Why companies are becoming a series of loops | Anish Acharya (a16z)

The End of the Linear Dream

For decades, we have been obsessed with the funnel. You pour users in the top, you squeeze them through a few filters, and a small percentage of them drop out the bottom as revenue. It is a linear, industrial way of looking at growth. But if you listen to the recent shifts in how firms like a16z are looking at the market, that model is dying. We are entering the era of the loop.

Anish Acharya recently laid out a thesis that challenges the typical doom-and-gloom narrative surrounding AI and labor. Instead of seeing AI as a tool that creates a permanent underclass of workers, he sees it as the engine that turns every business function into a self-reinforcing cycle. As a founder, you need to understand that your product is no longer a destination; it is a recurring process.

Why Funnels Fail in the AI Era

The problem with a funnel is that it is inefficient. Every time you want to grow, you have to find more top-of-funnel leads. It requires constant external pressure. A loop, however, uses the output of one cycle to fuel the input of the next. In a world where AI can automate the heavy lifting of content creation, data analysis, and customer interaction, the cost of completing a cycle has dropped to near zero.

When the cost of the process drops, the value shifts to the structure of the loop itself. If your company is still thinking about how to get a user from point A to point B, you are building for a world that no longer exists. You should be thinking about how the user's action at point B naturally brings them back to point A, while making the system smarter in the process.

The Feedback Loop as the New Product

In the crypto and AI space, we talk a lot about flywheels, but we rarely build them correctly. A true loop is not just a marketing trick; it is a fundamental shift in job functions. Take engineering or design. Traditionally, these were linear paths: you get a requirement, you build the feature, you ship it. Done.

In the new model, shipping the feature is just the start of the data collection phase. The AI monitors how the feature is used, identifies friction points, suggests optimizations, and in some cases, helps write the code for the next iteration. The human is no longer the laborer moving the brick from one pile to another. The human is the architect ensuring the loop stays closed and aligned with the business goals.

Addressing the Underclass Fear

There is a lot of noise right now about AI replacing junior workers and creating a gap where no one can learn the ropes. Acharya argues this is the wrong way to look at it. Instead of looking at it as a loss of entry-level jobs, we should see it as a shift in what entry-level work looks like. If the AI handles the repetitive, low-level execution, the junior's job becomes managing the loop.

The fear isn't that we will have fewer workers; it is that we will have a shortage of people who know how to manage automated systems.

For founders, this means your hiring strategy has to change. You aren't looking for people who can perform a task. You are looking for people who can design a process that performs the task repeatedly and improves with every iteration. This is the difference between a builder and a manager of builders.

The Consumer AI Pivot

We are seeing this play out in consumer AI right now. The first wave of AI apps were just wrappers—simple interfaces that gave you access to a model. They were linear tools. You ask a question, you get an answer. The next generation of successful consumer products will be loops. They will learn from your preferences, anticipate your needs, and proactively trigger actions that bring you back into the app.

If you are building a consumer app today, and you aren't thinking about how your AI agents create a recurring value cycle without manual user intervention, you are going to get disrupted by someone who is. The goal is to create a product that gets better the more it is ignored, because the background loops are doing the work for the user.

What This Means for Builders

• Stop measuring throughput, start measuring retention velocity. It does not matter how fast you can build if what you build does not feed the next cycle.
• Focus on the seams. The most important part of a loop is where the output meets the input. That is usually where the data gets lost. Fix the seams to ensure the cycle never breaks.
• Automate the execution, not the decision. Use AI to handle the volume, but keep the human at the center of the loop's direction.

The transition from companies that provide services to companies that operate loops is the defining trend of this decade. It is a more sustainable, more scalable way to build. It requires a skeptical eye toward traditional growth hacks and a founder's focus on long-term systemic health. The funnel is a bucket with a hole in it. The loop is a closed system. Choose wisely which one you want to spend your venture capital building.

The Takeaway

The future belongs to the architects of cycles, not the managers of funnels. If you want to survive the AI transition, you need to stop looking for the next customer and start looking for the next way to make your existing product work for itself. The permanent underclass isn't coming for the workers; it is coming for the businesses that refuse to evolve past linear thinking.

Read the original at Lenny's Newsletter →