The myth of "ETH killers" — why demand for blockchains will always outpace supply 🧵

For every important computing resource in history demand has outpaced supply. This includes CPUs, GPUs, memory, storage, and both wired and wireless bandwidth.

The core dynamic of computing movements is a mutually reinforcing feedback loop between applications and infrastructure.

Consider a modern smartphone. The phone itself is much better than a decade ago, and so are the apps. The apps drove the popularity of the phones which gave manufacturers more money to reinvest. Improved phones increased the app design space, and the cycle repeated.

In the 2000s, when broadband penetration reached minimally viable levels, entrepreneurs created video-sharing sites like YouTube, which spurred broadband demand even more, making video steaming work even better, in a reinforcing loop.

As GPUs got better, game developers invested in higher fidelity graphics and experiences, which grew the video games market, giving GPU makers more money to reinvest in chips, and so on.

This pattern happens over and over, and is one of the main reasons computers have become ubiquitous and so important in our lives.

Now let's talk about blockchains. Programmable blockchains are computers: you can write (near) Turing-complete code for them, store information in them, and so on.

It's very likely that similar to other important computing resources of the past, demand for blockchains will consistently outpace supply.

This is already true for Ethereum itself, and will likely also be the case for L2s, side chains, and other L1s.

Today, Ethereum has roughly 10M monthly users, which is roughly 0.2% of all internet users. Over the 2020s, usage will hopefully grow significantly both in terms of number of users and in terms of the number and frequency of transactions per user.

Many of the apps that will drive this growth probably haven’t even been launched yet. It’s possible they haven’t even been imagined yet.

The more blockchain infrastructure projects, the better. For one thing, it's important to fully explore the design space, including different approaches to scaling, security, governance, developer and user experience, etc.

Some of the designs will fail. Others will make fundamental trade offs.

For example it’s plausible to imagine a world where frequent-use, low-value transactions happen on blockchains that trade security for performance.

Assets that increase in value can then be moved to other blockchains that prioritize security over performance.

It's also important to have many blockchain infrastructure projects because when exponential growth kicks in, we will likely have a consistent shortage of high-quality infrastructure.

Computing movements are positive-sum games. There can be many winners at both the infrastructure and application layer.

It's important to run a variety of experiments — and stay focused on the bigger goal of bringing blockchains and web3 mainstream.

===

Originally published on November 21, 2021

The myth of "ETH killers" — why demand for blockchains will always outpace supply 🧵

For every important computing resource in history demand has outpaced supply. This includes CPUs, GPUs, memory, storage, and both wired and wireless bandwidth.

The core dynamic of computing movements is a mutually reinforcing feedback loop between applications and infrastructure.

Consider a modern smartphone. The phone itself is much better than a decade ago, and so are the apps. The apps drove the popularity of the phones which gave manufacturers more money to reinvest. Improved phones increased the app design space, and the cycle repeated.

In the 2000s, when broadband penetration reached minimally viable levels, entrepreneurs created video-sharing sites like YouTube, which spurred broadband demand even more, making video steaming work even better, in a reinforcing loop.

As GPUs got better, game developers invested in higher fidelity graphics and experiences, which grew the video games market, giving GPU makers more money to reinvest in chips, and so on.

This pattern happens over and over, and is one of the main reasons computers have become ubiquitous and so important in our lives.

Now let's talk about blockchains. Programmable blockchains are computers: you can write (near) Turing-complete code for them, store information in them, and so on.

It's very likely that similar to other important computing resources of the past, demand for blockchains will consistently outpace supply.

This is already true for Ethereum itself, and will likely also be the case for L2s, side chains, and other L1s.

Today, Ethereum has roughly 10M monthly users, which is roughly 0.2% of all internet users. Over the 2020s, usage will hopefully grow significantly both in terms of number of users and in terms of the number and frequency of transactions per user.

Many of the apps that will drive this growth probably haven’t even been launched yet. It’s possible they haven’t even been imagined yet.

The more blockchain infrastructure projects, the better. For one thing, it's important to fully explore the design space, including different approaches to scaling, security, governance, developer and user experience, etc.

Some of the designs will fail. Others will make fundamental trade offs.

For example it’s plausible to imagine a world where frequent-use, low-value transactions happen on blockchains that trade security for performance.

Assets that increase in value can then be moved to other blockchains that prioritize security over performance.

It's also important to have many blockchain infrastructure projects because when exponential growth kicks in, we will likely have a consistent shortage of high-quality infrastructure.

Computing movements are positive-sum games. There can be many winners at both the infrastructure and application layer.

It's important to run a variety of experiments — and stay focused on the bigger goal of bringing blockchains and web3 mainstream.

===

Originally published on November 21, 2021