Quote from: x3m on August 20, 2018, 04:09:03 PM

You're not relying on a centralized service but on a SC (decentralized) developed by a centralized service provider (i.e. airline company). The benefit is that the SC acts as an escrow. The airline company will be fully paid if the flight lands within the time interval otherwise the client will be automatically refunded (no forms to fill, no requests to submit, ...). In addition the SC is censorship resistant, no one can modify or cancel it.

The benefit of a smart contract acting as an escrow is rather limited from an airline company's perspective though. Sure, as a consumer it would be nice to get an automatic refund in case of flight delays. But there's no incentive for an airline company to (1) spend money on developing a smart contract for (2) keeping money in limbo that they otherwise would have instantly at their disposal while (3) giving up one of their favourite dark patterns (ie. it is not in an airline company's interest to make getting a refund easy). Put differently, smart contracts need to benefit all parties involved to see practical usage.

Quote from: Samarkand on August 20, 2018, 06:56:01 PM

Quote from: HeRetiK on August 20, 2018, 12:43:46 PM

...
In a more optimistic case, one could imagine a future iteration of current voice assistants taking care of more complex (financial) tasks. The latter being a more speculative use case, obviously.
...

I´m not sure if I understand your example correctly. Why is a smart contract necessary
for a future iteration of a current voice assistant? This seems like it could be done
without any smart contract and rather with some form of AI.

Sorry for being unclear -- I don't think that smart contracts are necessary for future voice assistants. What I mean is that smart contracts could be one of the many interfaces future voice assistants will be interacting with. To give an example: Right now voice assistants are acting on their user's behalf by either calling an API (eg. Alexa calling Spotify's API to play Despacito) or by making a phone call with a human (eg. Google Duplex making a reservation by calling the restaurant of your choosing). In the future such a voice assistant could additionally interact with public smart contracts (eg. placing an order on a DEX).

Come to think of it, it's a rather bad example though. It highlights how one could interact with smart contracts in the future, rather than what the smart contracts themselves would do.

Quote from: Samarkand on August 20, 2018, 06:56:01 PM

On an unrelated note, I think that a major problem with smart contracts is the fact
that all eventualities have to be put down in the code right at the beginning.
In reality contracts are constantly evolving, because of unexpected events that
weren´t specified in the contract. This seems like one of the major issues for now.

Excellent point that is often overlooked. Laws leave room for interpretation precisely because not all eventualities can be foreseen and not every factor that played part in an outcome can be quantified (eg. trolley problem).

To add to that, code being law adds another unintended side-effect: Unintended side-effects. From a puristic point of view they are of course simply part of the contract, in practice however they are simply bugs that could cost you a lot of money. Computers are less forgiving than human courts.

Quote from: LeGaulois on August 20, 2018, 08:04:58 PM

Quote from: HeRetiK on August 20, 2018, 03:18:40 PM

The challenge, when applying smart contracts, is that they should not only work -- which is already hard in its own right. Smart contracts need to be applied to use cases where they work better than what already exists.

I could agree with you if we were comparing with Ricardian contracts used by Openbazar for example but smart contracts aren't the same as written contracts. It can help to avoid a violation of an agreement, save time, automated process and reduce costs operating etc..
At the end it can be a more efficient way to deal with contracts

What I meant to say is this: From a practical perspective, a new decentralized solution needs to be significantly more convenient and / or cheaper than an established centralized one. Even if both solutions are equally viable, people will rather stick with an established solution rather than trying the new one. Put differently, smart contracts either need to find a business that they can vastly improve upon (like Uber / Lyft did with Taxis) or find an underserved niche (like Bitcoin did with global P2P payments).