Google’s Justification

At Google I/O 2016, CEO Sundar Pichai showed a future filled with their artificial intelligence (AI).   It is all very interesting, but I do have some questions.

How much data does Google’s AI need?

Google’s AI is backed by enormous amounts of data about us. Data that is collected from photos that are publicly posted onto the Internet, and from photos that we upload onto the Google Cloud services from our mobile phones. Data from our messages on Gmail or events on Google Calendar. Data from the GPSs on our Android phones which tell Google where we are every hour of the day. Data from our browsers which tell Google (often without us knowing it), which website we have been visiting. No other company has access to similar amounts of private information.

However, what has not been answered is how much data Google’s AI actually needs.

Can effective AI be created without too much data?

A recent article by Steve Kovach on Apple’s next generation AI system is very interesting.

Siri brings in 1 billion queries per week from users to help it get better. But VocalIQ was able to learn with just a few thousand queries and still beat Siri.

This suggests that it is possible to construct a advanced AI system with magnitudes smaller data sets; data sets that do not have to be aggregates of private user information, but can simply be collated from a relatively small number of people who were paid for the work.

Of course we need to see the results to be sure. At the same time, I find it interesting that IBM Watson was able to win Jeopardy without tapping into huge data sets like those that Google uses.

Does an intelligent assistant mean you have to give up your privacy?

Apple tries hard not to see your private data. Apple believes that your private data belongs to you only, and that you should be the only one who holds the keys. Many people have questioned this approach, based on the assumption that widespread access to private information from millions of people on the server level is the only way to create a sufficiently good AI system.

Apple’s approach does not preclude the storage and analysis of personal data, as long as it happens in a way that Apple itself cannot see. One way to do this is to handle analysis on the smartphone. This is what the NSDataDetector class in the Mac/iOS API does. It’s actually pretty neat, and Apple has a patent on it. Similar but more advanced approaches could easily be implemented in iOS, given the performance of today’s CPUs.

The question is, is this approach sufficient? Will analysing your private data on your device always be much less powerful than analysing it on the server? Furthermore, will there be a significant benefit in collating the private data from strangers to analyse your own? If so, then Google’s approach (which sacrifices your privacy) will remain significantly superior. If not, then Apple’s approach will suffice. That is, you will not necessarily have to give up your privacy to benefit from intelligent assistants.

Does Google need the data for other purposes?

Let us assume that there existed a technology that allowed you to create an effective intelligent assistant, but that did not require that you give up your personal data. Would Google still collect your personal data?

The answer to this question is quite obviously YES. Google ultimately needs your private information for ad targeting purposes.

Could Google be using the big data/AI argument to justify the collection of huge amounts of private data for ad targeting purposes? I think, very possibly YES.

Is India Really The Next Big Opportunity In Tech

A lot has been made about how important India is to tech, and what a big opportunity the 1.2 billion population is.

While that maybe true, I think it is also important to contemplate the possibility that this may not actually be the case; that despite its huge population, India may not yet be an attractive investment.

Rakuten Ventures had this to say at Tech in Asia Singapore 2016.

While India has a population of 1.2 billion, there are only about 40 million to 50 million people who actually have “real” smartphones – and not those weird Android permutations – and who are at least in the middle class, earning about US$10,000 a year.

If you’re looking at ecommerce alone, you’re talking about a demographic that has been shrunk from 1.2 billion to 40 million or 50 million. That’s basically the addressable market […] For us, when we look at a market, we ask ourselves: ‘Can we get in at the price point we want? Can we actually see a lot of these platforms accrue the value that they want?’ We don’t see that yet.

Objectively, the International Monetary Fund puts India’s GDP per capita for 2015 at 6,162 international dollars, which is less than half of China’s at 14,107. While obviously growing quite quickly, it isn’t necessarily growing that much faster compared to other countries with similar absolute levels. Although macro data obviously does not tell the full story, it does support Rakuten Venture’s view to a certain extent.

If we do accept Rakuten Venture’s view that ‘while India has a population of 1.2 billion, there are only about 40 million to 50 million people who actually have “real” smartphones’, then it does seem like other markets which aren’t receiving as much hype, might actually hold larger potential.

I think this is something worth thinking about. It might be more important to look at metrics of usage like Web usage or Twitter usage to understand how many people do have “real smartphones”, or use the ones they have as such.

Why Do Companies Outsource?

Last week, news broke out of Apple’s “McQueen” project; a plan to move Apple’s iCloud data away from Amazon’s AWS, Microsoft’s Azure, etc. and into its own data centres.

This isn’t really a surprise or an insidious plan by Apple to damage its competitors. It’s obvious if you think about what the benefit of outsourcing is for any company.

Companies outsource if;
1. The technology is not a core strength of the company.
2. The technology will not be a key differentiator going forward.
3. There are cost benefits (mostly due to scale) of outsourcing.

In the case of Apple, cloud infrastructure was not a core strength of the company so it made sense to outsource at the onset. However, it became clear that the cloud would be a key differentiator going forward. Additionally, the scale of Apple’s cloud operations became huge, and hence the cost benefits of outsourcing became negative.

The only thing that was surprising to me was the fact that Apple was outsourcing at all. I would have thought that Apple had had everything in house years ago.

Peak Google?

When I look back at 2015, I am reminded that we started out with the idea of “Peak Google”. It was not only Ben Thompson, but other analysts too chimed in with this theme.

However, this hasn’t happened. At least not yet.Google’s revenue and profits are as strong as ever, and the end of search volume is nowhere in sight.

Google s first quarter under Alphabet continues strong growth in revenue and profit The Verge

The Verge

So were Ben Thompson and other analysts wrong?

Well the question really isn’t who is wrong or right. The problem is that the analysts were not providing enough information to be proved wrong or right in the first place.

In the case of Ben Thompson, he provides a disclaimer;

Still, I hope the subtle point I’m trying to make is clear: I think Google is quite safe when it comes to search, and that they will be a very profitable company for the foreseeable future. I just suspect we will all think differently about that dominance when it’s a small percentage of total digital advertising, just as we thought differently about IBM’s dominance of mainframes in the age of the PC, or Microsoft’s dominance of PCs in the age of the smartphone.

He makes it clear that one will be able to neither validate nor disprove his statement based on Google’s finances, and that he is talking about the fuzzy term “dominance”. And as expected, no clear definition of how “dominance” should be measured is provided.

And that’s what most shrewd analysts would do to cover their behinds, so that’s OK.

So how should we really think about the situation?

Companies don’t live forever

IBM and Microsoft seemed utterly invincible in their heyday and yet, they have lost a lot of their power now. There is a lot of discussion about how long companies will live, and right now, the average lifespan of a company in the S&P 500 is a mere 15 years. If we looked at tech alone, the lifespan would probably be much shorter. Hence even if you were to randomly predict that a company would decline in the next 5-10 years, you would still probably be right.

It’s harder to predict how companies will survive

Given that companies will tend to die in a short span anyway, discussing how they will die isn’t very constructive. It doesn’t add much value to the prediction that companies will die pretty soon anyway. It’s much harder, and much more relevant to discuss how they could survive. That is, one should presume that a company’s present business will surely decline. On top of this, one should strive to seek out what it’s next business could be. Regarding Google, calling Peak Google based on a prediction of the demise of its search advertising business doesn’t add much. It will happen, and no matter how hard you discuss, you will not get an accurate prediction of when. Instead, one should look at what Google’s next business could be, and if they are making significant progress on that. One should look at how Google could survive when (and it will surely come) the search ad business collapses.

When is almost impossible to predict

Let’s look at some very nice charts from, showing revenue/profit trends for Apple, Microsoft and Google.



What you can see is that Apple posted record revenues up until 1995. Although they were starting to have issues with profits since 1993, there was little to suggest that they would experience a total collapse in 1996. Just like very few economists could see the Global Financial Crisis coming, complex systems often show catastrophic behaviour. That’s why predicting the when is so hard.


Proof of “Peak Google” has not come this year, and in my opinion, this shows more than anything that predicting the when is hard, almost impossible. Although Ben Thompson was shrewd enough to not make any predictions, some analysts were not. My advice would be to stop trying to predict “Peak Google” or peak anything for that matter, treat the decline of any business as a given, and focus on what that company may have in store for the future.

Google’s Response to Accidental Clicks

Google recently described on their blog, some of the ways that they have been trying to rectify the “accidental click” problem, which some 3rd party studies have estimated to amount to 50% of ad clicks.

This is obviously a welcome move from the advertisers which have ultimately pay for these accidental clicks, but it also raises the question of a) is this enough from the advertiser’s point of view, and b) how much will this impact Google’s revenues.

To understand this, we have to carefully look a the constraints that are carefully worded into this article.

  1. This only applies to AdSense: This article carefully mentions that all remedies and conversion rate improvements apply to display ads only. This does not apply to AdWords ads, the ads that show up on Google search result pages. Now the vast majority of Google’s revenue comes from AdWords. Google’s AdSense business is much smaller and hardly growing in revenue. Therefore we can conclude that Google’s remedies only apply to a small portion of their overall business, and it is a business that is decreasing in importance.
  2. AdSense text ads are not affected: AdSense is comprised of a image ads and text ads. It is unclear which of these is a larger business. However, since banner blindness is well documented and users have learnt to not click on anything that looks like a banner ad, it is possible that text ads actually attract the bulk of clicks for AdSense. This article again carefully makes it clear that the enhancements only apply to the image-containing banner ads.

It appears that Google has restricted the improvements for click quality to the segments of ads which will impact Google’s revenue the least. This is completely understandable given that Google, as any other company, wants to increase revenue and profit, but it is also important that the current efforts will only fix a minor segment of accidental clicks. This means that going forward, there will still remain a significant number of accidental clicks, and that Google may again be pressured by advertisers to reduce these in the not-too-distant future. Although not necessarily a likely possibility, it would be particularly damaging if Google was forced to modify its search ad format (AdWords) to prevent accidental clicks from users who did not recognise that the links were actually advertisements.

Google Now And The Priorities At Google

Mark Bergen (@mhbergen) wrote an interesting story on Google on how most of the original Google Now team has left the company, mainly due to a lack of prioritisation.

Although we need further reports to confirm this and to get a picture of what the consequences may be, I found this very interesting because it aligns with some predictions I had made previously.

  1. Predicting Android’s Change Of Direction: Thoughts from Andy Rubin’s Demotion (Apr. 2013, in Japanese)
  2. Who Is To Blame For Samsung’s Bad Fortune? (Nov. 2014)
  3. Android No Longer Competes With iOS (May. 2015)

Mark Bergen’s article in essence says;
1. Google Now was born within Android.
2. Larry Page heavily prioritised it, but then became too busy with moonshots.
3. Sundar Pichai de-prioritised Google Now as an independent intelligent assistant, and moved it from Android to Search.

My argument was that whereas Andy Rubin wanted Android to be the best mobile operating system in and of itself, Larry Page (and consequently Sundar Pichai) thought of it as a gateway to their Search cash cow. Hence priorities are determined based on potential contribution to the Search business, and not on the merits of Android itself. It seems to match Mark Bergen’s discussion, with the exception of heavy support from Larry Page, which I didn’t expect because of how he removed Andy Rubin.

At this point, it is difficult to say whether Google Now should be a priority for Google or not. I suspect that Android is more focused on emerging countries than developed ones, and if so, then de-prioritising Now makes sense. It’s also unclear whether Google Now will really provide good advertising opportunities. Despite theoretically being attractive for ads, I do not know of any mechanism provided by Google for purchasing ad space, and it is possible that it isn’t really very good. Sundar Pichai may be fully aware of the consequences of what he is doing, and that he is convinced that it is better for Google long-term. Google’s multiple projects were always in conflict in one way or another, and it’s possible that Sundar is simply clearing the mess up.

What I do worry about is the fate of Android OEMs. Android, if it is to compete with iOS at all, needs to have features that are unique to it. Tight integration of Google Now seemed to be a great opportunity. Without it, Android will struggle even more in the high-end. As a result, profits for Android OEMs will get even worse.

Chromebooks and iPads in U.S. Schools

A recent blog on the New York Times described how Chromebooks are gaining in the U.S. education market (K-12). I have wrote quite a lot about Chromebooks on this block, and this article tells us that progress has been made on the part of Google. Of course, the market that is described in this article is quite small with only 13.2 million units annually, in comparison to over 300 million PC units (excluding tablets) sold worldwide, and as far as I know, Chromebook’s success in K-12 education has not expanded to other markets (including international). Nonetheless, this is good news for Google.

The comments section is also very good, with some specific examples of why certain schools decided to purchase Chromebooks instead of iPads or Windows PCs.

My broad-view understanding is that Chromebooks are serving pre-existing needs that are mainly administrative by nature, and are best understood as sustaining or efficiency innovations. The blog post and the associated comments strengthen my view.

The real problem as I see, it is that there is a huge amount of potential in bringing technology to the classroom, but there is still too little investment in terms of hardware, software, curriculum and teaching skills. Sustaining and efficiency innovations won’t take us there. They don’t provide administration with good reasons to invest more; they only give us reasons to spend less. We need empowering innovations (such as which the iPad promises to bring) for that.

Data shown in this article

名称未設定 numbers

名称未設定 numbers

Does Google Now Really Need All Your Data?

On June 17th, 2015, I discussed Google Now, Apple’s Proactive and their respective approaches to obtaining your data. My main thesis was that as far as I could tell, the predictive assistant functions that most people come up with seem to be perfectly possible even with Apple’s approach. In conclusion, I said the following;

Although I certainly need to dig into this in a bit more detail, I am skeptical that invading your privacy is essential for providing a better personal assistant service. I would welcome any examples where the personal assistant must absolutely send all knowledge of everything about you to servers in the cloud to be analysed.

Phil Schiller made some comments in an interview with John Gruber which indicate that Apple thinks the same;

If ever there’s a modern definition of a Faustian bargain, this is it, right? Which is, that if you want to get the features, give us all this information about your life that you’d really rather not.

And we’ve believed for a very long time that that doesn’t have to be the case. And so we’ve built systems and processes all around the idea that, in order to help users, you can do things that are surprising and delightful and magical—but we don’t know your data.

So now the fight is on. On one side, we have Google which suggests that they need all your data to provide you with wonderful predictive assistant services (actually I haven’t seen anybody from Google actually say that, but it seems to be what the pundits are collectively thinking). On the other, we have Apple which believes that they don’t need all this data. Essentially, Apple is saying that there is an upper limit to what they need to know, and that limit is actually very low. It would be interesting to watch how good Apple’s Proactive turns out to be versus Google Now.

Of course, Google and Apple have very different business models and hence the business requirements for their predictive assistants are different. Google’s business is advertising so they need enough information to target you with ads. That might require much more personal information than what is required for just providing a predictive service. For example, if you’re just thinking of going to town, all that a predictive assistant needs to do is to give you the directions, time, lunch suggestions, etc. This can all be done anonymously. However, if Google needs to send you ads on behalf of advertisers, then the more they know about you, the higher they can sell you to advertisers. Advertisers love targeting information and would pay for a detailed profile of whom they are targeting. For example, they would love to know if you are married or have kids. They would like to know what kind of food you eat regularly. They may even like to know if you are fit or overweight. None of this is relevant for the predictive assistant task itself, but it is relevant for the ads given through the assistant. And in Google’s case, these ads are what financially support the service. Essentially, Google Now needs more personal information because they need to finance the service through ads. Even if a predictive assistant didn’t need this data to give users advice, Google would still need to collect data on behalf of the advertisers in order to sustain the service financially.

Google’s Slow Adoption Of Mobile First

I have recently commented (1, 2) on how I suspect groupthink to be creeping in at Google. Another area where I think we can find evidence of this is Google’s slow adoption of “mobile first”.

While my assertion that Google has been slow to “mobile first” may seem very odd given that Google developed the most popular smartphone operating system, I see that as a whole, Google has maintained a PC-centric view. That is to say, although the Android group is very mobile focused, it seems that not every division shares this view and this includes many of the strategically important ones. In many key areas, Google has failed to embrace the priorities and mentalities that are required to succeed in mobile.

One very evident example is the lack of focus on power efficiency. In a recent post by Google software engineer, Peter Kasting, Google mentions its commitment to make its Chrome browser more energy efficient. Kasting in particular mentions Chrome’s power efficiency on the Mac, but that is only because Safari on the Mac is extremely well optimised for battery life. Google’s lack of attention to power efficiency has also been evident on Windows.

This is very much in contrast to Apple which has been making power consumption a huge priority ever since 2005 when Steve Jobs announced their switch to Intel processors. Recent MacOS X and Safari updates have also frequently focused on improvements in energy efficiency. It is very obvious that the synergy with iOS development and the iPod before that has been a major factor in driving the Mac to power efficiency. In this sense, one can argue that the whole company, including the Mac division, shares the same mentality towards energy efficient design and optimizations. This apparently has not been the case with Google.

I suspect that inside Google, the groups that develop software for servers, the groups that develop for desktop application or web applications, and the groups that develop for Android are not talking to each other. The power-saving techniques developed for Android are not being transferred to their desktop applications. In fact, one observation that stands out in Android app development is how long it took for Chrome to be ported to Android in a version that did not suck (discussed here for example, among numerous other places, and also in my personal experience), which clearly shows that they had trouble developing Chrome for less powerful devices.

If this truly is the case, then Google can only be as strong as the sum of its parts. Synergy between Google’s divisions can only happen at the product level (lock-in etc.) and not at the development level.

Maybe given that Google’s real strength lies in its cloud business, this doesn’t really matter. If Google wishes to keep all the intelligence in the cloud and to keep computing on the client-side to a minimum, then all they have to do is keep improving their browser (which interestingly they haven’t been very successful at on mobile). They don’t really have to invest in creating the very best client software because their cloud services can make up for it (or at least that’s their plan).

If however the tides change and for some reason, computing on the client side becomes important (maybe due to increased awareness of privacy issues for example), I would be rather pessimistic of the transfer of knowledge between Google’s AI experts and their mobile development experts. I don’t think they would be able to create very good client software (maybe Chrome for Android quality).

Of course, in Apple’s case, it is the other way around right now. Apple is strongest when the device teams and the software and services teams work together and create synergies. It all depends on how technology and customer expectations evolve.

Did Google Expect Apple to Focus This Much On Privacy?

In relation to my previous post where I raised the possibility that groupthink might be happening within Google’s top management, I would like to bring attention to the possibility that Google might have totally failed to predict Apple’s very strong focus on privacy.

In the WWDC keynote, Craig Federighi repeatedly mentioned Apple’s focus on privacy. Proactive Assistant has privacy, Photos has privacy, and the News app has privacy. Furthermore, in addition to focusing on privacy on its own apps, Apple has build in what one could call an adblocker API into Safari so that users can install adblocker extensions that would also protect them from the user tracking activities that ad networks often conduct behind the scenes. Apple has also added many capabilities to the Spotlight feature so that you don’t have to visit Google and help them create a ad targeting profile of you nearly as often.

Although it is still unclear whether for example Proactive Assistant will work as well as Google Now, or whether normal users will install and adblocker extension, it is evident that the features that Apple introduced are potentially very damaging to Google if users embrace them. It would result in significantly reduced traffic to Google search, which is Google’s sole monetisation engine.

My question is, are there any clues that Google predicted these features from Apple? Do they have a plan already in place to counter any negative publicity on privacy (other than the heavy lobbing that they are doing in Washington), especially outside of the US? What was their contingency plan in the event that Apple would embrace adblockers? What Apple introduced was not completely unexpected given their previous comments on privacy and Tim Cook’s emphasis on human rights, and if Google’s management had been paying attention, they could easily have predicted them. If we actually end up seeing clues that Google completely missed these cues, then it reinforces my argument that Google management might be in groupthink.

I will be keeping an close eye on how Google executives react.