Startup Improbable.IO claims to solve the parallelization problem for distributed computing

For those who call bullshit on what Improbable claims to have achieved, there is an open invitation to come and see the technology. “Come into the office,” says Button-Brown. “We’ve got something working. It’s not perfect, but it’s hard to say it’s not working when it is working.”
I understand their objective – to write simulator code as though for one processor. I’ve no idea how they implement the abstraction (of entities that interact in space) on a distributed computing platform with fault tolerance. Here’s a snippet from a 2014 Wired UK article:

…Improbable was set up and funded (to the tune of £1.2 million so far) by 26-year-old Herman Narula — son of billionaire construction mogul Harpinder Singh Narula — with some of his friends after graduating in computer science at Cambridge. Their main aim was to take distributed systems used in high-frequency trading and apply them to games to enable massively multiplayer experiences that have the richness of gameplay of a first-person shooter. “You could have a Call of Duty experience with an entire army. You can have hundreds of thousands of entities in the world with a simulated city with traffic infrastructure,” Narula explains.

Wired.co.uk first visited Improbable at the end of 2013 in a large, opulent residential house called Hyver Hall — owned by Narula’s family — in High Barnet. A dozen or so guys were set up in stations throughout the property working on server architecture, simulation software and games ideas. Many of them were sleeping there at night and there was a palpable sense of optimism, spearheaded by Narula.

Wired.co.uk sits down with Narula, CTO Rob Whitehead, software engineer Peter Lipka and studio head Nick Button-Brown. Narula is effervescent and speaks with the self-belief of a seasoned entrepreneur. Within minutes of meeting he’s illustrating how the technology works on a white board, wiping away previous workings with the sleeve of his pristine white shirt. Every now and then his teammates interject with “be humble”. Narula will temper his words for the next few sentences before lurching back into game-changing, world-dominating hyperbole. The over-arching ambition, he says, is to be the “Google of simulation”. His evangelism is both infectious and a little suspicious — one part cult leader, one part CEO.

Later, in a phone interview, Wired.co.uk asks why his team kept on asking for humility. “Because we started this company off with a level of audacity that was borderline insane,” Narula says. 

The gaming application may be a big deal. But what got my attention is simulating the real world. Especially the real world instrumented with IOT (internet of things) like cities and vehicles.
“If people want to do large-scale traffic simulations, military simulations, see how a disease might spread or anything where you are simulating outcomes with large numbers of autonomous entities, our technology can help,” Narula says, adding that he’s speaking to a “whole bunch of professors” as well as a number of “really unlikely non-gaming application areas”, including an airline. Deals, he assures, will be imminent.
Andreessen Horowitz invested $20M. Here is a relevant a16z podcast The Cool Stuff Only Happens at Scale.