Chris Espinosa: Fire | Hacker News Reader

Chris Espinosa: Fire

cdespinosa.posterous.com

245 pointssiglesias44 comments

One could argue that Opera Mini has been doing the very same thing for a while now, and it did not appear to cause any major uproar between the privacy proponents. This might be a typical "Ah, that's Opera, cool stuff, but who cares" attitude, but it might also indicate a profound shift in surfer's attitude towards their privacy. You press them long enough and they will grow to accept that intrusive Web surfing is a damn norm.

That's the shift I am personally really afraid of, but by the looks of it and as upsetting as it sounds, it is inevitable.

I believe Opera is not generating any uproar simply because they are acting as a "dumb pipe" in the truest sense - they're not trying to mess with the stream of data that they're entrusted (even though they're in a best possible position to do so), and so we trust them with the stream of data.

I think it would be different if they were intruding on the stream, and I imagine the backlash against Amazon if they do will be major.

But the thing is that Opera now sits on a trove of highly valuable information, and I can't imagine them not doing anything about it. As benign and non-evil as they are, they must feel pressure - be it internal, external or both - to capitalize on this "asset". Do you trust them enough to keep resisting this pressure, which is only guaranteed to rise?

All we know is that they've resisted it so far, right? I know the bar for ethics here is pretty low, but give credit where due.

One of the things they use the data for is State of the Mobile Web (http://www.opera.com/smw/), which is generated based on log files stored in a HDFS cluster and processed with Hadoop. At least it was like this a few years ago, and as far as I know they haven't changed it. Having all this information available is certainly valuable to Opera and both their current and potential customers.

Any pressure to be more evil is countered by a few things: the scandinavian culture and the developers. In general, it is in scandinavians' nature to be nice to other people, so doing evil things don't come naturally in most cases. Really. Afer all, we're socialists. Of course there are plenty of non-scandinavians at Opera, but when I was there a few years ago the corporate culture was still scandinavian. Also, the developers have traditionally had a lot to say at Opera, probably partly because of the scandinavian culture, and partly due to corporate culture instilled by the founders. They are also very vocal about their opinions, and are against doing evil things. Actually, the Opera Mini team has been especially concerned about this.

However, Opera Mini servers are located in countries around the world, and will be subject to local laws, so use and abuse of this may not always be in Opera's hands.

Confused: I read here (HN) that Opera is substituting ads on web pages, subverting every websites revenue scheme with their own. Not true?

Opera Mini is likely not causing any uproar because they're a 2nd or 3rd tier browser. They're not a 1st tier browser, and they're certainly not the default browser on a 1st tier platform.

Amazon, however, is making a play to be a 1st tier Apple-style comprehensive platform for mobile devices, media, and applications, and they're shipping "Silk" as the default browser.

Opera Mini is indeed the default browser on several phones. It's one of the few discernable ways that Opera makes any money. Granted, nothing as popular as the Fire is likely to be.

> This is the first shot in the new war for replacing the Internet with a privatized merchant data-aggregation network.

Doesn't this already exist? Google has the information from toolbar data. DoubleClick has already classified everyone on the internet based on their interests based on what sites they visit. How much better are Amazon's product recommendations going to get based on this new source of data? My guess is some, but not much.

A cheap tablet with a fast browser will bring new people into the Amazon ecosystem, and that will be worth far more than the clickstream data.

Actually I was going to say the first shot was iCloud from Apple, they just didn't extend it to browsing.

Doesn't this already exist? Google has the information from toolbar data.

... and if instead, Google built reporting of visited sites and all their contents, cookies, and your personal information into Chrome?

I don't know why this keeps coming up. It would violate their privacy policy to do anything like that, they've said they're not going to do anything like that, and if they did do something like that it would be easily detected and there would be a massive, unrecoverable backlash.

If you want to play conspiracy theory, why don't you ponder what percentage of sites have Google Analytics or +1 javascript installed?

I think you've misread the parent. nupark2 is distinguishing between what Amazon is planning and what Google do. jcampbell is saying "doesn't this already exist" and nupark2 is saying no.

Like the "unrecoverable backlash" they suffered for the App Engine pricing scandal, the net neutrality scandal, "If you have something to hide..." and their making a big stink about pulling out of China and then not actually doing it until China started hacking their servers?

Google could do anything they wanted, no matter how evil and there will still be an endless steam of fanboys defending them.

Opera Mini's been doing this for more than a year. And Opera Mini is one of the main browsers on Android phones.

Works fine with Opera for iPad. Better than Safari imho.

Opera doesn't have the other concerns raised here. They aren't vendors of anything else themselves, and have no particular reason to use their data to further other goals they may have.

The question of whether this technically "works" is uninteresting. The answer is yes.

I would worry about Opera precisely because they aren't vendors of anything else themselves. Amazon can run these servers just to increase the cachet of the tablets they sell.

Opera needs to make sure that the servers they're running are bringing in income.

Opera inserts ads via this proxy. They bought AdMarvel last year. I'd say that's a good reason to check out your data.

http://www.admarvel.com/OMAE.html

Oh dear. I'm surprised someone hasn't had a major-league lawyers-involved hissy fit about this. Thanks for the info!

Opera Mini's been doing this for more than five years. The J2ME Mini has been available worldwide since January 24, 2006.

I doubt end users will care as long as the TOS aren't absurd (e.g. Amazon has the right to sell your non-anonymized data probably wouldn't fly). People _will_ care that their Silk browser is screaming fast, though.

ISPs have been doing this for ages, maybe not as fine grained as what Amazon plans, but all traffic in general is data-mined and sold.

Is this true? If it is, why don't the RIAA and MPAA just buy the information they seem so desperate to get so they can sue you with it?

What about pageviews: If Amazon stores one copy of NYTimes and serves it to 1000 users, will NYTimes record it as a single hit?

Theoretically, Amazon could "replay" the 1000 hits back to the NYT in the background while the users are getting served the cached page. In addition, it appears that JavaScript still runs on the client (though compiled down to bytecode before reaching the device), so any JS-based analytics would still get fired off.

And Amazon could also do filter out any malware distributing websites.

Would it be wrong to call the Fire some sort of thin client? (just for the browser)

Think about this: So Amazon can learn about people's browsing habits and can infer a user's interests. Now, based on this, it can put ads in the screensaver (say with "$20 off on the Kindle fire if you let us do this"). An extreme case would be to modify the webpage with their own ads (hopefully will not happen and is unethical). But nevertheless good move, Amazon.

I think their IPs will start getting blocked if they substitute ads. However, the Kindle ads would work wonders. Just inferring the main categories in which the user makes purchases from their Amazon history and showing the top five best sellers in those categories would be sufficient, but having the web history is huge. That said, the potential abuses are also massive if they start getting hit by warrants/national security letters.

Or they could start a SilkAds service like AdWords/AdSense so content providers can display ads using Amazon.

Android Apache's license allows just that. Google was aware that was going to happen. There have already been android phones released that were locked on Bing as the search engine and many other devices with no android market. So nothing unexpected here for google.

Sensationalist.

Is it? Sounds like a strong privacy argument to me. Especially if Silk doesn't offer an opt-out. There will be wayyy too much useful data in Amazon's lap for them to simply ignore it.

Apparently it does offer an opt out. Which is nice.

Especially if Silk doesn't offer an opt-out

Which it does. "Fortunately Amazon will support an "off-cloud" mode for Silk. This lets users opt-out of the benefits of using EC2 while retaining the traditional privacy benefits of connecting directly to remote web sites." - http://nakedsecurity.sophos.com/2011/09/28/amazon-kindle-fir...

Is there a point where even "anonymized" data can be mined and recombined into something that is used to personalize and sell things better?

Absolutely. Even if you only know what people are doing/buying in aggregate, that's still very valuable. And you can still correlate between various anonymized streams of data if you know how they correlate -- i.e. 32-year-old male manga fans are more likely to buy Pokemon and Pocky than average.

Amazon already gets (and uses, vigorously) a lot of this kind of data from Amazon.com, but I'll bet they're salivating at the idea of getting to cross-compare against all the non-Amazon.com data on the same topics. If they're not, they should be.

Yep, not much logic in this article. 'All of this on Google's dime'? ZNope, it's all on Amazon's dime, server-wise, and there's very little association between this supposed web-crawling effort and Android.

I think you're misinterpreting the original paragraph, which is admittedly quite terse, even telegraphic. Read the lines that follow:

they can intermediate user click through on Google search results so Google doesn’t see the actual user behavior. Google’s whole play of promoting Android in order to aggregate user behavior patterns to sell to advertisers is completely subverted by Amazon’s intermediation.

Google's business model is to spend a metric ton of money on servers and software in order to deliver a better search engine than anyone else. They do so in order to attract users. Part of the payoff is that Google gets detailed data on user behavior -- what they search for, what they click on -- and uses that to increase their value to advertisers.

Well, now Amazon has a brand-new potential business model: They've put a man in the middle between the customer and Google. They serve the customer the search results that Google is paying to generate, and they capture all the same data about customer behavior that Google gets. To the extent that Google's tracking data has any value, Amazon has just acquired that value, all in exchange for merely proxying a search engine that Google builds and maintains "on Google's dime".

Of course, any browser vendor could do the same. But if Apple built a phone-home feature into Safari that uploaded your clickstream (including, perhaps, the decrypted end of your SSL connections?) directly to Apple headquarters they would be roasted alive -- rhetorically, anyway -- by every geek on earth. Presumably we should give the same scrutiny to this move by Amazon.

Speaking of caching, are browsers already doing all the caching that is practical on modern machines? I vaguely remember years ago downloading a special program that sped up the internet by performing more aggressive caching than the browser was capable of (basically it was just a caching proxy). Everything was noticeably quicker.

Though these days things probably feel slower because every website is loaded down with 3rd party JS from 100 different sites...

Just a guess, but that program might have had more to do with adjusting Windows' TCP/IP settings, which used some very antiquated defaults until rather recently.

Apart from the technology provider, amazon is a giant e-commerce outlet. The implications - Amazon knowing what sells, can be misused to foster its growth on e-commerce. And such things are not even related to user's privacy at all.

User might benefit for a short while, by getting their products under the same roof. But in long run it only serves to create e-commerce monopoly and drive mom-pop/small businesses out of the market, in turn eliminating the competition.

Isn't this the same data stream that Microsoft is grabbing via most IE users?

It's a big privacy violation, but not a major advantage for amazon.

What is stopping you from installing another browser on the Fire? Wouldn't that entirely bypass Silk?

Silk has an option to turn off the cloud proxy browsing mode.

Reply on news.ycombinator.com