A truly stupendous crowd gathered at Melbourne Business School on Tuesday night this week. Google’s chief economist, Professor Hal Varian, was in town. I was pretty excited to go along: Varian is a legend, and the issues facing Google could hardly be more pressing and contemporary.
Varian began his address not on privacy, and not on regulation. He started with machine learning. Google can now do some cool stuff with image recognition, he explained. If you visit Google’s Cloud Vision, you can upload a photo and an algorithm will tell you, with striking accuracy, what it believes the photo depicts. It was impressive to see that yes, Google can not only recognise Varian’s cat as a cat, but tell you the breed, too. The rest of the speech, whoever, was less compelling.
Adding up the numbers
“General purpose search is a tough business. Because you can only sell 6% of what you produce,” Varian said, explaining how Google makes money from selling ads on searches. Google’s search engine business spends 94% of its time providing answers to searches for which it can discern no commercial intent, Varian explained. For the rest it sells ads.
This is cute, disarming. But let’s glance for a moment at the books of Alphabet, Google’s corporate parent. In its most recent three month period, Alphabet shows revenues of $33 billion — just over $28 billion of which came from advertising. Cost of revenues was $14 billion. Margins like that don’t emerge in competitive industries.
Alphabet also spent $5 billion on R&D and another $5 billion in fines to the European Union, managing to make $3 billion in profit.
The EU fines were for the abuse of market power related to the android operating system. Asked about anti-trust actions, Varian said: “It is not illegal to be dominant. It is not illegal to be successful.”
A pleasant stroll
Varian spoke before an impressively credentialled crowd. RBA board member Professor Ian Harper was on stage with him. Productivity Commissioner Professor Stephen King was in the audience. Other professors, businesspeople and economic journalists were milling about.
And yet Varian’s speech never got into the tough stuff. This was an amble through a pleasant open meadow. It barely recognised the existence of bramble patches, let alone made any attempt to hack at them.
But brambles are getting thicker. Major technology firms are now able to exert awesome control across multiple platforms to bleed data from us. The terrible possibilities of that data are at the heart of the contemporary conversation about Silicon Valley. Yet on Tuesday, Varian wanted to look elsewhere.
Varian is here to lobby governments and win friends. Perhaps he has an impressive array of sophisticated defences for the dominance of Google, but if so he chose not to air them on Tuesday.
Privacy and arrogance
“I claim Google is the best thing that ever happened to privacy,” Varian said at one point.
He had been hit with a good question from the audience. It was about the security of data in a medical setting — posed by an audience member whose professional life had thrust them deep into the thorny issues involved.
“People ask questions to Google that they would never ask [in public],” Varian explained, smiling. “They feel they can get a more secure, better answer. These questions can be really important — how do I know if I’m pregnant, for example.”
It’s not that Varian is wrong. It’s that he is wilfully ignoring the very real issues of data privacy at the heart of the question.
Varian continued to make the case for Google that we all already grasp: it’s very useful. But he showed scarcely any indication he has begun to understand the problems it faces going forward. Signs of separation from reality are everywhere. For example, amid a discussion about venture capital, Varian made that classic Silicon Valley Freudian slip of referring to capital raised as “revenue”. Hey, it’s all money coming across the threshold, right?
A similar profundity of thought was visible in his response to a pointed question about how little tax Google pays around the world.
“It is pretty clear to us that the value to Google is created in the US because that is where the vast majority of our engineers are,” said Varian. “If I write a book and the book is sold in England, I pay taxes in the US.”
To deliver such a glib answer to an assembled crowd of powerful minds felt almost like an insult. If this complex problem could be solved by reference to the most elementary principles of tax design it would not presently be roiling the entire world.
If Google thinks deploying this man as global ambassador will smooth their path into the future, I fear for them. They may have a deeper lack of self-awareness than even I supposed.
It’s also not illegal to be immoral, unethical, a fossil fuel dinosaur, a refugee-bashing ignoramus, lead an un-examined life, treat people very badly etc etc.
Reminds me of another Titan of corporate America, may have been Jeff Bezos, who said something along the lines of ‘Competition is for losers’.
Sounds like you had an occasion with the devil’s right hand man, Jason.
Regarding your remark concerning competition, DB, you may be interested in Taplin, J. “Move Fast and Break Thinks” (a quote
attributed to Zuckerman) where the first sentence of the introduction states “I thought I was going to write a history of a culture war”.
but it is about how the emerging modern multi-nationals counduct themselves.
Hardly a surprise that Varian delivered an anodyne chat. His first year textbook of recent vintage is barely distinguishable from the 4th edition of Samuelson‘s first year text of 1962 vintage. Economists of the corporately acceptable kind from the US of A are a bit like figure skaters, elaborating on conventional displays of artistry without advancing the art in any significant fashion. But in the case of economists, failing to deal with the failings of our economic world, as we are so well aware.
We can all winge and bleat Jason but was the function a seminar where one could ask questions and if so just how well prepared were you. Policy as to Asia and China in particular could be examined along with a few recent books concerning Google itself.
Did you actually expect anything profound? Indeed, just what did you anticipate based upon previous visits (e.g. Gates) of various IT notables to Australia.
With those preliminaries why don’t we take a moment to think this event through. I expect you, along with the majority of the audience, were fascinated with the example of machine learning that Varian provided. Did it occur to you that there may be a coded message attached to the illustration? Apparently not.
I have mentioned that robots (particularly in Japan) are capable of processing insurance claims. A human vets a sample produced by the bots but by roughly 2020 the ‘bots’ will do doing it all. Ditto for translation services and similar. There is a marriage between Machine Learning and Big Data. Both are recoginised disciplines in their own right but the marriage makes for an explosive combo in governing a country
or even anticipating the next government of a country. Considerably more could be written but, with deference to AR and a few others (who prefer brevity), I’ll leave you with that thought and some basic research that you could undertake on (cough) Google.
Oh – thanks for the article; I didn’t know Varian was in town and I’m indebted to the description of the seminar.
Thanks, I think, for this exposition – I wish that I knew enough about the issues raised to be aware of why you were less than assured that all was well & we could go back to sleep.
I look forward to further elucidation.