In the heated summer of 2024, while the crypto market nursed its bear wounds, two names from the AI frontier quietly dominated Chinese tech media: Liang Wenfeng of DeepSeek, and Yang Zhilin of Moonshot AI. The headline was simple, almost cruel: “Liang Wenfeng has no life. Yang Zhilin has no exit.” No market panic, no liquidation cascade – yet the story carried a weight reminiscent of the darkest days of the ICO bust. For those of us who have watched founders burn out under the weight of impossible promises, the parallel was unmistakable. We built crypto on the illusion that code replaces trust. Now AI is reminding us that trust – and the human cost behind it – remains the only real architecture.
Context: The Two Faces of a Single Fever Liang Wenfeng, a quant-turned-AI researcher, built DeepSeek as an open-source, low-cost alternative to GPT-4. His team works in near-silence, shunning media, pouring 16-hour days into model optimization. “No life” is not hyperbole – it is a badge of extreme focus. Yang Zhilin, by contrast, leads Moonshot AI, the creator of Kimi, a chatbot famous for handling 200,000-character contexts. His company raised hundreds of millions from Alibaba and others, but market pressure leaves him in a zero-sum game: win the long-context race or face existential obsolescence. “No exit” is the reality of a founder whose every move is watched by investors expecting a unicorn exit. Both stories, though rooted in AI, are textbook examples of a phenomenon crypto knows intimately: the cult of founder sacrifice. Verify the code, trust the community – but what code protects the founder?
Core: The Fragile Architecture of All-In Founders When I audit a Layer2, I look for liquidity fragmentation, not community hype. When I evaluate a DAO, I check whether “code is law” is actually enforced by a handful of multisig keys. Founders are no different – their personal bandwidth, mental health, and strategic optionality are the true ‘multisig’ of a startup’s survival. Liang’s “no life” model echoes the relentless grind of early crypto builders who coded through the night during the 2017 ICO bubble. I remember auditing 150 whitepapers only to find most had no sustainable incentive design; they ran on founder charisma until charisma ran out. Yang’s “no exit” scenario mirrors the predicament of DeFi protocols that bet everything on one yield curve – if the market turns, there is no fallback. In crypto, we learned the hard way that over-indexing on a single founder’s stamina is a systemic risk. Bulls react to gains. Bears reflect on fragility. We build systems that outlast any individual. Tech changes. Values remain.
My own journey through the 2022 bear market taught me that solitude and reflection are not weaknesses – they are the foundation of resilience. Retreating to a Virginia cabin for two months, disconnected from Twitter, I re-read Hayek and Turing. I realized that the industry’s growth had outpaced its ethical infrastructure. The same is true for AI founders today. Liang and Yang are not just building models; they are being forced into a hyper-competitive arena where the only accepted narrative is total sacrifice. This narrative, if unchecked, will produce not breakthroughs but burnouts. The crypto community must see this as a warning. We have already lost talented developers to the cult of “number go up.” We cannot afford to lose a generation of AI leaders to the cult of “no life, no exit.”
Contrarian: The Hidden Cost of ‘No Exit’ But is this narrative entirely true? Or is it a convenient story spun by investors to extract maximum output from founders? Consider the data: DeepSeek’s API pricing is 1/100th of GPT-4, a pricing strategy that sacrifices short-term revenue for market share – a luxury funded by Liang’s quant background. Moonshot AI’s long-context lead is being eroded by competitors within months. In both cases, the “no life” and “no exit” framing might be a PR tool to justify aggressive tactics and mask underlying strategic fragilities. In crypto, we have seen similar narratives used to justify locking liquidity for years in illiquid tokens. The contrarian view is that the most resilient systems are built by founders who maintain balance, who build teams, who create governance that survives their own absence. A protocol that depends on a single founder’s 80-hour work week is not decentralized – it is a dictatorship of insomnia. We must question whether we are romanticizing exhaustion instead of engineering sustainability.
Imagine a crypto-parallel: Satoshi Nakamoto disappeared, but Bitcoin survived because its code and community were designed for independence. If Liang or Yang were to step back tomorrow, would their companies survive? If the answer is no, then the architecture is flawed. The real innovation is not in building a better model, but in building a system that does not require a founder to have “no life.” That is the covenant we need – a covenant that protects the builder, not just the build.
Takeaway: Build the Exit, Guard the Soul So what is the takeaway for both the crypto and AI communities? It is this: celebrate the builder, but do not worship the sacrifice. The most durable protocols are those with a clear path to decentralization – where the founder’s role becomes optional, not existential. For AI founders, this means investing in team depth, open standards, and governance that allows for rest, reflection, and even failure. For the crypto community that hosts them, it means applying our own principles: verify the code, trust the community – but never trust a system that requires a human to have no life and no exit. As we move toward a future where AI and crypto converge, we must ensure that the infrastructure we build is resilient enough to withstand the absence of any single individual. Because tech changes, but values remain. And the value of a human life is not a resource to be mined – it is the bedrock upon which any covenant must rest.