We didn't need Jeff Yan to tell us the crypto industry is in a talent crisis—the GitHub commit history screams it. The Hyperliquid co-founder recently sat down for a podcast and dropped a grenade: the industry hasn't attracted top-tier entrepreneurial talent, and young engineers are gravitating toward AI because of the prestige and pressure. He's right on the symptoms, but his prescription—a vague 'on-chain financial renaissance'—is the kind of narrative that sounds good at a conference but misses the structural rot underneath.
Let's start with the data. According to the 2024 Electric Capital Developer Report, the number of monthly active developers in crypto has been declining since November 2022, with a 25% drop year-over-year. Meanwhile, the number of developers contributing to AI-related open-source projects has nearly doubled. Venture capital distribution tells the same story: in Q2 2024, AI companies raised $15.6 billion, while crypto raised just $2.3 billion. The capital follows the talent, and the talent is voting with their commits.
But here's the twist—the same data shows that the number of deployed smart contracts is actually rising, especially on L2s. So what's happening? We're seeing a bifurcation: more code being written, but less of it is novel. Most new contracts are forks of Uniswap V3 or Compound, with a new token name and a different yield curve. The real engineering talent—the people who build new primitives, who understand zero-knowledge proofs at the level of circuit optimization, who can design a decentralized order book that doesn't suck—those people are scarce. And the ones we do have are being poached by AI labs offering $500k+ packages.
I've seen this movie before. In 2017, I was a junior analyst in Tokyo, cranking out ICO whitepaper deep dives at three per day. The 'tech' was often a PDF and a dream. The talent pool was even smaller—mostly crypto-native kids who had never built a production system. But the industry grew because it promised something new: financial inclusion, decentralization, peer-to-peer value transfer. It attracted idealists. Now, after the collapses of Terra, FTX, and a dozen bridges, the idealism is replaced by cynicism. Young engineers look at crypto and see a casino run by anon founders with yacht photos. They look at AI and see a chance to build something that actually impacts people's lives—from drug discovery to autonomous driving.
Hyperliquid is a useful case study. As a decentralized perpetual exchange, it sits at the intersection of DeFi and CeFi. Jeff Yan's call for a 'rebuilding of financial engineering from first principles' is exactly what the industry needs—but it's also exactly what the industry has been failing to do. The reality is that most DEXs are still clunky copies of centralized order books, with latency issues and liquidity fragmentation that make them unusable for serious traders. We didn't need a podcast to tell us that the chain-based financial renaissance is stillborn—the TVL numbers for most L2s show a round peg in a square hole: the same user base sliced across dozens of networks.
My contrarian take: the talent war isn't being lost to AI. It's being lost to crypto's own addiction to speculation. Top engineers don't want to build the next yield farm; they want to build infrastructure that solves real problems—identity, supply chain, energy credits, decentralized science. Yet the industry's narrative is still dominated by price calls and meme coins. Jeff Yan's 'renaissance' rhetoric implies that if we just rebuild finance on-chain, the talent will come. But that's backwards. The talent will come when we show them that crypto can be more than a casino.
From my years auditing tokenomics, I've seen this pattern repeat: a project raises $50 million on a 'DeFi 2.0' pitch, hires a marketing team of 20, and has exactly three engineers. The codebase is a fork with a few lines changed. The result? A product that works in a bull market and dies in a bear. Meanwhile, the two engineers who could have built something novel are working on weekends because the company can't afford more. The industry's evolution from ICO to DeFi to L2s is a story of diminishing returns on innovation—each wave promises a breakthrough but delivers a copy.
Let's talk about the elephant in the room: why would a PhD in distributed systems choose crypto over AI? In 2021, I interviewed a researcher who had worked on Solana's consensus algorithm. He left in 2023 to join an AI startup. His reason? 'In crypto, my work is measured by TVL and token price. In AI, my work is measured by model accuracy and real-world utility. I want my career to mean something.' That's not a failure of crypto's technology—it's a failure of its culture.
So what does the 'on-chain renaissance' need? It needs to stop being a marketing slogan and start being an engineering roadmap. It needs to address the real reasons talent is leaving: unclear regulatory paths, toxic communities, short-term thinking, and a lack of impact beyond speculation. Hyperliquid can be a part of that if it focuses on building a robust, low-latency DEX that actually competes with Binance—not just in TVL but in user experience. But that requires hiring people who understand high-frequency trading, networking, and insurance—skills that are also in demand at traditional hedge funds and AI labs.
The takeaway is uncomfortable. If crypto wants to win the talent war, it must stop talking about 'renaissance' and start building infrastructure that serves real economic needs. That means prioritizing engineering over marketing, sustainability over hype, and utility over speculation. Until then, the smartest minds will stay in AI—and they'll be right.
The industry's evolution into 2025 will be defined not by how many L2s we can launch, but by whether we can retain the people who can build the next paradigm. Jeff Yan's warning is a canary in the coal mine—but the coal mine is on fire, and we're still arguing about who should hold the hose.