That's awesome. And I think it scales for just long enough, too. I don't think hiring real, intelligent humans to make on-the-fly shopping decisions like this (think, an old-school concierge desk at hotel) costs more than the incremental value they'd bring in. If 90% of the requests can be handled automatically, then you can afford to make magic happen (and delighted customers) in the 10%.

It falls apart at a certain global scale, but I'm sure the business has pivoted twice by that point.

Nicely done.

Awesome demo! Can you update the shot while swiping left to show the timestamps?

Sure! http://brianbeck.com/images/magic-times.png

The first couple responses were pretty slow, sounds like they were getting absolutely hammered with requests. I could have done it a lot faster myself for sure, seems promising though!

That actually doesn't seem very good. If I'm craving sushi as 7pm, getting delivered at 9:15 isn't really going to do it for me.

Isn't that because he had no credit card attached to his account, and needed one before anything could progress?

In addition to the high demand they're going through, right now, I'd say it's understandable.

It was 3 minutes from CC request to CC info done.

Definitely not in the long run. But cmikec's latest update here put it in perspective – they had to limit access while this request was being handled because they were getting totally flooded.

The more trained and properly staffed they are, the faster it'll be, I don't think it's representative

One and a half hours to order something (plus delivery time - which they set would be another hour!) for 33$. Wow.

Given their explosion in requests around this time, I'm not fazed by it. Try the same request a month from now...if it's the same turnaround time, then I'd be worried that the company can't handle the volume long-term. Which would be a shame, this seems really useful.

What were the extra pieces?

Edamame, some sushi rice, and escolar sashimi which I've never tried before. (edit: some kinda soup as well)

Escolar, delicious in moderation - http://www.thekitchn.com/use-caution-when-eating-escola-6660...

This is horrible compared to my daily experience using Seamless. Why would a person craving sushi want to exchange 8 text messages and then click on a web link and get sushi 2+ hours later?

They had to click a web link because it was their first time using the service so they had to add their credit card. How else are they supposed to charge you?

Well right. But my description of the experience still stands. It doesn't seem like an improvement.

Because it's seamless when you want takeout. And instacart when you want groceries. And something else when you want office supplies. And a fourth thing to make dinner reservations...

Indeed: https://news.ycombinator.com/item?id=9094791

Because they want to feel they have their own personal servant.

And how was it? :)

Not sure about surprise additions, but isn't it bit overpriced for one person quick lunch?

Well, the additions weren't really a surprise given I told them to just meet the delivery minimum for this particular sushi place ($25).

Offtopic: your iPhone is pretty long :) or you used something like Stitch It! ?

did it manually, I'm afraid.

Get the app Tailor; it'll do it for you. :)

https://itunes.apple.com/us/app/tailor-automatic-screenshot/...

Maybe the Magic people should promote a tool, since word-of-mouth promotion can be hard without.

Thank you. Every once in awhile I need stich some mobile screenshots together and this seems perfect.

How much did they surcharge?

I just put the same items into an order on eat24.com, which is how I'd normally do this. The order came to $30.38, which is exactly $2 less than what Magic quoted. That's before you have the opportunity to add a tip.

So either the Magic surcharge was $2 and they tipped $0, or they tipped $2 and were doing some free introductory thing. ($2 off coupons for eat24 are also plentiful, so I guess Magic could have also made $4.)