The current state of blockchain development is that most public blockchains still rely on the Bitcoin-like proof of work algorithm. While it certainly works at reconciling the data across nodes, it's by design and incredibly inefficient way to achieve consensus. It burns way too much energy for practically nothing and as a result, has some serious scaling issues. Consequently, a main cutting edge into development of blockchain technology is new consensus protocols that aim to achieve the same results as proof of work, but dramatically enhance it's efficiency, and do so without relying on private blockchains or expensive vetting of nodes. The most talked about of these emerging developments is called proof-of-stake, which is currently used in the block chain platform Neal, and was originally proposed to be used by aetherium, but as of now, development is still ongoing. In this video, we'll have a high-level conceptual overview of proof of stake, and look at its advantages and drawbacks, both of which are quite obvious. I'll include some references to the other new developments brightly labeled as practical byzantine fault tolerance protocols, at the end of the video. While it sounds quite complicated, proof-of-stake is essentially Virtual mining. Let's go back to proof-of-work for a second. The idea is that, to prevent easy faking of identities in the random consensus process, each node in the network should be backed up by actual computing power. Mining, therefore, is equivalent to proving who has invested the most money in hardware and energy. Now, I want you to pause a bit and think about the process. There's one component of the process that's not necessary. The proof-of-work mining game is nodes that have invested more money in hardware and energy, have a higher chance of winning the block. Well, we can cut out the hardware and energy part and get exactly the same result. Simply modify the mining game to the nodes that have invested more money, have a higher chance of winning the block. This is proof-of-stake in a nutshell. Instead of proving who has spent the most money to buy the most amount of computing power, to get the same result, it's in fact much more straightforward to directly prove who has spent the most money. This can be done completely virtually, skipping the hardware and energy costs altogether. Let's use the same illustration to see how we might implement proof-of-stake. Same board, same four miners. Now, instead of allocating the board space to miners based on their computing power, let's just ask them to directly buy the board space instead. So for each new block, they're going to build them individually as usual, but instead of appending a nonce and hashing, let's just required the nodes to place a monetary bet on the block with the cryptocurrency that they have. The bet will be used to purchase for space. Once the bet is put down, is locked up for multiple periods, regardless of whether you win or lose, and you'll get it back after a certain amount of time. Now, same process as before. Andrew has put down 50 coins for his block, Bob put down 25, Carol and David put down 12.5 coins each. So each of them rented a board space of different size. Notice now, this is exactly the same board, as in the proof-of-work example. Now, we're going to spin the needle as before. Notice that the probability for each of the miners winning this block is exactly the same as in the proof-of-work case. Yet this is done completely virtually, using cryptocurrencies, without using any real resources like computers or energy whatsoever. Now, one caveat about this process is that, it's easier to implement if no new coins are minted. So if the minor wins the block with this bet, he'll be rewarded by transaction fees only. You can probably immediately see the benefit of proof-of-stake. It's virtual mining without real resources. So it's infinitely more efficient than proof-of-work. You can design your block however you like. it could be as short as you like, and there could be as much data in the block as you like as well. Therefore, the system is much more scalable than the proof-of-work-based block chains, and theoretically could rival existing systems like Visa in processing transactions, while remaining a completely decentralized system. The other benefit is that if the blockchain carries a cryptocurrency, proof-of-stake also incentivize miners to behave and not carry out bad things like 51 percent attacks. The intuition behind this is that large miners hold a large stake of the cryptocurrency. If they do something to compromise the system, users will quickly lose trust of it and abandon the blockchain, leaving its large stake worthless. So for his own benefit, a large stakeholder would tend to behave in a way that's beneficial to the system's long-term survival. Now, the exact flip side of this benefit is also proof-of-stakes' biggest straw back. Under this protocol, we can easily have a rich gets richer problem. To see this, suppose that someone is initially in doubt with a larger stake than any other minor. Then, this node has more resources to bet in each round. So theoretically, he could win more rounds than anyone else, or any more money than anyone else. Therefore, asymptotically at the limit, this guy will end up owning every single coin, and the system would obviously collapse. Of course, this extreme case is not likely to happen due to the survival incentive that we just talked about, but he could easily lead to a large degree of state concentration, just like in proof-of-work, reducing the system to more of an oligarchy rather than a true decentralized network. This is just one of the challenges facing the new alternative consensus algorithms. Nevertheless, researchers from all corners of the industry, computer scientists, game theorists, and economist, are all working on addressing these challenges to produce truly efficient, decentralized consensus protocols, which I think is critical for the long-term survival of the crypto and blockchain industry.
