Mohit Mamoria


Here’s why decentralization is inevitable…

…and how not to miss the bus?

If you are anyone like me, you might have had a lot of people talking about blockchain and decentralization in the year 2017, but as we made progress through the year 2018, all of them went silent one by one. The prices kept going down, and so did the voice of several advocates. Relatable? And as the silence grew, you started thinking, blockchain is dead!

On top of that, there were frequent announcements from various countries who were taking a negative stance, and there were regular blog posts from who’s who talking about how Bitcoin is the greatest scam in history.

Amid such chaos, decentralization must already be dead, you thought. If you did, you wouldn’t be further from the truth.

In this blog post, without going too much into the philosophical ideas about governments, intermediaries, or trust, I’ll try to bring out the reasons why decentralization is an inevitable milestone in the way of always-improving technology.

Here’s the deal. We’ll walk together through some of the examples that we all agree upon to change the world, and then see how decentralization has an integral role to play in it.

When Machines Think 💭

“Forget artificial intelligence — in the brave new world of big data, it’s artificial idiocy we should be looking out for.” — Tom Chatfield

Since the dawn of the industrial revolution and machines, there were always humans operating them and making them work. But over last few years, we have seen them becoming more of our thinking partners. Speaking of identifying faces in a CCTV’s video footage, or translating a sentence from Korean to Mandarin, or playing a match of Go, machines have continuously beaten humans.

Still, what we see now in the name of artificial intelligence (AI) is merely the tip of the iceberg. We have organizations building AI models in silos and training them on their historical data. But the Pandora’s box will open the day we can have all these organizations collaborate on AI models that do better than what any of those individual AI models could do.

We are yet to see the moment in AI along the lines of ‘the whole is greater than the sum of its parts.’

Time for an example, I believe. Almost every bank has an AI that detects the probability of payment being fraud. Each bank has been training their version of the models on their historical data. It makes some banks better at fraud detection than others, which sometimes they position as one of their strengths to their customers. While only some people might buy banking services from a bank solely based on their expertise in detecting a potential fraud, the bigger problem of fraud payments is an issue for society at large. But for banks, their business and profits are equally (if not more) important than just doing good for the community.

If in an imaginary world, banks choose to work together and build a standard fraud detection AI model that serves all the banks equally, they will not be able to capture the economic value of their data. If banks choose otherwise, there would still be people suffering from fraud payments.

Interestingly, in such scenarios, decentralization allows us to capture the economic value as well as do good for the society at large. How exactly? Here’s how — banks together can train a common AI model that is stored on a peer-to-peer network (blockchain). Anyone can take the latest copy of the model from the network, train it further with their data, and submit it back to the network with a proof-of-training. If the network accepts that the additional training improved the model’s performance, it recognizes it and marks it as latest for others to train it further. Whenever the network accepts a trained model, the trainer gets paid in the network-specific tokens that can be used in return to use the model. That’s how the trainer captures the economic value from its data, and at the same time, the common trained AI model makes it possible for every bank — even a new one — to use the trained model — therefore, doing good to the society.

How machines will think in the future.

When Machines Talk 🛰

“The real problem is not whether machines think but whether men do.” — B. F. Skinner

Talking about intelligent machines, let us talk about self-driving cars. In a world where cars are driving themselves, they would need to communicate with other vehicles on the road too. That communication cannot happen over a centralized network. Why? Because, right now in 2018, if a centralized server goes down, a handful of websites go down. But in a world of self-driving cars, if a server would go down, there would be accidents on the road.

With machines talking to each other, dependency on centralized servers would be risky.

Also, with the rise of self-driving cars, interesting new economic models will come up. For example, how would a car decide should it give way to the other vehicle on the road or not?

I reckon vehicles can transact with each other on the way to optimize the rider’s preferences? If a passenger is in a rush, he/she would be willing to pay other cars on the road to give way. And riders who are not in a hurry can give way to this passenger to earn some incentives.

How machines will talk/transact to each other.

It will not be unusual to see riders picking one of the following options whenever they get in their cars before a trip:

  • Reach the destination as quickly as possible by paying others on the way
  • Reach the destination before the time but also be kind to others on the trip

These kinds of transactions will need to happen directly between peers. A server cannot be guaranteed to be up 24/7. But a decentralized network can.

When the stakes at which machines would be operating will increase, the world will drift towards decentralization on its own.

How not to miss the bus? 🚌

“Opportunities are like sunrises. If you wait too long, you miss them.” — William Arthur Ward

Be a learner while others are just being ignorant. As an engineer myself, I would urge you to think and understand decentralization beyond the volatile prices of tokens and coins. Building your own mental map of blockchain will help you make business decisions that will make your business future-proof. It will also lead you to foray into areas that didn’t seem inspirational earlier. And you don’t really have to be an engineer to understand these concepts and apply them. Especially, when the world’s leading universities have taken the plunge to create online programmes. The Oxford Blockchain Strategy Programme is one such programme by Saïd Business School, University of Oxford in collaboration with GetSmarter.

I have had an opportunity to take a sneak peek at the contents of the programme and I am thrilled to see that leading universities are taking the responsibility of educating us mere mortals. I plan to join the programme and tweet my learnings over a period of six weeks. Follow me around 🐦 if you want to see what progress I am making the through the programme.

Thanks for reading! :) If you liked it, please support by clapping 👏🏻 and sharing the post. Feel free to leave a comment 💬 below.

Mohit Mamoria

Mohit Mamoria is the CEO of Authorito Capital (a crypto fund) and the editor of a weekly newsletter, Last Week In Crypto, which delivers an email every Monday containing everything that has happened in crypto space in the previous week.

Have feedback? Let’s be friends on Twitter. 🙌🏻

More by Mohit Mamoria

Topics of interest

More Related Stories