The Human Layer: How PublicAI Turns AI Disruption Into Collaboration
AI won't destroy humanity, but it will redefine how we work. The question isn't whether humans will be replaced, it's how we'll collaborate.
When Google Engineering Director Matt Welsh predicted that "the job of a programmer will vanish in three years," he wasn't describing an apocalypse. He described evolution. AI will write code, but humans will train models, validate outputs, and teach AI new skills. The challenge isn't stopping this transformation, it's ensuring humans remain essential participants, not passive observers.
This is why we built PublicAI: to create the Human Layer of AI.
The Origin Story: From Gaming to Groundbreaking Research
Steven Wong's journey began in 2019 when he left his position as a game developer for NetEase's Fantasy Westward Journey to pursue a Ph.D. in AI under Professor Shuguang Cui. His research focused on the intersection of deep learning and biomedical applications.
During his Ph.D., he designed a protein structure prediction algorithm—DSM-Distil—tailored for low-homology sequences. It outperformed DeepMind's Nobel Prize-winning AlphaFold in multiple benchmark tasks and later earned him the U.S. EB1A Extraordinary Talent Green Card.
The real breakthrough wasn't algorithmic, it was philosophical. Steven realized that in AI, the upper bound of model performance is determined far more by data quality and diversity than by algorithmic innovation.
The Data Challenge and Web3 Solution
In early 2022, our lab began constructing the world's largest 3D version of ImageNet. To fuel this effort, we needed massive amounts of real-world data. That's when I first proposed using crypto staking + incentive + penalty mechanisms to drive global contributors. This idea eventually became the prototype for PublicAI's business model.
We successfully commercialized the dataset with clients such as Adobe and LumaAI. After graduating, I closed a seed funding round participated in by the Solana Foundation. PublicAI officially stepped out of the lab and into the real world.
Revenue-Backed Reality: A New Web3 Model
Rather than hosting flashy conferences or chasing hype cycles, PublicAI follows one principle: "Mint tokens only with real customer revenue."
The Business Model:
-
Demand Side: PublicAI has signed over $14 million in commercial deals with Web2 AI leaders like Adobe, LumaAI, AWS, and Abaka.AI
-
Supply Side: PublicAI's DataHub platform converts client payments into USDT rewards distributed directly to global data contributors
Real Impact: In some countries, contributors to the Mother Tongue Audio Collection S2 campaign are earning over $10/hour, far exceeding their local average wage.
The Innovation: Revenue-Backed Token Issuance
At the core of our model is what we call "revenue-backed token issuance." For every dollar earned, only an equivalent worth of tokens is minted. The action is triggered by a DAO-based governance.
If a client pays X USD and the current token price is P, then only X/P tokens are released to verified contributors. Every newly minted token is backed 1:1 by real-world revenue, fully redeemable, and inflation-resistant.
Why Web3 Fits AI Training Data
Web3's fundamental strengths lie in globally distributed participation and permissionless openness, both of which align perfectly with the needs of modern AI. AI training requires diverse, high-quality datasets, but traditional centralized data collection often leads to homogeneity and overfitting.
PublicAI solves this by incentivizing users around the world to contribute and validate data, unlocking linguistic, cultural, and environmental diversity. Combined with on-chain reputation systems and traceable data provenance, PublicAI guarantees verifiable data quality and transparent reward distribution.
This isn't Web3 riding the AI hype—it's AI relying on Web3 to scale sustainably.
The Human Layer: Always in the Loop
As Matt Welsh remarked, "The job of a programmer will vanish in three years. AI will write code, and humans will be tasked with training models and teaching AI new skills."
At PublicAI, we're not just creating a platform for humans to do grunt work for AI—we're building the Human Layer of AI. A layer where real people participate across the full AI lifecycle, not just data labeling, but also inference validation and decision alignment.
Our vision is a future where human and AI collaboration is the norm, not the exception. Where humans are "always in the loop," not phased out of it.
The Bigger Picture: Redefining Work
Throughout history, every technological leap has forced humanity to renegotiate its relationship with tools. Take an example, in 2018, 60% of existing jobs didn't even exist before 1940. This time, however, we're not dealing with a new tool, we're facing an evolving, learning system.
That's why PublicAI is building a human-AI symbiosis, a future where every individual has the opportunity to work with AI, not be replaced by it.
Build the AI Future with Web3
Crypto will only return to mainstream relevance not by stacking buzzwords or flexing tech jargon, but by building truly useful Web3 applications with real business loops, applications that bring off-chain value into on-chain economies. That's the only way to regain market trust and rebuild conviction among traders and builders.
PublicAI represents that future, where blockchain technology solves actual problems, creates sustainable economic models, and empowers human participation in the AI revolution.
After all, the vanishing programmer job is just the tip of the iceberg.
Steven Wong, Ph.D., San Francisco