"Interview with Dongqu" 0G Founder Michael: The Risks of Centralized AI Monopoly and How Decentralization AI Can Solve the "Terminator" Dilemma.

In the face of the monopoly and “terminator” dilemma that centralized AI may bring, Michael Heinrich, founder of 0G Labs, proposed a solution in an interview with the moving area. (Synopsis: 0G Binance up more than 500%: dual-track strategy to build the next generation of decentralized AI infrastructure) (Background supplement: Flora Growth completes its first 0G purchase, buying 772,200 0G coins at an average price of $2.59) From the Internet boom in Silicon Valley, to Wall Street's largest hedge fund, to the creation of Web2 unicorns, 0G Labs CEO Michael Heinrich has a proud resume. Today, his All in Web3 goal is how to prevent AI from becoming a “black box” controlled by a few companies and ensure that it becomes a public good for the benefit of all. The answer he gave was - 0G, a full-stack “decentralized AI operating system”. This exclusive interview with Michael explores why he regards centralized AI as a potential “terminator”, how 0G can build a transparent, verifiable, community-owned AI new future through blockchain technology, and their complete layout in go-to-market strategy, technical architecture and ecosystem construction. This is not just a conversation, but a deep reflection on two possible paths for AI in the future. Moving Zone: Hello Michael! Your background spans Silicon Valley and Wall Street, from SAP, Microsoft, Bain, to Bridgewater Fund and founding unicorns. What made you finally go into Web3 and choose the direction of “decentralized AI”? Michael: I grew up in Silicon Valley and was introduced to technology very early. What really made me take crypto seriously was when I was in graduate school at Stanford, listening to Tim Draper, Marc Andreessen, and my classmates talking about Bitcoin. In 2016-17, I worked on a lot of ICOs, saw a huge explosion of creativity in this space, and thought that one day I would have to dedicate myself to it. In 2023, this opportunity came. I sat down with my co-founder — I think the best engineer I've ever worked with — and thought, what are we going to do? At that time, ChatGPT was ushering in an explosive moment, and we realized the huge potential of AI, but we were also very worried: ten years from now, when AI starts running social-level systems, if it is controlled by a few companies, it becomes an opaque “black box” that could bring very negative results to humans, even a “terminator-like” scenario. So, we decided to create something completely opposite – decentralized AI. A transparent, verifiable, community-owned, and ultimately public (public good) secure AI. Dynamic area: You call traditional AI “black box AI”, and the core of 0G is “decentralized AI”. What is the most fundamental difference between the two in your definition? Michael: The fundamental difference is the trust assumption and the engagement model. In “black box” AI, you don't know anything about data sources, labelers, censorship mechanisms, model versions, and you can only “trust” a company unilaterally. Privacy is also not guaranteed. And in “decentralized AI”, based on the immutable ledger of the blockchain, you no longer need to “trust” an entity, you can personally “verify” everything: where does the data come from? Who marked it? Which version of the model am I using? Everything is transparent. This also leads to a whole new model of engagement. In a world of centralized AI, where corporations seize all value, ordinary people may face job losses. And in our model, anyone can directly own and participate in the construction of AI. For example, you can contribute data and hash power with your friends to train an expert model, tokenize it, and benefit from it. It's an ecosystem of opportunities and participation. Dynamic Zone: 0G's recent TGE has been so successful that you completed a NASDAQ-listed DAT (Digital Asset Treasury) collaboration even before the TGE, the first of its kind in the industry. Can you share the strategy behind this? Michael: Our TGE does have a lot of success. Regarding DAT, we are now the Executive Chairman of a NASDAQ-listed (FLGC) whose goal is to continue to buy and accumulate 0G tokens to provide natural buying pressure to the market. Doing this before TGE requires a lot of legal, financial, tax, and even relationship engineering, such as how to value an asset that has not yet been publicly traded. I'm proud that our team has done just that, and the core members of this team are the same people who previously built the Solana Treasury (DFDV), which is second only to Microsoft's strategy. As for TGE itself, success requires a lot of micro-decisions. From selecting a trusted market maker to designing option structures and liquidity parameters; From exchange strategies, to market budget allocation; Then work with the right communities and influencers. We were also fortunate to have received strong support in the Korean market, with Bithumb and Upbit launching on their first day. Dynamic area: 0G claims that its architecture can achieve an impressive throughput of 50 GB/s, how is this achieved in the context of the blockchain “impossible triangle”? Are there any disruptive technology trade-offs? Michael: That's a good question, it gets to the heart of our technology architecture. To achieve extreme performance while remaining decentralized, our core design principle is to minimize the amount of information that must be processed by the consensus layer. You can imagine it as a national governance system. If all the trivial things were to be decided by the Supreme Court ( the consensus layer ), the whole system would collapse. Our approach is that for massive amounts of data and computations that can be processed outside of consensus systems ( such as most ) of the process of AI model training, we use sophisticated sharding techniques that allow it to scale horizontally almost infinitely. It's like local governments and functional departments that can deal with a large number of things in parallel. So how do you ensure that the work of these “local authorities” is credible? That's the key to our technology – verifiable storage and computing technology. It acts as a bridge between the non-consensus layer and the consensus layer, ensuring that all off-chain processing results can be verified quickly and cheaply by the on-chain consensus layer. In addition, our innovative shared staking (shared-staking) consensus architecture enables the consensus network itself to scale while maintaining decentralization. This “decentralized” but “verifiable” system is the key to breaking through the impossible triangle without sacrificing security or decentralization. On the decentralized AI track, there are projects like Bittensor that create a marketplace for AI models through incentives, as well as infrastructure like Celestia (DA), Akash ( compute ). How does 0G view its competitive landscape? What is your moat? Michael: It's a multidimensional question. First of all, it is difficult to say who our real competitors are at the moment. Inside Web3, Celestia or Akash just intersect with us on a single link,…