 Sudhakar asks, is a smart contract legally accepted? What happens if one party defaults? That's a really great question. I think the word smart contract is very confusing for most people. What is a smart contract? Well, it's not smart, and it's not a contract. A smart contract is a dumb program. A smart contract is a program written in a language that is executed inside a blockchain-based virtual machine. It's a program, and they're not particularly smart. They're fairly unsophisticated software programs. The important part being that they manage money. Could they have legal implications? Could they be legally binding? Perhaps, especially if combined with some kind of written legal contract. But there are many jurisdictions in which various forms of agreement, whether written or not, are considered legally binding. There's no reason why a smart contract could not be legally binding. However, this depends on the jurisdiction, and it also depends on what the legal precedents are. Until there is more established case law in this case, and I'm not a lawyer, so I'm just giving you my opinion on this, until there's more established case work on this, the term smart contract is very loosely used and should not be considered a contract. For the second part of your question, what happens if one party defaults? Well, if the smart contract handles that particular case in its program, then something happens. Something happens on the blockchain. Is that legally binding or not? I don't know. That's a question you have to ask a lawyer. This is a really interesting thing, that this is not a matter of law. Smart contracts are not a legal contract. They are a software contract. They do things on blockchains. Whether those things are within the law, or accepted by the law, or recognized by the law, has nothing to do with how the smart contract works. Catarina asks, how is a token single-doubted and designated to represent, for example, ownership of a house? I can decide that one particular token will represent ownership, and then code that into the token. Can you explain how this works? Catarina, usually the way this works is you have some kind of external data store. This could be something like a decentralized file store like IPFS, or Ethereum Swarm. You would store some PDF document, for example, the deed of the house, or some other title, or a reference to a title registry, such as a municipal state or government registry of titles. You could take all of that information and encode it into a token. The real question is, if you transfer that token to someone else, does that also transfer the title? In most jurisdictions, no. A lot of people are trying to get laws passed that recognize the transfer of ownership of the asset, the house, when the token that contains the title or hash of the title is transferred. That is a bit harder, really. This is the difference between software and law. Just because you can transfer token and software doesn't mean that you can take that token ownership to your local planning office and say, hey, I own this land, I own this house, and have that legally recognized. That is one leap further down. Rural asks, where are Ethereum transactions stored after executing the contract code? That is a really great question, Rural. Ethereum transactions are stored on the blockchain the same way that Bitcoin transactions are stored on the blockchain. The difference is that Bitcoin transactions track the state of ownership of specific coins, or unspent transaction outputs, as they are called. In Ethereum, the transactions may trigger smart contract execution, and that can modify the state of a contract, or the state of an account within Ethereum. Ethereum transactions track or cause changes, or state transitions, in any form of state that could be data within a contract, or the balance of an account, a much broader set of state. What is actually stored in the Ethereum blockchain are two things. The first thing is the transaction itself, and the second thing is the transaction logs, or receipts, as they are called, which show what changes this transaction caused to the global state. The contract execution itself isn't stored on the blockchain, so you store what started the contract, and you store what state changes happened because of the execution of the contract, but the execution of the contract itself is not stored on the blockchain. You essentially store the inputs of contracts, which are transactions, and the outputs of contracts, which are state changes. Those are the two things that are stored on the Ethereum blockchain. Anyone can validate this information by taking that same transaction that is stored on the blockchain, playing it through the smart contract execution by running it on the Ethereum virtual machine, and then looking at the output of the Ethereum virtual machine, the logs, and seeing what state transitions occurred during the execution of the smart contract, and what state changes occurred, and comparing those with the receipts or logs that are stored in the Ethereum blockchain. Every single validating node on Ethereum should arrive at exactly the same answer. The same transaction should cause the same state changes and produce the same receipts and logs. That's how Ethereum validation is done. Kirily asks, is there enough being done to confirm that smart contracts will work in the way that people have envisioned they will work? Kirily, the problem here is that the vision is far more expansive than what can currently be done with smart contracts. People envision all kinds of things. Whether you can do all of those kinds of things is a whole other question. I think people sometimes go way too far in coming up with these crazy ideas to replace things, even things that currently work with smart contracts. The term smart contracts is beginning to become as vague and full of bullshit as the term blockchain or AI. It doesn't really mean what most people think it means. Simon asks, do you think it's possible to combine smart contracts with AI and machine learning? Maybe in very small ways, but not in any practical way and not anytime soon. These are both very mature, very new technologies, and when I say both, I mean smart contracts and AI and machine learning. While they are progressing really fast, they are not necessarily easy to combine. What happens if you take smart contracts and combine it with AI? Now you took all of the risks and problems of one immature technology, and you multiplied them by all of the risks and problems of another immature technology. It makes for a fantastic marketing brochure, and you will probably raise a lot of venture capital money from investors who don't really understand what AI or smart contracts are and are very impressed by these buzzwords. But the truth is that in practice, doing anything that involves two experimental technologies increases your risks exponentially. I don't expect we're going to see any useful or practical, relevant, interesting applications combining AI and smart contracts in the next several years, perhaps even decade.