#cdixon

Có một bài toán nổi tiếng trong Khoa học máy tính, đó là bài toán leo đồi. Mục tiêu của bạn là tìm ra ngọn đồi cao nhất và leo lên đấy.

Thuật toán cơ bản nhất là bạn sẽ đi theo các bước để ngày càng đạt vị trí cao. Rủi ro của nó là bạn sẽ có thể leo lên một ngọn đồi thấp hơn, chứ không phải là ngọn đồi cao nhất.

Thuật toán thứ 2 có thể giúp bạn giảm thiểu rủi ro kể trên bằng cách thực hiện một số bước đi ngẫu nhiên và giảm thiểu số lượng bước đi ngẫu nhiên theo thời gian.

Thuật toán tốt nhất là bạn sẽ thực hiện một số bước leo đồi cơ bản, sau đó  lùi lại và ước lượng ngọn đồi nào cao nhất rồi sau đó mới leo lên ngọn đồi đó.

Sức hấp dẫn của "ngọn đồi hiện tại" là rất lớn, một khi đã leo lên một ngọn đồi, rất khó để bạn có thể đi xuống để leo lên một ngọn đồi khác. Điều này có thể liên tưởng tới việc lựa chọn con đường sự nghiệp của mình.  Con người thường đánh giá cao những giá trị trước mắt hơn là những thứ lâu dài. Hiệu ứng này có vẻ còn mạnh mẽ hơn đối với những người tham vọng, nó khiến cho những người tham vọng khó từ bỏ những bước đi hiện tại hơn.

I've had Gruber's Rethinking What We Mean by 'Mobile Web' open in my browser for a couple days now, thinking over its contents.  I think that his main point is a semantic separation and broader shift between (or back to) 'mobile web' and 'Internet', which adequately addresses the point raised by Dixon.

Gruber dances around the point that I think he wants to drive at, but doesn't explicitly make.

We shouldn’t think of “the web” as only what renders in web browsers. We should think of the web as anything transmitted using HTTP and HTTPS. Apps and websites are peers, not competitors. They’re all just clients to the same services.

If we're looking at the broader trend away from 'mobile web', it's larger than just a shift back to "the web" - it's a shift to platforms.  And since I'm sure we could get very semantic about it, I'm talking specifically about Internet applications and ecosystems which are accessible through an API, whether open, closed, or somewhere in between.

Consider Facebook, the single biggest winner in Flurry’s “time spent” statistics. On the desktop, all Facebook usage takes place in browser windows. On mobile, most of it takes place in a native app. Same with YouTube and Twitter: on the desktop, they’re in the browser; on mobile, they’re in apps. It’s not about the politics of open-vs.-closed platforms. It’s simply about providing the best possible experience for users.

It's not a question of browser or app.  Taking the examples of Facebook, Twitter, and YouTube, they started in the previous era of computing, before the "Post-PC era".  They started on the web.  And when they did go mobile, they didn't start with apps.  There are a range of factors which affected their evolution, and I'm sure that mundane items like initial infrastructure setup and resource allotment was a huge part of it.

Compare instead to something that started in the Post-PC era.  We'll look at an example that correlates to the above examples - Instagram, which overlaps a lot of the same categories that Facebook, Twitter, and YouTube do.  Instagram started as an iOS-only app.  Eventually, it moved to include Android.  It took them a while longer to bake in some web (browser) capabilities.

Instagram may not be the norm in this, but the broader trend is that we are moving to cloud-based platforms, networks, and applications with robust APIs to access their functionalities on an increasing number of devices in an increasing variety of ways.  We may view websites as the norm, the fallback, the standard - but that's because we come from the pre-Post-PC era, when that was how things worked.

A web interface is no longer the authoritative version or the "true face" of a service.  For me, the face of Twitter is Tweetbot.  The face of Reddit is Alien Blue.  I consume most of my news through Feedly, not their respective homepages.  I use the Yahoo Weather app almost every day, but have never been to their website.

My main concern, though, is not this move to cloud-accessible services (which presents its own concerns), but the fact that many of them are now restricted by various platforms.  Any service that wants to build an app to put on iOS devices has to be approved for that App Store.  Mac OS X is also starting to move in this direction with the Mac App Store and Gatekeeper.  Android, of course, is less restrictive.

Yes, Apple and Google (and Amazon, and Microsoft) control their respective app stores. But the difference from Dixon’s AOL analogy is that they don’t control the internet — and they don’t control each other. Apple doesn’t want cool new apps launching Android-only, and it surely bothers Google that so many cool new apps launch iOS-first. Apple’s stance on Bitcoin hasn’t exactly kept Bitcoin from growing explosively. App Stores are walled gardens, but the apps themselves are just clients to the open web/internet[1].

[1] Note too, for example, that Twitter itself has imposed far more limitations on mobile Twitter client apps than any app store has. Or consider the spat between Google and Microsoft regarding Microsoft’s YouTube app for Windows Phone. The jostling for control is not limited to the app stores themselves.

Let's focus on Apple specifically, since they are to most people's eyes - including mine - the most restrictive.  Yes, Apple doesn't want cool new apps launching Android-only - but what's cool in their eyes isn't necessarily what's cool in the eyes of the customers.  And they clearly balance "cool" against their security/control priorities.  It obviously doesn't matter how "cool" an app or service is if it goes against their policies.

Apple's stance on Bitcoin hasn't kept it from growing explosively - but how might it have affected BTC if Apple had allowed it on their devices?  How might it have affected BTC if Apple had baked Bitcoin right into every iOS device, and accepted it on the App Store?  Sure, that's wild speculation, but what I'm saying is that we're no longer dealing with the idea of the open web.  Android is a lot closer to that ideal, but Apple is far from it.  

You might argue that people can still access the open web from a browser on iOS - but you'd know just as well as I that these relegate most of those services as second-class citizens on that device where apps rule supreme.

And yes, Twitter may have imposed far more restrictions on apps than Apple does on App Store submissions - but this is a whole different class of restriction.  These limitations don't affect a customer's access to the open web, they restrict what an app that accesses the Twitter API (and undoubtedly carries the Twitter name to it in some way) and displays information from Twitter can or cannot do.

Now this ties back to the original post that Gruber was addressing, and while Dixon spends most of his post lamenting the rise of apps and the downfall of the mobile web, the most valid and critical of his arguments lies in the last paragraph:

Most worrisome: they reject entire classes of apps without stated reasons or allowing for recourse (e.g. Apple has rejected all apps related to Bitcoin). The open architecture of the web led to an incredible era of experimentation.

In the comments from this post by @cdixon

I just reblogged Fred Wilson quoting Chris Dixon. Dixon's point was that mass produced items like books and DVDs will always be cheaper at and delivered faster from Amazon than any startup could hope to achieve.

In my response, I used Dollar Shave Club as one of the examples of companies that could beat Amazon because they are providing an item (razers and blades) that has a UPC in a way that provides more value to the customer than Amazon can presently do - and doesn't seem like they want to move into that space.

Dollar Shave Club is providing a consumable product as a service. I'm proud to call myself a member, and I haven't had to think about buying razer blades since. I also haven't had to think about the cost of razer blades since. Every month, 3 dollars is taken out of my bank account, and 5 blades arrive in the mail. I use a blade for a week (maybe 3 shaves?), and then I throw it away. I'm actually shaving more because of the service, I love it for that reason, but I love it mostly because it has solved a problem for me in a way that I never have to think about again.

ManPacks is another example of a startup providing items that their customers go through pretty quickly. If you often find yourself getting low on socks, undershirts, underwear, or condoms, than ManPacks is solving your problem.

These aren't the only items that could be turned into a subscription service so we never have to think about it again. The winner at this year's SXSW was Cerealize. They want to make sure your shelf is always stocked with a custom made cereal of your choosing.

Diapers.com was the first that I can think of to enter into this space, and I think the space come a long way since then, but it still has a long way to go, before it really starts solving a lot of problems.

Whenever I finish making a sandwich, I go into the snacks cabinet in my kitchen looking for a handful (ok, maybe two) of chips to put on my plate. Imagine the torment when I open the snacks cabinet and all I find is granola bars and despair? Chips are vacuum packed! Why isn't there a service that is sending me two bags of chips every week?

The home run in this department occurs when someone figures out how to do this legitimately for milk and eggs and bread, but those items are highly perishable and it's probably not the best place for this nascent industry to go to.

Take a moment, though, and look around your house. How many non-perishable or semi-non-perishable consumables do you see? Light bulbs? Plastic cups? Laundry soap? Shoe laces? Break-pads? Golf balls? Printer paper? Duct-tape?