Billions over billions are burned in expensive GPU-processes designed to structure unstructured data and information.

I’m not a technical engineer; I’m a logic person.

And I like the library example, where billions of books lie on the floor, never structured on bookshelves, always there, but one above the other. Now to find the right book – a horrible searching scenario.

And I asked myself: why not rebuild everything from zero? First determine a structure, thus organizing the library with a clearly structured searching method, and then build a new wonderful library with beautiful bookshelves.

Possible.

Horrible work too – but possible.

I thought: it could be possible to start with the consumer market. The consumer is searching for a 4-star hotel and the structured library presents all 4-star hotels related to some consumer-made indicators.

I think many platforms already work that way – don’t they? But still one question disturbed me: if the consumer market works that easily (demand question – supply answer = fit), why then do we really need those expensive GPU processes (mostly implemented in hybrid systems – however very expensive and economically – bleeding)?

Why?

Because: the entire market is built upon fragments.

Fragments in consumer footprints and thus fragments in consumer understanding. Fragments in supply information and data and thus fragments in supply understanding. AI should fill the gap – with expensive statistical and probabilistic processes. What could be interesting? What did others do? What could be a fit?

Sorry – so much money for probabilism.

Bleeding again – from the perspective of an economist.

And this is why I build SEPPY.

The crux is: SEPPY instantly understands Demand without AI, Third-Party Data or historical Data. It puts Demand in the same numerical structure as Supply and makes it thus instantly comparable and matchable.

But numbers would not be relevant if they were not explainable. And the explanation must be comprehensive.

SEPPY is:

→ Harmonizing the market with demand and supply twins all of the same mathematical structure

→ Demand twins modeled from zero, in near real time, with few user inputs in that numerical structure – a CPU-native, deterministic, O(1) transformation process.

→ Supply in the same numerical structure put in the common numerical state space for instant comparison and a trustworthy match

→ The numerical twins instantly explained through properly architected, patent-pending ontologies

→ Demand Twins instantly decouplable from any real user input data – thus fully GDPR-by-design anonymous, yet hyper-personalized

→ Never overwritten – but growing and scaling with the SEPPY patent-pending approach

→ Thus, explainable development and evolvement and relationships of every singular twin

→ 31 billion twins can be modelled

This is SEPPY – built by a non-technical person to prove that mathematical comparison and explainability are the counterpart of GPU-expensive approaches.

The Verge switches up its audience team to prepare for Google Zero