a16z: The history of YouTube is the future of the internet

Source: Anish@a16z

Fifteen years ago, if you asked “What do smart people do on weekends?” a good answer would be “Make YouTube videos.”

YouTube celebrated its twentieth anniversary earlier this year. While we all love reminiscing about those early videos that made us laugh, we don't always realize how countercultural it was to be a YouTuber in the early days. Even after the YouTube Partner Program launched in 2007, the idea of earning money from a channel seemed distant. With so much video content already available, from major studios to long-tail content in the 110th channel of cable TV packages: how could there possibly be more commercial demand?

Looking back now, we realize that the world 15 years ago was actually a world of short videos. The reason we know this is due to the development on YouTube since then. When anyone with a camera and editing tools can find their audience, we discover a variety of channels and businesses— from Hot Ones to Mr Beast to Dwarkesh— all of which clearly deserve a prominent place in our content world. The so-called “long tail” is much larger than anyone ever imagined.

Perhaps this is the correct historical guide to thinking about large language models (LLM), web applications, and the future of the internet. YouTube has transformed content by simplifying “creatively producing content” and “running a small business” into a series of straightforward steps. You still need creativity and drive, but the rest has become much simpler. So, why shouldn't we expect the same to happen with software? The internet has always excelled at facilitating creation without permission for anyone. But it wasn't until the emergence of the large model (LLM) that the definition of “anyone” shifted from “developers” to “anyone with an idea who can access code agents.” Five years from now, we may look back and realize that this world is lacking in software (short software), because the only people capable of building software were engineers. In other words, for other areas of the internet, this is a YouTube moment:

• The argument that “the world lacks software (short software)” is equivalent to the “the world lacks content” from 2006. Back in 2006, you would point to 100 cable TV channels and say “that's enough.” Today's software and websites face the same dilemma.
• Large language models (LLM) ultimately make it possible for niche software and applications that could not be brought to market before development. You wouldn't hire a team of engineers to develop a product for 100 people, but you can use application generation and programming tools to develop (and profit from!) smaller-scale products.
• Long-tail web applications will be built by specific types of “professionals,” and Youtubers are the best template of such professionals we have.

The content is now the application; the application is now the content

Paul Bakaus ( pointed out that when people talk about the internet, they are actually talking about three different things: the content layer of the web, which includes blogs, YouTube, Substack, and traditional publishers; the commercial layer of the web, which includes marketplaces like Amazon and Shopify stores; and the application layer of the web, which for much of internet history has consisted of “serious” cloud-based software, such as enterprise platforms and social networks.

Large language models )LLM( influence all these aspects in different ways. The commercial aspect is the subject of another article, but we do see that large language models play an important role in product recommendations and purchases; this is reflected not only in new AI discovery mechanisms such as search, but more importantly in the recommendation engines themselves. However, aside from business, our understanding of “content” on the internet and how we are rewarded for it is undergoing a significant transformation.

Content has always been the “traditional long tail” of internet participation, which dates back 30 years to the birth of the first web browsers. Every decade, there has been a significant story about the long tail of content being controlled by some centralized power: initially AOL, and later Facebook.

This time, the biggest highlight is that LLM scraping websites are a new and even worse form of content capture—because they have effectively turned into an application that allows you to consume information that should have previously brought traffic to content creators. Publishers are desperate about a phenomenon called “Google Zero” (the day when organic search traffic approaches zero), while some interesting new forces are fighting back against this phenomenon, such as Cloudflare's entry into pay-per-scrape and new micropayment standards like x402.

Meanwhile, interesting things are happening on the other side of the internet: applications are becoming the new content.

In the long run, the internet will become more participatory.

There are some categories of software that have never been developed, and the reasons are simple: insufficient return on investment, high costs, or because the preferences of about 20 million developers dictate the software we all use. But soon we will discover what they really are.

![uMMQLiScSrktcM6qn929O9PH65jLElVRZHJlFXwM.png])https://img-cdn.gateio.im/webp-social/moments-53abc15934cc33b27d7d01661d9dfc42.webp “7409885”(

Source: SK Ventures

Now, with new application generation tools like Replit, v0, Loveable, Figma Make, Bolt, and Base44, it's easy to create prototypes, build, and launch new applications. Previously, this required engineering expertise and might even have needed an entire development team. Now, you only need $200 a month (or even less!) and a good idea.

There is a famous joke that most Generation Z aspire to become professional YouTubers or TikTokers when they grow up. People love to argue and blame whether this is a realistic goal, yet they overlook a more fundamental observation: children crave to start businesses at the forefront of the internet and achieve the American Dream. And so far, the main channels to realize this wish are YouTube and TikTok.

LLM provides a brand new perspective for creators on the internet. If you are passionate and eager to build your app ideas and turn them into reality, you can now make it happen. In the previous era of the internet, you had to invest huge upfront costs, which were very high for realizing an idea, and you needed to continually increase your customer base to prove your existence. With LLM, you can easily launch products and quickly gain paying users.

For example: My wife has recently become obsessed with manifestation (note: a practice based on mindfulness and a positive mindset). She is very good at it (to the point that I am now extra careful not to upset her). She is currently developing an online application for her friends to learn about manifestation. A few years ago, she would have promoted this service on Facebook, working hard to make the content engaging and lively for the audience. Now, she can develop an app that directly connects with customers.

I am sure this is not just happening in my home: I have seen a variety of applications, from low-fidelity traffic control streaming websites in Brazil to an AI band improvisation platform built with Replit. The powerful capabilities of Replit are evident: earlier this year, Replit's ARR reached 150 million dollars, and the demand for its coding agent product is also skyrocketing.

We also hope to see this phenomenon extend beyond the network. Companies like Wabi are making it easier to build brand new personal mobile applications that cover features such as weightlifting, collage generation, fasting, and reminders to touch the grass.

Creating Applications: A New Type of Entertainment Industry

A few years ago, Nadia Asparouhova collaborated with Stripe Press to publish an excellent book titled “Working in Public”. This book tells an interesting story about how open source software development work has changed in the era of Github: maintainers of popular projects have to spend a lot of time managing those fans who want to participate and contribute. Public creation has started to become a kind of “performance”. It's not enough to just do good deeds—you have to think like a streamer.

Today, with the long-tail effect of vibe coding applications and software, we may see similar things happening on a larger scale. The challenges may not be the same, but the required skill sets may be similar: successful individuals will be some type of professional. They are completely different from the previous “professional” software developers, as they have a different intuition about the medium and how to build things that resonate.

It's worth humbly remembering that very few veterans of the traditional television industry have become major stars on YouTube. For a long time, mainstream television programs (like late-night comedy shows) have viewed YouTube as a dumping ground for “extra content,” such as bloopers or deleted scenes. Traditional software developers and investors could very well misunderstand vibe coding applications in the same way—not completely dismissing them, but overlooking their purpose.

Just like what we see on YouTube, the most successful new applications are likely driven by individuals and personality: these people have existing distribution channels, have created natural Shearing points for their communities, and they make “unboxing videos, but for software.” But platforms like Wabi also point to another possible outcome. We might see ultra-personalized applications targeting smaller niche audiences online. This greatly liberates people: software no longer needs to be practical. It doesn't require billions of dollars in revenue to justify its existence. It just needs a good idea behind it, and some people with the professional intuition to understand its value and appreciate the success it requires (just like professional YouTube creators).