The following content does not involve the present or future, but purely analyzes and reflects on past, already established facts.
Right after graduating, I accidentally entered a company that was incubating an alliance blockchain internally. At first, I interned in another department of UnionPay Advantage, something like the Mobile Recharge Services Department, mainly writing SQL queries to pull data. Later, they hired an outsourced employee, let me go, and passed my resume to other departments, which is how I ended up in the Tech Development Department working on blockchain.
After becoming a full-time employee, I did frontend development for a while, displaying Bitcoin market data and blockchain industry data on a big screen. Later, I shifted to regular Java backend development, building a business system. After that, I started working on application layer blockchain development using Java to interact with the underlying blockchain developed by another team. Eventually, because someone from the underlying blockchain team left, I transferred there and began learning and developing alliance blockchain technologies in Go.
In this first phase of my career, I was understandably very lost, with almost no planning for my future career path—basically doing whatever the company assigned to me.
During this time, I tried learning various programming languages, apart from Java (which I studied in university), including PHP, Python, Scala, Go, Kotlin, Rust, Haskell, Erlang, etc., but only superficially—mainly syntax and small demos. After dabbling in multiple languages, I realized programming concepts and syntax are often quite similar.
I also tried learning other fields, such as big data analysis and advanced Java topics, but I never persisted or delved deeply.
I actively switched jobs after some time, because after working on the underlying blockchain for a while, I realized that alliance blockchains were largely built on top of Tendermint consensus, with layers of miscellaneous additions like simultaneous support for MySQL, TiDB, MongoDB, LevelDB, etc., as data sources. Our company’s focus shifted to certification tests by the China Academy of Information and Communications Technology (CAICT), complying with various test standards to earn certificates.
During work, I also explored Hyperledger Fabric, developed a DID program based on Hyperledger Indy, and kept records of the PPTs I created.
After these experiences, I gradually realized that alliance chains and public blockchains are entirely different technical systems. Alliance chains aren’t real blockchains, which later led to my strong prejudice against alliance chains. Thus, I voluntarily moved to a company working on public chain projects.
Onchain is actually a very powerful company. In China, if you were involved in blockchain (at least two years ago) and hadn’t heard of Onchain, there were only two possibilities: 1. You weren’t really doing blockchain development. 2. You hadn’t truly entered the industry.
When I first joined Onchain, I still didn’t really understand blockchain, even though I had grasped the PBFT process and knew how VBFT consensus worked. Technically, I knew the “big four” components of blockchains—consensus, database, storage, and one more.
You might not know Onchain, but you can’t be unaware of NEO, ONT, and projects like WING and FLM—they all originated from Onchain. Our project was based on ONT nodes.
Once you learn about NEO, you inevitably hear about its founder Da Shu, then trace back to the 2017 ICO boom, Ethereum co-founder Vitalik’s fundraising in China, angel investor and Bitcoin billionaire Li Xiaolai, Guo Hongcai who forked Ethereum and helped Binance, and eventually get familiar with names like Zhao Changpeng, Li Lin, Sun Yuchen, Du Jun—and truly understand the crypto world.
The biggest issue in this job was that Da Shu was too wealthy, giving our project team great salaries with almost zero workload or pressure. I learned a lot during this time, including concepts like PDP Proof and PoST. Unfortunately, the project never launched, and we were laid off when Da Shu abandoned the Chinese market to move to the U.S.
My experience at UnionPay Advantage and Onchain made me mistakenly believe I had blockchain development skills, and I constantly planned my career around blockchain development.
This was a startup I left after just two weeks. Normally, such a short stint would be negligible, but the company’s mission and its appeal are worth mentioning.
A professor from the University of Finance and Economics, Zhou Pingzuo, who was also an expert at the central bank, proposed a method using Nash Equilibrium (game theory) to achieve blockchain consensus. This company aimed to engineer this theory. If it had been feasible, it would’ve been my dream job. But as it turned out, it wasn’t reliable.
You may not have heard of Fusionist, but you might know about ACE, which launched on Binance at the end of 2023, peaked at $16, and now sits around $0.6, ranked around 600th in market cap. To be clear, the game developers at Fusionist were genuinely hardworking; the coin price wasn’t their fault—blame the investors.
Ironically, I left Fusionist because I felt I was merely using blockchain, not developing it. Yet to this day, the most tangible result of my career is helping build the ACE mainnet.
I’ve always believed using blockchain is much simpler than developing it. Running an open-source node is like using an iPhone—you wouldn’t call iPhone users “tech geniuses,” right? Similarly, running an Ethereum node doesn’t make one a great engineer.
I left Fusionist deliberately, accepting the maximum loss to show my determination to correct my career path. Having both the will and ability to pivot is, I think, pretty cool.
Not being able to contribute more significantly to a major blockchain like ACE remains a regret.
After Fusionist, I firmly believed in not doing smart contract development and not doing secondary development on Ethereum. These two decisions essentially ruled out most Web3 jobs. So what was left? I already had the answer in mind, and coincidentally found a job opportunity: Cosmos development.
I wasn’t keen to actively learn Cosmos SDK, but was happy to use it if the work demanded. Using Cosmos SDK to develop a chain, however, still falls into the trap of “using blockchain instead of developing blockchain.” Cosmos SDK is fundamentally an SDK designed to lower the barrier for developers—it doesn’t require deep knowledge of consensus algorithms, transaction mechanics, or blockchain architecture, much like smart contract platforms.
The Pell Chain project wasn’t difficult. I even arrogantly wrote about how to understand the technical architecture of any blockchain project. Pell Chain inherited much of its complexity from Zeta Chain’s cross-chain logic since it was forked from Zeta Chain.
I learned a lot during this time, especially from my Leader’s “big tech” style of management, which gave me a taste of what working in a big tech company might be like. Even though I never joined a big tech company, after seeing the money floating in the crypto industry, I lost all desire for it.
I left Pell Network voluntarily too. When an environment is bad for you, having the will and ability to leave is, again, pretty cool.
Throughout my career, there were many times when I voluntarily gave up potential opportunities.
If I had seriously learned smart contract development from the start, maybe I’d now be proficient in EVM, SOL, APT, and other platforms’ smart contract languages, with several live contracts under my belt—there’s a huge demand for smart contract developers in Web3.
If I had jumped into DeFi development early and stuck with it, I might now deeply understand how DEXes work, from Uniswap to AAVE to cross-chain bridges—skills needed by both CEX and DEX projects.
If I had committed to Cosmos SDK from the beginning, I could have become a seasoned Cosmos developer by now, familiar with all its quirks and contributing code to the ecosystem.
If I had persistently studied Ethereum’s source code, I might now be capable of optimizing parts of Ethereum clients and contributing meaningfully to Layer 2 development efforts.
If I had stayed in alliance chain development, I might have ended up at a major project like Chang’an Chain or Tencent Blockchain, working quietly but happily.
If I had never encountered blockchain at all and just focused on being a backend developer in the traditional internet sector, I’d likely now be proficient with databases, caching, middleware, and wouldn’t fumble on interview questions about Go’s GMP scheduling principles.
What if…
Overall, my past career experiences have been a huge failure. I spent years trying to become a true blockchain developer but never even touched the threshold—only working on the peripheries.
And with the industry’s evolution, after 2019 the Layer 1 blockchain wars were basically over. By around 2023, with Binance, Tron, and Huobi leaving China, and Da Shu abandoning the Chinese market, there were almost no legitimate blockchain development companies left here.
The world no longer needs “blockchain developers.”
UChains is a blockchain project incubated within a traditional payment company. The project’s Team Leader left the company in 2019 and later co-founded ChainMaker (长安链) as a partner. When ChainMaker was first launched, many technical features bore strong similarities to UChains. Later on, ChainMaker developed quite successfully, collaborated with Tencent, and gained significant recognition.
The technical architecture of UChains was previously mentioned in “VRF + BFT Consensus Causing Transaction Failures“.
The core of UChains is based on Tendermint Core, with the code directly integrated into the project and heavily customized. The main improvements include:
SaveFS, also known as saveio, was a blockchain project incubated within Onchain. When Onchain was dissolved in 2022, the entire SaveFS project was shut down.
Technical aspects related to SaveFS were previously discussed in “A Year in Review“.
The SaveFS project was relatively complex, essentially integrating elements from the following projects:
Why the term “essentially integrating”? Back then, NEO became a very successful project under the banner of “China’s Ethereum,” and it is speculated that SaveFS inherited this kind of ambition, aiming to become something like “China’s Filecoin.”