[babaeth]

1. To accept DAI you simply need a phone and a dedicated app that would generate the payment QR code. No new device needed. (67% of payments in convenience stores in China are being made with QR codes: https://technode.com/2018/03/16/qr-codes-nfc-china/)

2. There is definitely a chicken and the egg problem as there wouldn't be as many people ready to pay with DAI at first but it has never stopped good projects to take off. In my opinion the main issue (once the tech is ready with Plasma running on the main net) is to get merchants to accept this new payment method which should possible since it's literally free to try.

3. Paying with Bitcoin is completely different since most of the people who own some Bitcoins don't want to spend it as it's mostly used as store of value/speculative asset.

[richardknop]

1. What are the costs of your solution? Who's running the nodes, how much does it cost? Gas? Give me some numbers please.

2. Scaling. Current state - it doesn't scale. My own opinion is it will never scale, centralised systems will always be much more performant.

3. Not really. The only difference is Bitcoin is very volatile and most people are expecting it will rise in value so they are not spending, you are right. But as a payment system integration wise it would be the same as your proposed solution, some mobile app and scanning of QR codes.

Finally, give me a single reason why it would be better than current system (as I believe the answer to 1. it's not cheaper, it's actually more expensive as centralised system). And that was the only reason you have given. Is there anything else than theoretically lower fees (which I don't believe would be the case).

[babaeth]

So it seems that our main disagreement is about the cost of the implementation of such system.

While I agree that the current state of Blockchain and Ethereum doesn't allow this I also believe that a part of the scaling issue will soon be solve with Plasma, you can learn more about it here (https://plasma.io/plasma.pdf). In this case the transaction fees and processing time are drastically reduced and Plasma could support (in theory) 1,000,000 transactions / seconds. The sender will pay for the tx fees (probably around half a cent) and the payment will be delivered almost instantly.

So "the single reason why it would be better than current system", since you asked for it is really because transactions operated on an efficient Blockchain would be less expensive than credit-card payments.

[richardknop]

In theory what you are describing sounds appealing. One more issue I thought about is in a case this system ever became a thing, Visa/Mastercard could just lower their fees. Pretty sure they are still profitable even with much lower fees so there's a chance they could still outperform this plasma system since their centralised system will be more efficient and will allow them to have even lower fees. It would be good to have a competition to force Visa/MC to lower their fees though. And other payment processors. They are getting away with insane profits right now.

Also I don't think it's possible to rollout a system like this in China due to the government there. They would not allow a a system they don't control to be used. So it would have to be in some relatively free country where you would have to implement this.