. For an engineer, the job changes from writing code by hand to orchestrating a dozen instances of coding bots. The most valuable people in the future will be those who can harness AI to design and build entire products from scratch. This transition moves tech beyond the siloed roles that have defined the industry for the last thirty years.

The atomic unit of work is the task rather than the job. A job is a bundle of tasks. While people often fear job loss, the real change is task loss. For example, executives 50 years ago never used typewriters. They dictated notes to secretaries. Today, those executives type their own emails. The job of the administrative assistant still exists, but the tasks involve complex planning instead of typing memos. The jobs remain, but the tasks within them change.

The job persists longer than the individual tasks. As the tasks change enough, that is when the jobs change at the level of an individual. You want to think of your job as a bundle of tasks. You must be ready to swap those tasks out and adapt to new technology.

This evolution is clear in the history of programming. The word calculator originally referred to people who performed math by hand in large rooms. Programming then moved from machine code to punch cards and then to languages like C. Later, scripting languages like Python took over. Each new step added a layer of abstraction. Marc notes that AI is simply the next layer in this history.

AI coding abstracts away the process of actually writing the scripting code. This is the next layer of task redefinition under the job of programmer. Now, the job is not necessarily to write code by hand.

Engineers are moving toward a world where they do not write code at all. Instead, they will act as orchestrators. The best programmers are already managing several AI coding bots at once. Their work involves directing these bots to get the correct output. The role is shifting from manual labor to high level orchestration.

The role of a programmer is changing from writing every line of code to managing and arguing with coding bots. This shift does not make traditional coding knowledge obsolete. Marc emphasizes that if you do not know how to write code yourself, you cannot evaluate the quality or speed of what an AI produces. He encourages his own son to learn the fundamentals even while using tools like Claude and ChatGPT to build games.

You need to still fully understand and learn how to write and understand code, because the coding bots are giving you code if it doesn't work or if it's not doing what you expect or it's not fast enough or whatever. You need to be able to understand the results of what the AI is giving you in the same way that somebody who's writing scripting language code does need to understand ultimately how the microprocessor works.

This up-leveling of capability makes technical depth more important than ever. While AI handles the bulk of the manual work, a top-tier software developer needs to understand every layer of the stack. This includes assembly, machine code, and how the physical chip functions. Knowing how the machine works allows you to get more value out of it. Marc suggests that while AI can generate endless amounts of mediocre code, creating truly significant technology requires a deep understanding of the underlying systems.

If the goal is to be a mediocre coder, then just let the AI do it. It's fine. The AI is going to be perfectly good at generating infinite amounts of mediocre code. If the goal is I want to be one of the best software people in the world and I want to build new software products and technologies that really matter, then you 100% want to go all the way down to assembly and machine code.

The current shift mirrors the historical rise of scripting languages like Perl. In the past, developers who wrote in C often looked down on scripting languages. Eventually, scripting languages became the industry standard. The most successful people were those who understood both the high-level scripting and the low-level systems. AI now acts as an incredibly fast tutor for these concepts. If a programmer needs to understand a complex topic like memory management, they can spend ten minutes with an AI to learn the theory while the AI helps implement the solution.

AI will soon handle task-level design work like creating icons. This shift allows human designers to focus on capital D design. This higher level of work involves questioning what a product is for and how it functions for human beings. Great designers think about whether a tool makes people happy or fits into the rest of their lives. AI is a tool that allows a young designer to spend more time on these important components instead of getting bogged down in basic tasks.

The job of designer will involve much more of those higher level, more important components and then again with AI doing a lot more of the underlying tasks. A young designer can be so super empowered by this technology to be able to do so much more. So much more of my time and attention is going to be able to be focused on these higher level things that most designers never get to.

Marc suggests that modern technology allows designers to reach the level of the best in their field more quickly. By harnessing AI, a designer can focus their attention on the higher level things that most designers never reach. This fits into a broader T-shaped strategy for success. In this model, an individual becomes an expert in their specific role, such as engineering or design, while gaining enough proficiency in other areas to collaborate effectively.

Success often comes from combining two or more distinct skills rather than being the absolute best at just one. Marc explains how Scott Adams became spectacularly successful with Dilbert not because he was the world's greatest cartoonist or the best business person, but because he was a cartoonist who deeply understood business. This additive effect of being good at multiple things is more than double or triple the value because it creates a rare specialist in the combination of domains.

The additive effect of being good at two things is more than double. The additive effect of being good at three things is more than triple because you become a super relevant specialist in the combination of the domains.

This principle is evident in Hollywood with auteurs who can both write and direct. AI is now accelerating this trend by allowing individuals to expand laterally. A director might use AI to help with scripts or acting, while a writer could use it to direct or perform. This shifts the traditional T-shaped career model. Instead of just having depth in one area and a thin layer of knowledge across others, individuals can use AI to become triple threats with deep knowledge in several areas at once.

Economist Larry Summers describes this as being non-fungible. If you are only a designer or only a coder, you are a replaceable cog in a system. However, when you combine those skills, you become a unicorn that is nearly impossible to replace. AI makes it easier than ever to bridge these gaps. Beyond just performing tasks, AI can act as a personal tutor. It can teach you the skills you lack so you can super empower yourself and perform feats of magic in your field.

There has never been a technology before where you can ask it to teach you how to do this thing. People who really want to improve themselves and develop their career should be spending every spare hour talking to an AI and saying, train me up.

AI functions as an exceptionally patient tutor for those looking to switch roles or gain new skills. A coder can ask an AI to train them to be a product manager, and it will generate problems and assignments to evaluate their progress. One effective way to learn is by watching the output and logic of an AI agent as it works. Observing its decision-making process helps users understand complex concepts like software architecture even if they do not have an engineering background.

It is almost become this layer on top of learning to code is learning to see what the agent is doing and thinking, because that teaches you about architecture.

When errors occur, asking the AI what could have been done differently provides a path to improvement. This interaction mirrors working with human colleagues. Effective feedback requires understanding the logic behind an action. Marc explains that just like managing people, you need a theory of mind for the AI to understand why it went sideways. AI is unique because it will happily explain its reasoning or even critique its own work without fatigue.

I need to actually understand what you were thinking in order to really give you the right feedback. And so, and again, the great thing with AI is AI will happily sit there and explain all day long why it's doing what it's doing.

Advanced techniques include playing multiple AI models against each other. One AI can write code while another debugs it, creating a council of agents that debate and refine results. This shift in capability opens doors for those with latent creative interests. Marc shares that while he lacks traditional art skills, he has always wanted to be a cartoonist. These tools might finally allow him to pursue that goal.

Marc identifies three distinct layers where AI is currently transforming how startups are built. The first layer focuses on whether AI simply adds new features to existing products or if it redefines entire categories. In some cases, a new technology is just an additive ingredient, like flash storage replacing hard disks. In other cases, the shift is total, similar to how web software replaced old school on-premise tools. Marc questions if traditional tools like Photoshop will just add AI features or if generative AI will make manual editing obsolete because it is easier to just create a new image from scratch.

Sometimes you get the additive to an existing thing. Sometimes you get the actually it redefines an entire product category, redefines an industry. The actual, in many cases the companies themselves turn over.

The second layer involves how AI changes individual jobs. Founders are looking at how to empower their teams so that a small group can do the work of a much larger one. If a company has a budget for 100 coders, the question is whether they use AI to make those 100 people ten times more productive or if they shrink the team to ten super-empowered individuals who do the same amount of work. The leading founders are already experimenting with these ratios to maximize output.

The third and most radical layer is the potential to redefine the very idea of a company. This involves the pursuit of the one-person billion-dollar outcome. While we have seen small teams achieve massive success with companies like Instagram or WhatsApp, AI might allow a single founder to oversee an army of autonomous bots. Marc suggests we might even see businesses that exist entirely as AI bots on a blockchain, operating independently and issuing dividends to their owners.

Can you have entire companies where you have basically the founder does everything because what the founder is doing is like overseeing an army of AI bots? There is this holy grail in our industry that has been running for a long time, which is like, can you have like the one person, billion dollar outcome?

The idea of a one person billion dollar company is a popular topic of discussion among tech leaders. However, practical challenges like support tickets, bugs, and administrative forms often get in the way. Even if AI handles much of the workload, edge cases still require human attention. This makes it hard to see how a single individual can manage everything alone.

There are so many little annoying things that I have to deal with with just support tickets and issues and bugs. It is hard for me to imagine actually a one person billion dollar company.

Defining what counts as a one person company is also tricky. Marc points out that Satoshi Nakamoto created Bitcoin alone, but the project grew through an open source community. It is unclear if a founder who relies on contractors or a community truly fits the definition of a one person business.

Large technological shifts take a long time to unfold. People often rush to make definitive claims about which companies will win or where the moats exist. This happened with the early internet. Most predictions made in the first fifteen years were completely wrong. Media outlets prefer guests who provide certain answers, but the reality is much more complex.

The really big technological transformations take a long time to play out and there's all of these structural implications that just cascade out over time. And then there's this rush to judgment up front where people say it is therefore obvious that xyz.

The question of whether AI models have moats is still open. At first, it seemed like high costs and talent requirements would create a monopoly or oligopoly. Instead, models commoditized rapidly. Within a short time, multiple companies in the US and China released comparable products. Open source options also became available and can run on less hardware.

The debate continues regarding the value of applications. Some believe the model will eventually do everything. Others argue that fitting a model to a specific industry, like law or medicine, is where the real value lies. Marc views this as a complex adaptive system. Many factors like regulation, capital, and individual choices will determine the outcome.

The technology itself provides one of the inputs. The legal and regulatory process is another input. Actual individual choices made by entrepreneurs matter a lot. The economics matter a lot. Availability of investor capital varies over time. That matters a lot.

The term GPT wrapper used to be an insult. Now, some of the fastest growing companies are effectively wrappers or agents built on top of existing models. For example, Claude Code is an app built to harness the Claude model specifically for coding. It is better to remain open to surprises rather than assuming the current leaders have an unbeatable lead.

A new coding tool was built in just a week and a half using AI. While the functionality is impressive, it raises a serious question about defensibility. If a magical product can be created in such a short time, the barrier to entry might be lower than previously thought. Major model companies are likely watching these developments and preparing to build their own versions. History suggests that fundamental breakthroughs in this field are often replicated or surpassed within months.

The other way to look at it is how much of a barrier to entry can there be in something that was developed in a week and a half? People are all over the world every day now saying they cannot believe what they can do with this. It feels like the most magical product ever, but at the same time, it took a week and a half.

There is a theory that major AI labs do not have many secrets. They seem to share similar knowledge and frequently overtake each other. The emergence of DeepSeek shows that a relatively unknown group can reimplement major ideas and even add original concepts. Despite this, labs continue to pay engineers like rock stars. These creative individuals may be working on nascent ideas that will eventually become hard to replicate.

Marc believes that forecasting the future of this industry is currently impossible. Trying to predict which specific company or killer app will win in five years is likely a waste of time. Instead, the focus should be on staying adaptable in a rapidly changing environment.

I need to put a big discount on my forecasting ability on this one. For me, it is much less interesting to try to say industry structure in five years is going to be X. I think a much better use of my time is being very flexible and adaptable at a time like this.

Peter Thiel uses a framework to categorize how people view the future. He divides them into optimists or pessimists and determinate or indeterminate. A determinate optimist believes the world will improve because they are going to do a specific thing. Elon Musk is a prime example because he has concrete goals for electric cars and Mars. Thiel historically argued that Silicon Valley has too much indeterminate optimism. This is the belief that the future will be better without knowing exactly how it happens.

The indeterminate optimist thinks the world is going to be better but can't explain why. Some combination of things is going to happen to make the world better, even if we don't know what those things are.

Marc identifies as an indeterminate optimist. He argues that this is the best strategy for a venture capital firm. While every founder needs to be a determinate optimist with a specific plan, the ecosystem as a whole should not try to pick just one path. The strength of the American economy is that thousands of capable people are all running their own experiments. By supporting many different founders, the system creates the best chance for success.

The nature of the future is that we just don't know all the answers and that's okay. The right way to deal with that is to run as many experiments as possible and have as many smart people try to do as many interesting things as possible.

Founders often point out that they take more risk because they only get to make one bet. They are the ones who drive the companies and deserve the credit. History remembers individuals like Henry Ford rather than the investors who funded them. However, investors play an important role by providing the capital that keeps the cycle of innovation moving. They help many smart people try to solve problems at the same time.

Marc struggles with the standard definitions of AGI. He sees two main versions. The cosmic version is the singularity, where machines self-improve and human judgment becomes irrelevant. Marc does not believe we live in that world yet. The prosaic version defines AGI as the point when AI can perform valuable economic tasks as well as a human. Marc thinks this second definition actually understates the future because it treats human ability as a cap.

Human intelligence is limited by biology. Most high-level professionals like doctors or research scientists have an IQ around 140. The absolute peak for humans is roughly 160, which is where people like Einstein or Feynman sit. AI models are already approaching the 130 to 140 range and will soon pass the 160 mark. There is no theoretical limit to where AI intelligence can go once it is released from biological constraints.

I think we're used to living in a world where we just don't understand how good, good can get because we've been capped by our own biology. And we're going to get to experience what it's like when you have the capability at your fingertips that's actually better than human in these domains.

Marc suggests that reaching human-level capability will be a minor event. He views it as a footnote that might happen on a random Tuesday in 2026. The more important phase is the exploration of what happens when we have machines with an IQ of 200 or 300. We will soon have AI coders, doctors, and lawyers who are better than the best humans in those fields.

I think this idea of human equivalent is just going to be like a footnote. It's like, oh, yeah, that was just on Tuesday in 2026 is when they hit that. And it kind of didn't matter because the next question was like, okay, what do we get to do in a world in which we actually have machines that are better than that.

Marc experiences constant frustration with his own cognitive limits. He feels like he should be able to do more, but he often lacks the time or the mental capacity to handle complex tasks alone. Even after reading ten books on a subject, he often forgets most of the details. This gap between what he wants to achieve and what his brain can retain is a source of regular struggle.

I sort of live in this kind of state of endless frustration. If I could just be smarter than I was, I would be so much better at what I do, but I am not.

This feeling extends to his interactions with others. Marc frequently meets people who are clearly more intelligent than he is. During these conversations, he finds himself taking notes just to keep up, eventually realizing that they are simply outthinking him. These human limitations make the prospect of AI very compelling. Machines do not have these same boundaries. They can handle the tasks that human brains find exhausting or impossible.

We are just so used to having those limitations that the idea of having machines that work for us that do not have those limitations is much more exciting than people are giving credit for.

Marc follows a barbell strategy for his media consumption. He focuses on two extremes while ignoring the middle. He reads X for real-time information about what is happening right now. For deeper insights, he turns to books written at least 50 years ago. These older works have stood the test of time and contain ideas that remain relevant across generations. He is much more skeptical of media that falls between these two categories.

I have a perfect barbell strategy, which is I read X and I read old books. It is basically either up to the minute what is happening right now, or it is a book that was written 50 years ago that has stood the test of time.

Standard news sources like newspapers and magazines often fail to provide lasting value. If someone reads a newspaper from just a week ago, they will likely find that many of the predictions were incorrect. These publications often lack the necessary information to make accurate forecasts. Magazines are even less reliable because their long production cycles often make content obsolete before it reaches the reader.

If you go back and you read old newspapers, you realize none of this happened. None of what they predicted played out the way that they said that it would. None of this turned out to actually be that relevant or correct.

Direct exposure to practitioners who are actually creating content is still dramatically underrated. In the past, information was filtered through mass media like television and newspapers. Now, platforms like Substack and podcasts allow smart people to explain their work directly to an audience. This provides a level of depth and expertise that was previously inaccessible to the general public.

There is just actually just tremendous amounts of alpha in listening to the world's leading experts in the space who actually just show up and talk about what they're doing.

While some critics argue that experts are merely promoting their own interests, most people genuinely enjoy explaining their work. This sharing of knowledge contributes to human progress and offers unique insights. Silicon Valley operates like a company town where the entire industry acts as one collective entity. This environment allows for a constant flow of talent and knowledge between different fields. Marc notes that AI is now the ninth major technology wave in the region's history.

The indeterminate optimism of the ecosystem meant that Silicon Valley could morph into all these categories. Nobody had to sit and plan and say, okay, in the 1990s, Silicon Valley is going to do the internet.

The ecosystem is defined by its flexibility. It has successfully pivoted from silicon chips to the cloud and now to AI without a central plan. There is also a common joke in the industry that everything being built is essentially just a wrapper over another layer of technology. Since everything eventually traces back to silicon chips, Marc and the hosts laugh about how everyone in tech is ultimately just a sand wrapper.

The movie Eddington stands out as a significant piece of modern cinema. Set in a tiny New Mexico town of 600 people, it pits a conservative sheriff against a progressive mayor during the chaotic events of 2020. The story unfolds as the town deals with the onset of the pandemic and the social unrest that followed. A tech company resembling Meta also looms in the background as it builds a massive AI data center on the outskirts of town.

Marc describes why the film is so effective at capturing the current American experience. It shows how people in small, isolated communities were fully consumed by global events through their screens. While movies often struggle to integrate smartphones and social media naturally, this film succeeds by making these digital experiences central to the plot.

The reason I love the movie so much is it is the first movie that directly grapples with what happened in 2020. It does a really good job of showing what it was like to live in a world in which there were things happening in the real world and people were experiencing events online in a way that was very central in their lives.

The director engages with controversial topics and social dynamics that many other talented filmmakers avoid. Marc notes that while he might disagree with the director on many points, he appreciates the effort to honestly portray what it is like to live as a human being in modern America. The film identifies cultural flashpoints and addresses them directly.

This guy for some reason is just like, I'm just going to find all the third rails and I'm just going to grab them.

Marc observes how children often find technology more appealing when they discover it independently rather than through their parents. His ten-year-old son recently became obsessed with Replit and vibe coding after discovering it on his own. This form of coding allows users to create games and applications through natural interaction, and it has captured his interest for hours at a time.

The current state of AI voice technology is particularly impressive. From emotional voice experiences to humorous AI personas, the capabilities of these models are expanding. There is also a growing revolution in wearables and voice input devices, such as smart glasses and pendants, that change how people interact with computers.

I have this app on my phone now called Whisper Flow, which is voice transcription, which works staggeringly well. It kind of understands when you are telling it, no, I want bullet points over there and I want this and that. It understands that you are not telling it to type in the words I want bullet points. It just actually understands that you want bullet points.

This level of intelligence in transcription apps marks a significant shift. The software no longer just records words but understands the intent behind them. This allows for a more fluid way of creating documents and notes where the user can direct the AI while speaking naturally.

Marc describes how his son uses Replit to build Star Trek simulators. While the original series used random knobs and lights on set, Star Trek: The Next Generation introduced a formal UI design language. Marc explains the LCARS design system used in the show as a specific aesthetic that can now be recreated through modern tools.

One of the fun things you can do vibe coding is you can say, give me a Star Trek Next Generation user interface for whatever this, that or whatever. It actually uses the LCARS design language. It will actually build you Star Trek Next Generation Bridge consoles using that design language, but with your choice of a Star Trek game.

This approach allows users to build functional software by describing the specific aesthetic and functionality they want. Beyond personal projects, Marc points to a recent article by Packy McCormick as a comprehensive explanation of how his firm thinks about the future. He also suggests following their team's YouTube channel for more updates on these emerging technologies.