Pressing a button is milliseconds, and can be done as a complete afterthought, while still doing your current task. Taking out your phone, navigating to a website and placing an order means you have to stop what you are doing, and do something else.

When the time to complete tasks changes by an order of magnitude - even going from seconds down to milliseconds - our behavior tends to change. We easily accept this notion with developer related tasks such as compiling and testing. I don't see a difference here.

Imagine if developers only built and tested once or twice a month.

Now I come out with a new product where the entire pitch is, that build-and-compile you do once or twice a month now takes one second instead of ten! Oh, and it only works for one project. If you work on multiple projects, buy multiple products.

Would this be worthwhile? I don't see it.

Developer tasks like compiling and testing benefit from small efficiency gains because we do them constantly. Or the efficiency gains allow us to move from a model where we do them rarely to a model where we do them constantly. Neither one applies here. You won't buy detergent any more frequently with this button, and it's not something you do often.

I won't buy detergent more frequently, but because it's easier, I'm more likely to be in a situation where I actually have it where I need it. If I could buy either at the moment I realize I'm about to run out, I would. Instead, I have to remember when I'm out at the grocery store, and I often forget, and then end up having to make a special trip the next day just to pick up that one thing.

It's not just about time saved. I tend to forget things I don't do often, so I automate them, even though setting up the automation might take more time than it will save me. As soon as the automation is set up, I no longer need to worry about it, and next time I can just push the button.

Phones are annoying, and the UI of most apps is user-hostile. Here are the equivalent steps you have to do to accomplish the same thing as this button does:

    - Wipe your hands from moisture/whatever you had on them.
    - Find your phone in your pocket.
    - Orient it.
    - Unlock it.
    - Exit whatever app you left running in the foreground.
    - Open the app drawer.
    - Find the ordering app.
    - Click through several layers of UI to get to the right product.

And only now you're in the same position as after just pressing a button. This is a significant cognitive load. When I realize I need to order something, I want to order it, not navigate through countless steps. Having to do more than two actions is too much.

Exactly, a cellphone app can get the job done but it is a great cognitive deviation from whatever task you had at hand at the moment you realize you ran out of something(1). So amazon's dash button is not an innovation on getting something done, but get it done faster and more conveniently, and that alone makes sense for an enterprise whose profit depends greatly from understanding consumer behavior. So I agree with TeMPOraL, it is not ridiculous at all.

And there's a bigger picture outside of comparing the convenience between using this little button as opposed to mobile apps, but that's off topic for this branch of the conversation, if you're interested IMHO & my brief reasoning, read this other brach[2] (also because DRY).

(1): Optimally you realize you are about to run out of something before it happens. [2]: https://news.ycombinator.com/item?id=9299505

I, for one, don't have my phone handy when I'm doing laundry in my pajamas, or while shaving in my bathrobe.

You have probably never worn women's jeans. They have ridiculously small pockets, barely big enough to squeeze my S5 into.