Ideas
Things we can achieve
Things that we can achieve with current technology (or maybe with some tech that is coming our way).
What if you could feed in the audio book to this machine? What if you were an author of a creative content? You could simply speak to the machine while it listens inventively to your creative ideas and creates a book or a movie out of it! Greater the details, better it is. But if you are short of them, it will plug in the gaps for you so that the output is coherent.
Though this is not a possibility today, it will be the holy grail of AI. As it stands today, AI is good at solving small tasks, replacing humans where humans would take quick decisions i.e. take decisions in extremely short duration of time. Even one of the most complex task of AI driving a car is actually a series of extremely quick decisions taken while being on the road.
The Ideas to Reality machine
June 30, 2019: Great story books often get adapted as movies. Each story book inspires as many movies as the number of its readers. Every reader has his own imagination and plays the storyline of the movie in the head as he goes through the pages of the book. What if a machine could do the same? It takes the book, understands the storyline, context and meaning. If required, the machine fills in the gaps and details of the storyline, exactly like a reader does - imagining in his head! Let the machine create a movie for you right away. No waiting for movie adaptation of your favorite books. The reader has turned into a movie creator.What if you could feed in the audio book to this machine? What if you were an author of a creative content? You could simply speak to the machine while it listens inventively to your creative ideas and creates a book or a movie out of it! Greater the details, better it is. But if you are short of them, it will plug in the gaps for you so that the output is coherent.
Though this is not a possibility today, it will be the holy grail of AI. As it stands today, AI is good at solving small tasks, replacing humans where humans would take quick decisions i.e. take decisions in extremely short duration of time. Even one of the most complex task of AI driving a car is actually a series of extremely quick decisions taken while being on the road.
Intertwined blockchains
July 28, 2019: Below synopsis discusses the intertwining of blockchains and its practical use. While this concept is valid for both public and enterprise blockchains, it is more useful in case of enterprise blockchains that are not backed by well-known entities.
Acclaimed Blockchain - Enterprise blockchains by industry well-knowns
We now have several uses case of enterprise blockchains, but most of those blockchains are promoted by consortia that have well-known participants, where users inherently trust them. This is important because the original inspiration of blockchain – the Nakamoto Consensus Blockchain – is an extreme case that assumes no trusted participant. However, enterprise blockchain – which are efficient implementation of blockchains (and hence currently have use cases) – put this trust in a few trusted participants in the blockchain. Hence, it is important that the trusted participant is indeed trustworthy – where the participant has earned that trust in the industry through its behaviour or actions over a period of time. Indeed, trust is earned. Examples of blockchain use cases that are backed by such “acclaimed” trusted parties include:
• Quorum by JPMorgan based on Ethereum implementation
• Lygon by ANZ, CBA, Westpac and IBM based on the Hyperledger blockchain
• Blockchain based financial auditing where Deloitte, KPMG, EY and PwC have partnered with Taiwanese banks.
• Winding Tree blockchain in partnership with Nordic Choice hotels, as well as Lufthansa, Swiss Air, and Eurowings.
Of, course these are use cases created by well known experts of that use case target industry. For example, Lygon brings blockchain technology to the world of bank guarantees.
Trust begets trust and continues to build on trust
Why would industry participants and consumers trust a blockchain solution to store their transactions including an exchange of money or value?
Apart from implementing a blockchain-based solution, the other common thing about the above solutions is that they are promoted by well-known reputed parties which most participants will generally trust even outside their association with a particular blockchain solution. Trust in the well known, trusted industry player automatically gets transferred to a blockchain solution where the player is a part of the consortium or the one that promotes the blockchain solution.
This trust cannot be bought. It is earned. It can be transferred, though, through association, when the trusted player associates itself with a blockchain solution and seeds it.
What if a relatively unknown market player or group of such players bring a blockchain solution to the market. Will that solution gain enough audience to become a serious contender?
Well established players seeding a blockchain can create an oligopoly
Enterprise blockchains require permission to join and use the solution. A blockchain created by an established industry player or players can stifle competition by either preventing other players from becoming part of the blockchain solution or by crowding out smaller competitors from the market as customers start to gravitate towards a trusted implementation of the solution. A key reason for this behaviour is that blockchain is inherently difficult for most users to comprehend and when money is involved such users would often go with the trusted player. This effectively has the potential of creating an oligopoly in the market.
Why would industry participants and consumers trust a blockchain solution to store their transactions including an exchange of money or value?
Apart from implementing a blockchain-based solution, the other common thing about the above solutions is that they are promoted by well-known reputed parties which most participants will generally trust even outside their association with a particular blockchain solution. Trust in the well known, trusted industry player automatically gets transferred to a blockchain solution where the player is a part of the consortium or the one that promotes the blockchain solution.
This trust cannot be bought. It is earned. It can be transferred, though, through association, when the trusted player associates itself with a blockchain solution and seeds it.
What if a relatively unknown market player or group of such players bring a blockchain solution to the market. Will that solution gain enough audience to become a serious contender?
Well established players seeding a blockchain can create an oligopoly
Enterprise blockchains require permission to join and use the solution. A blockchain created by an established industry player or players can stifle competition by either preventing other players from becoming part of the blockchain solution or by crowding out smaller competitors from the market as customers start to gravitate towards a trusted implementation of the solution. A key reason for this behaviour is that blockchain is inherently difficult for most users to comprehend and when money is involved such users would often go with the trusted player. This effectively has the potential of creating an oligopoly in the market.
Blockchain implementation by a non-acclaimed party
Let us consider an example of a logistics provider who wants to use blockchain to keep track of the movements of goods within its organization or partners organizations. The player implements a blockchain solution and seeds it. Why should you trust its implementation of the blockchain? After all, the company can simply abandon a blockchain or the last few transactions on the blockchain, or create a parallel fraudulent blockchain fork and replace the original with that at a later date. What if one was to dispute the timing of the transaction recorded on the blockchain? What if the company decides to invalidate all the blocks created after a certain time or a certain number of transactions? Is there a way to avoid these from happening? Is there a way for the non-acclaimed party trying to establish a foothold to explicitly build trust in its implementation of the blockchain? Especially, if that blockchain contains only a few trusted parties that are not well established or acclaimed in the industry as trust-worthy parties. They have not spent enough time to demonstrate their behaviour and earn trust, yet.
Intertwining of blockchains to achieve commit and time-synchronization
In situations like the one discussed above, the blockchain can “commit” the signature of the current state of the blockchain to a more well-known blockchain (let us call this the acclaimed blockchain). This in effect freezes the transactions concluded up to that point in time (let us call this an epoch) and also provides temporal evidence (time synchronization)– the transactions represented by the signature would have happened on or before it was committed on the acclaimed blockchain.
Thus, we achieve two key things with the intertwining of the blockchains:
- Transactions up to an epoch getting committed to an acclaimed blockchain
- Time synchronization with an acclaimed blockchain's timeline
If the blockchain gets intertwined with several acclaimed blockchains, the transactions in the blockchain will be as good as committed for all practical purposes. This is because:
- Well known trusted parties across different blockchains will have to act in concert to repudiate the acclaimed blockchain which in turn will then repudiate the commits made on that blockchain.
- Not all the acclaimed blockchains will go offline or be corrupted at the same time.
The risk is still with the developing epoch that is not yet committed to any acclaimed blockchain.
There are some solutions to build trust for such developing epoch. One is to commit the transactions as soon as possible to the acclaimed blockchain. The other one is to incorporate smart-contracts in those blocks that will “back” the transaction by a guarantee e.g. cash (cash collateralize) until they are committed on one or two acclaimed blockchains. This ensures that collateral is tied up in the transactions only for the duration of the epoch development, and once it is committed to an acclaimed blockchain, it is released, thus minimizing the time that it is blocked or tied-up due to the blockchain transactions.
Blockchain memory
Jan 15, 2020: Memory is chronological. Blockchain records chronological events. Ideal memory is permanent, and blockchain provides an immutable record. Hence, blockchains are best placed where events are to be recorded on a permanent basis. The "memory" property of the blockchain is best suited for transactions (e.g. in banking, supply chain etc.).
Another great example is use of blockchain in notary services involving verification of ownership. If ownership of asset is recorded on blockchain when it is created (e.g. intellectual property) or when transacted (the transaction itself), then it can be used as a proof of existence - when it was created, by whom it was created and what was created.
A simple algorithm to implement a proof of existence system is as follows:
Step 1: Take a hash (e.g. SHA256 hash) of the intellectual property document. Keep a copy of the document in backup as well.
Step 2: If there are several documents, then collect the hash of the documents generated in Step 1 in a new file.
Step 3: Take a hash of the document containing the hash (that was generated in Step 2)
Step 4: Write the hash generated in Step 3 on a reputed blockchain like Bitcoin. Once this is confirmed and post 6 to 7 transactions, we can reasonably say that the proof of existence has been recorded.
AI assisted trading system
Aug 15, 2020: It's nearly impossible to figure out an algorithm for an effective stock trading system. AI systems are best placed to come up with some form of algorithm to assist trading.
One can argue that stock prices incorporates all information there is that can influence a stock. However, this assumes an efficient market. If the price information is assisted by other information especially, leading indicators, then the output can be much better.