GRO MONEY NEWS - 🧠 AI was meant to come before blockchain

For the real estate applications we were building at Consortia, the missing piece was the computational capacity to make fragmented information useful before asking institutions to rely on a shared record of it.

🔗 We were presenting property identity and provenance in 2018. In 2024, I told RESO: “AI and networks are finally where I need them to be for blockchain to be efficient.” The dependency was already clear.

⚡ What we are building in stealth starts with abundant supercomputing. We have the capacity to read, parse, sanitize and organize authorized data at the most granular meaningful level, connecting individual claims to their sources and identifying contradictions that a summary would conceal.

🔍 AI helps interpret the information. Supercomputing enables that work at scale, while blockchain provides a verifiable history of what was submitted and how it changed. Together, they create a stronger foundation for decisions, provided the evidence and permissions remain attached.

🚀 Consortia was early, and I am carrying that thesis forward. The opportunity is to turn fragmented records into intelligence people can inspect and use, with greater control over the information that creates their value. The compute is here, and we are building.

#ArtificialIntelligence #Blockchain #Supercomputing #RealEstateTechnology

Duration: 8m
Broadcast on: 16 Oct 2026
Audio Format: other

AI was meant to come before blockchain in the commercial sequence that matters for real estate. First, make the underlying information intelligible, then make its history and permitted use verifiable across institutions. We built consortia around the second problem, while the first remained expensive, fragmented, and dependent on human labor. My view is that AI paired with supercomputing changes the economics of that sequence and allows the original blockchain thesis to reach much further. In my 2018 presentation to the Real Estate Standards Organization, consortia described a digital identity for property that connected its history with the relationships and events surrounding it. In a February 2019 interview with the FinTech Times, I explained how public records, private information, and documents contributed by homeowners could converge around that identity. The opportunity was to make the information surrounding a property more useful to the people making decisions about it. That argument continued through my writing on personal data in 2022 and property records in 2023. Then, in October 2024, I told RISO, AI and networks are finally where I need them to be for blockchain to be efficient. The dependency was already explicit. Consortia was early to an infrastructure problem whose computational requirements were becoming increasingly clear. A property has records scattered across institutions, formats, and decades. An address changes, a parcel is divided, an owner uses several legal entities, or a renovation appears in an invoice but never reaches another database. Before any system can use those records together, it must establish which asset they describe, when each claim applied, and what evidence supports it. Putting the documents on a ledger leaves that interpretive work to be done. Blockchain can preserve evidence of what was submitted and make subsequent alteration detectable under the network's rules. It cannot, by itself, establish that a statement about the physical world is true. NIST identified this boundary in its 2018 blockchain overview. Information supplied by people or sensors may fail to reflect real events. Preserving an incorrect statement creates a durable record of the error. The argument is specific to applications built around complex information about the real world. Blockchain already supports transactions involving assets native to its own networks. Property Intelligence asks more of the connection between a digital system and an external asset, which is why the quality of that connection matters so much. AI helps interpret the material crossing that boundary. Accelerated computing makes much more of that work practical. The computational work extends well beyond answering questions in a chat window. It includes extracting information from documents, resolving duplicate records, comparing sources, and rebuilding indexes as information changes. NVIDIA's data processing platform documents acceleration for both structured analytics and unstructured vector search. These capabilities support an approach in which preparation and retrieval receive the same attention as the model that produces an answer. In our stealth company, abundant supercomputing is the starting condition. It gives us room to read, parse, sanitize, and organize the data we are authorized to process at the smallest meaningful level, then revisit it when the source changes. We are building around the capacity to examine the underlying material comprehensively, with the evidence attached to each useful claim. The performance standard is how much validated information the system can produce, at what cost, and with what delay. Granularity changes the product. Consider a hypothetical property with an inspection describing an aging roof, a later contractor invoice describing a replacement, and a permit record that lacks a final sign off. A summary saying that the roof was replaced conceals the unresolved question. A useful system separates the contractor's claim from evidence that the work was completed and identifies what remains to be verified. At that level, each material assertion carries its own source, relevant date, and verification status. An underwriter can inspect the evidence behind the roof claim, while a homeowner can see which document would resolve the uncertainty. A corrected record can trigger a new analysis without silently rewriting the history of the earlier conclusion. That is a much stronger foundation for a decision than a collection of files or an unexplained score. Sanitizing the data must preserve that distinction. Standardizing an address, removing duplicate copies,

Transcription truncated.

#ArtificialIntelligence #Blockchain #Supercomputing #RealEstateTechnology