| HN Mirror

I'm not up to speed on the mathematics, but a matchmaking algorithm that keeps preference data secret until all participants in a trade web commit to a deal that satisfies at least one of those preferences could drive a lot of middlemen and scam-like businesses out of the market.

For instance, you might find an algorithmic pickle agent in the network, and tell it that you could eat one big jar of crisp dill cucumber pickles every two weeks if it costs less than 8 money units, delivered to your door, or one per week if it costs less than 3 money units, committing some number of money units greater than 8 to back a promise to buy at those prices. A small-time pickle-making farmer might tell the agent that they can ship at most 500 jars a week, as long as they get at least 2 money units per jar, or as many as 800 if they can get 4 per jar (cost of hiring a dedicated packer, perhaps). The pickle agent consults with a commodity shipping agent, calculates a billion different ways to move pickles from suppliers to consumers, and then starts moving money and pickles around. Everyone who promised to buy at a certain price is guaranteed to get the goods at that or a lower price, and everyone who promised to sell at a certain price is guaranteed to ship the goods at that or a higher price. The shippers get their fee for moving a package from point A to point B. The agents take their cut to pay for their computation, and for insurance against failed shipments or bad pickles. The system would also need to include distributor/importer/resellers, because some trades just aren't possible unless you pack a whole pallet of pickles, or a whole truckload/shipping container, and break that out for individual orders closer to the consumers.

That's all technically possible with smart contracts, as far as I know, but it would require a huge amount of programming effort to even get the basics correct. And Wal-Mart already has their supply chain, inventory, and distribution software in place.