Wang Feng's New Year's Eve Dialogue with Tim Gong: On Information Sorting, Entropy, and the Tomorrow of Web3

Source: MarsBit

Element Founder Wang Feng

SIG China Partner Tim Gong

I have known Tim Gong for quite a while now. I have always thought of him as a non-mainstream figure among Chinese VCs.

I first heard about him in 2007 when I started my first venture. At that time, SIG China and his name did not have the same influence in the industry, far from being as famous as Sequoia and IDG. However, his way of speaking left a deep impression on me. We haven't had many conversations, whether in SIG's office or at gatherings in this circle, but when Tim talks to me, there are rarely serious discussions; it's mostly jokes and banter.

People influence each other; if you meet someone serious, it's hard to joke around with them. However, when you meet someone like Tim, who seems to always be lighthearted, it's also difficult to have a serious conversation.

One summer in 2021, after SIG and Sequoia China co-led the investment in Element, Tim invited me to have drinks with him and some VC friends, waiting for me until midnight. By the time I found him, the others at the table had already left. Only Tim was left waiting for me, and when I arrived, he was still laughing and chatting with me, never missing a chance to tease.

That day, I heard him share some insightful comments about investing that I've also seen online recently: "Being a person is more important than doing things, emotional intelligence is more important than IQ, experience is more important than knowledge, being emotional is more important than being rational, courage is more important than brains, and luck is more important than skills," as well as "Do business with time, and make friends with trends." It's hard to connect these words with a Princeton PhD student in engineering.

In Tim's WeChat Moments, you wouldn't guess he is one of the top VC partners in China, as he rarely talks about investments or technology. He is a classical music, wine, and tea enthusiast.

This time, however, might be the most serious conversation between Tim and me, and it could even be a discussion that some might find obscure. In Tim's self-deprecating words: it's rare to use unimportant IQ, knowledge, and brains in investing to talk about Web3. Undeniably, Tim has always had his own original insights on Web3, built on foundational logic.

Our conversation took place on WeChat, and I'm sharing it here.

Wang Feng: As a successful VC investor in the Web2 era, having a conversation about Web3 with me just before the Chinese New Year is quite timely. Today, the most talked-about topics around us are either AI or Web3, regardless of whether ChatGPT is overrated or how much of the Web3 concept is marketing. A unique atmosphere in the field of internet technology seems to be emerging, hinting that the next information age is about to arrive. From an investor's perspective, can the successful models of leading internet companies over the past twenty years be abstracted and summarized? Do these summaries still inspire future development?

Tim Gong: Yes, I also take this opportunity to wish everyone a Happy New Year in advance. In fact, from the very beginning of the internet, starting with Yahoo, the underlying logic of the most valuable business models online has always been information sorting. In the Yahoo era, "experts" helped everyone sort through the same information.

Google sorted information based on search algorithms and query keywords. Facebook, Twitter, and WeChat sorted information based on each person's social relationships. By the time we reached ByteDance, information was sorted and recommended to you based on your interests and history through algorithms. This is what I often say: ByteDance achieved a leap from "people finding information" to "information finding people," where "content finds people," "products find people," and "services find people" all involve sorting and distributing massive amounts of digital information based on certain logics (like interests). All of this relies on exponential growth in computing power, communication speed, and the full digitization of information (basic sorting).

In fact, the next generation of internet applications we see today, such as ChatGPT, Web3, or Web5, are also innovations in information sorting.

Wang Feng: The first principle, frequently quoted by Elon Musk, has become popular among tech entrepreneurs and investors in recent years. However, your perspective on analyzing the commercial value of the internet through information sorting is indeed interesting. This is something worth deep analysis in our industry, allowing us to reflect on ourselves. From this perspective, the strength of an internet company's ability to handle information sorting and its innovation capability should become an important indicator for judging the excellence of that company. Do you think this value is unique to the internet industry?

Tim Gong: Musk learned some physics along the way, and before entering the startup and venture capital industry, I also majored in physics; I pursued a PhD in Electrical Engineering (Applied Physics) at Princeton University. Currently, at ByteTrade, where I serve as Chairman, six people, including the CEO and CIO, have backgrounds in physics, so we like to use physical principles to analyze the underlying logic of the world. The process of "sorting" is essentially the process of generating "negative entropy" in physics.

The real universe, all life, from beginning to end, follows physical principles. The human spiritual world and the digital information world also quietly follow physical principles.

In physics, there is the concept of entropy, which quantifies the idea of "disorder" using statistical methods. Simply put, the more disordered and random the arrangement, the higher the entropy value; conversely, the more ordered or higher the sorting level, the lower the entropy value.

In a closed system, the system itself will evolve to reach the most disordered random state, which is the state of maximum entropy. Increasing the order of a system, or reducing entropy, requires energy consumption. From a microscopic perspective, this involves sorting some form of matter (particles, atoms, molecules, electrons, DNA, cells, proteins, etc.). For example, the reason that advanced life forms, including us, can evolve on Earth is that the sun provides us with energy and negative entropy.

Let me give a few examples that everyone can understand to illustrate the application of entropy and sorting in human society.

The first example is gold. Gold is the currency chosen by human civilization and the cornerstone of the financial world. This choice is not accidental but a result of sorting. When a supernova explodes, various simple low-sorting elements at the front of the periodic table form first, and after forming iron nuclei, gold nuclei are formed through countless collisions with high-energy neutrons.

Since uncombined neutrons do not exist on Earth, gold can only be produced in the laboratory using enormous energy to reconstruct high-energy neutron beams. Gold ranks lower in the periodic table, meaning it is a higher-level element that requires immense energy to form. Its reactivity is very low, existing almost exclusively as a monatomic substance, does not oxidize, has high malleability, a mid-to-high melting point, and high density, making it the ideal currency. For thousands of years, humans have expended enormous energy to mine gold from various places on Earth, gradually forming the gold bars and nuggets stored in national treasuries today. This is also an energy-consuming sorting process, but this sorting has built a stable foundation for international finance.

The second example is diamonds. Over a hundred years ago, the Jewish community chose diamonds as a supplementary financial currency. Diamonds represent the most stable high-level sorting among all crystal structures, formed under high pressure and temperature from the Earth's interior or from collisions with meteorites, requiring high energy to create this advanced sorting. Humans also expend enormous energy to mine, cut, and polish diamonds, performing further sorting, with the 4C standard being a form of aesthetic sorting.

The third example is oil. Oil is formed from organic matter deep within the Earth's surface, undergoing enormous heat consumption, where atoms and molecules are re-sorted into hydrocarbon clusters. Its combustibility is one of the foundations of modern industry, with the purpose of burning oil being to generate electricity. Electricity is a form of ordered arrangement of electrons, and electrical energy is the most ordered form of energy, elevating human civilization to new heights and serving as the foundational cornerstone of today's digital information age.

All of the above processes of sorting people represent a process of entropy reduction, where nature or human life expends energy to create new material sorting. Therefore, in both the material and digital worlds, the underlying logic of my investments has always been to pursue the generation (sorting) and distribution of negative entropy (ordered information).

Wang Feng: After experiencing significant development over the past 20 years, Web2 platform companies are now facing similar social issues in both China and the United States. People are increasingly questioning them, such as data monopolies, excessive concentration of power, and violations of user privacy, etc. Even tech elites are beginning to reflect that today's Web2 giants have deviated from the original technological ideals of the World Wide Web, such as decentralization of power, freedom of speech, and protection of personal data ownership. Do you agree with this viewpoint?

Tim Gong: First of all, information is a form of negative entropy; the more precise the information, the lower the entropy value. Early traditional media, goods, and services had relatively high entropy values. The continuous upgrading of information through digitization is a sorting evolution from low-level to high-level, and it is also a process of entropy reduction. The purpose of sorting information based on search keywords or personal interests is to distribute precise information to individuals and reduce entropy. The competition among internet companies is essentially a competition to reduce more entropy with a unit of energy. Thus, -dS/dU is a quantification of their speed in creating commercial value.

However, in the Web2 world, whether "people find information" or "information finds people," this sorting of information is completed by centralized platforms. Centralized platforms insert advertisements into the sorting results, which is a form of "pollution" of the sorting results, increasing the information entropy received by users. Essentially, it transfers the users' negative entropy to the platform itself. This is why users are increasingly dissatisfied with Web2 platforms.

The competition among Web2 platforms is a competition for entropy reduction efficiency. The more efficient the entropy-reducing service, the more advertising pollution it can bear. However, their business models dictate that the purpose of improving platform efficiency is to make users endure more advertisements. Users do not benefit from technological advancements and efficiency improvements.

This issue is common among centralized platforms. For example, we see that ChatGPT may refuse to answer certain questions based on the platform's will, and in the future, it is also likely to insert advertisements or commercial information into AI responses.

The entropy in information theory can be calculated using Shannon's formula, which is very similar to the entropy calculation formula in statistical mechanics. Implementing many of these ideas at the application level in these calculations is quite challenging; part of the difficulty lies in the fact that information theory involves a lot of knowledge from social and humanities sciences, with many more variable factors than in natural sciences. These are the topics we are currently researching at ByteTrade Labs.

Wang Feng: Do you believe that Web3 will solve the problems arising from the increasing centralization of Web2 platform companies? Please list a few points.

Tim Gong: The ideal of Web3 has always been "personal sovereignty." The most talked-about aspect is personal data ownership. However, from my analysis above, I believe that personal data sorting rights may be even more important. In other words, both storage and computation need to be decentralized.

This is also why we invested in ByteTrade. ByteTrade is about "personal cloud," controlled by individuals on edge servers. It can not only store personal information but also compute. For example:

1) Search engines serving individuals 2) Subscription services serving individuals 3) Recommendation engines serving individuals 4) Large language models serving individuals

Regarding the last point, it may be necessary to store personalized model prompts in personal clouds.

ByteTrade reconstructs today's cloud computing architecture. Each user's node server replaces the platform's server. Algorithm providers are no longer platforms but submit code to user nodes, run algorithms, and create value for users, who then pay.

Wang Feng: Speaking of which, can the negative entropy theory you mentioned earlier be further explained in today's Web3 ecological practice?

Tim Gong: In fact, the Web3 and Crypto industries are built on the negative entropy generated by information sorting. The core task of all blockchain nodes is to "produce blocks." What does this involve? It involves sorting the information that needs to be put on the chain!

Bitcoin is a form of digital sorting implemented through software, sorting through computational power via digital calculations, generating hashes and blocks, and "mining" Bitcoin. The computational power and energy expended in mining Bitcoin result in Bitcoin being formed through sorting, similar to how humans expend energy to mine gold and mint it into bars.

Thus, Bitcoin is digital gold, and like gold in traditional finance, it serves as the cornerstone of digital finance and the digital information world.

Ethereum represents another form of sorting; both POW and POS are sorting the results of smart contract computations by block-producing nodes. Ethereum is digital oil, akin to oil in the real world. The inefficiency of the L1 and POW era resembles the low combustion rate of unrefined crude oil. The formation of Ethereum 2.0 and L2 gives Ethereum the opportunity to transform itself from digital oil into digital electricity. Users pay network computing nodes with ETH as Gas. This is essentially exchanging ETH for sorting services. The circulation of ETH represents the flow of negative entropy.

I strongly agree with the basic logic of Crypto. Of course, the development of the Crypto industry has not been satisfactory so far. Due to the ease of speculation on Tokens and the lack of regulation in this industry, Crypto has been overly financialized before generating valuable applications. This is unhealthy and is currently the reason for the industry's major reshuffling.

Wang Feng: Based on this negative entropy theory, how do you view the future of public chain ecosystems?

Tim Gong: As we mentioned earlier, the circulation of ETH represents the flow of negative entropy, and the characteristics of the ETH chain determine that it mainly supports financial applications. Therefore, ETH quantifies the negative entropy within financial information.

However, financial information in this world is only a small part of the information world, while Web3 aims to address the decentralization of the entire information world. In the blockchain world, there has always been a vision of multi-chain/application chains; for example, well-known projects like Polkadot and Cosmos are centered around the multi-chain concept. Ethereum 2.0 is also moving in the direction of multi-chain development. In a multi-chain ecosystem, each chain has its own native Gas token, which serves to measure and exchange the negative entropy flow required by that application. Nodes providing computing and sorting services will receive tokens issued by that chain.

If in the future the digital world only has Crosschain, it would be like the human world having only oil as an energy source. The digital world needs Multichain just as humanity needs Alternative Energy.

I believe that public chains and their Utility Tokens will have a flourishing future.

Wang Feng: Yes, the future of Web3 is not just about public chain technology itself, but also about the controversial Utility Tokens. Many people still find it confusing; could you elaborate on your views?

Tim Gong: For example, our ByteTrade personal cloud nodes have a cooperative relationship with public chains. Personal nodes belong to individuals, and there is no need for consensus or even information synchronization between them. However, these personal nodes will utilize others' data and services and will use Crypto to pay for others' negative entropy. Public chain nodes may even run on cloud nodes. Specifically, Crypto provides immense value in this application scenario:

1) Crypto is the currency for machines. Payments between machines must be based on automated, programmable rules. I believe the blockchain smart contract concept of "code is law" refers to the laws and rules between machines.

2) Fiat currency is defined by nations, while multi-chain/application chain Crypto is defined by applications. The exchange rates between Cryptos express not the relative strength of nations but the relative value of application information negative entropy.

ByteTrade's personal cloud nodes will have users' Crypto wallets, holding and frequently exchanging various application chain Tokens. This is why we named the company ByteTrade!

Wang Feng: Centralized platforms have mature business models, which have been validated over many years. Therefore, Web3 practitioners view them as traditional industries. However, from a business perspective, how can decentralized networks support themselves in the long term?

Tim Gong: The best platforms like Google and ByteDance strongly associate (sort) advertisements with users' keywords and interests, allowing users to obtain precise information while also receiving targeted ads. However, advertisements are information that increases entropy, driven by the interests of advertisers. Centralized platforms insert advertisements into sorting results to generate revenue, partially sacrificing users' interests. This disrupts sorting in a non-transparent manner, leading to increased user entropy. A more transparent, just, and fair business model should be one where the network provides services to users, charging users for those services.

The development of the Crypto industry has validated that transparently charging Gas or transaction fees in asset trading is an excellent business model. In the ByteTrade ecosystem, we are already building decentralized digital asset management tools, based on peer-to-peer RFQ trading markets, automated trading bots, and more. We firmly believe that users will definitely pay for products that generate negative entropy and value.

Wang Feng: Decentralized compliance and regulation have always been a challenging topic. Today, Web2 platforms face enormous regulatory pressure; is this issue even more severe in Web3?

Tim Gong: The digital world is composed of software, and software implements contracts through programs, so:

Code = Execution of Contract = Contract.

When we say Code is Law, we are actually saying:

Code = Contract = Law.

However, in the human world, Law is to Enforce Contract to be Executed. Therefore, the role of true digital world Law has already been replaced by Code.

The current issue is that digital currencies have been assigned traditional monetary values, forming traditional asset values. Once there are unfair gains or losses, traditional financial Law will come into play to determine right from wrong. Thus, the problem with digital currencies is that they are too prematurely traditionalized and financialized, and they have advanced too quickly into traditional asset trading.

These problems arise because all digital currencies and trading networks are software. In fact, apart from Bitcoin (which has no Contract, only addition and subtraction) and Ethereum (which has a relatively perfect community-maintained Smart Contract execution software), the current software foundations are not capable of or suitable for such transactions colored by traditional assets. To put it more simply: 99% of past digital currency projects resemble alchemy and turning water into oil, and assets resembling fake gold and fake oil have begun to be traded on a large scale without proving their value.

Decentralized Web3 shifts the compliance responsibility from platforms to individuals. Since individuals have data sovereignty and sorting sovereignty, they naturally must also bear legal responsibilities. As the saying goes, "With Great Power Comes Great Responsibility." Companies like ByteTrade will provide compliance tools, similar to how TurboTax helps Americans file taxes legally.

For example, ByteTrade will support voluntary KYC for nodes, and after KYC, American nodes can participate in DAOs or DEXs organized according to U.S. law, etc.

Wang Feng: I heard that you have been supporting and funding research in life sciences, especially regarding cell aging, death, and cancer. Can your investment methodology in the internet field inform investment decisions in the medical field?

Tim Gong: The emergence and extinction of life is indeed a sorting process. High-level sorting of DNA, cells, and proteins has produced humanity. The growth period that a person goes through from birth is also a process of acquiring energy through food and breathing, sorting within the body according to its own rules, which is an entropy reduction process. This sorting begins to disintegrate as human organs, blood, and brains start to age, and the final death is a state of maximum entropy, the most disordered state. Last week, a Harvard lab proved after 13 years of effort that sorting (Sequence) has a mechanism that can be accelerated or reversed in mice.

My alma mater, Princeton University, has a bioengineering professor, Cliff Brangwynne, who won last year's Breakthrough Prize for Life Sciences for his research on "phase transitions in biological cells." This is one of the most influential awards in biomedical science, co-founded by founders of companies like Google, Facebook, and Alibaba.

Dr. Brangwynne's work uses physical chemistry to study living cells. He revealed that physical entropy can also be well applied at the level of life sciences. The process of cellular aging is an entropy-increasing process. These physical and chemical principles have significant guiding roles in treating cell aging diseases like ALS, cancer, and Alzheimer's.

Of course, Web3 also has a direct impact on human immortality—could it be that one day, immutable and everlasting memories of individuals could exist on ByteTrade's edge nodes, accessible to friends and future generations' computing programs?

Harvard professor George Church has said that there may come a day when scientific advancements can extend human lifespan by two years for every year that passes. This essentially means that the negative entropy input into our bodies exceeds the entropy increase caused by natural aging. This singularity of human immortality may be just a few decades away!