Category Archives: Society

Innovation vs Dogma

Image August 13, 2017 harrouet

« I disapprove of what you say, but I will defend to the death your right to say it ». This maxim is wrongly-attributed to Voltaire. And it might just be what Google needs. Indeed, the current Google situation is a typical situation of Innovation vs Dogma.

Back to facts

In the course of the summer 2017, a memo by Google employee James Damore leaks out, stating that (in short) women are biologically not meant to be engineers…
James Damore gets fired
Google CEO Sundai Pichar needs to cancel a company-wide interactive session about gender discrimination, after several employees complained that they would not be able to express their views without retaliation from fellow employees.

Quick disclaimer

Just to set things straight: this blog is not about opinions. However, we think it relevant to point to scientific research which prove James Damore wrong. Also, we can’t emphasize enough how women greatly contributed to science and computing. Marie Curie and Dame Stephanie Shirley are enough to ridicule the whole career achievements of James Damore.

Google wanted so much to be inclusive that it got exclusive

But back to Voltaire: an innovative company must allow anyone to feel comfortable being who they are, regardless of political opinions. The Google situation is a meaningful management lesson: Google wanted so much to be inclusive that it got exclusive.

It is hard for a company that excels so much in not paying taxes to pretend to work for the good of humanity. Kittens don’t replace schools, hospitals or roads — and these are paid by taxes. “Don’t do evil” was Google’s mantra — which Steve Jobs rightfully called “Bullshit”.

Nevertheless, from a very cynical business perspective, Google needs inclusive values, because:

It is good for its image. When your whole business is based on spying on people, and that the NSA leverages that to get info on millions of citizens, you needs to work on how you are perceived by society.
It promotes a culture of performance. Not matter who you are, we only value what you produce.
It increases resources in the long term. Exclude women and your world is 50% smaller. Add to this non-caucasians, homosexuals, and republicans, and you’ll be quickly in shortage of workforce and customers.

The problem is that embracing inclusiveness may be dogmatic, when it means that alternative opinions are excluded. For example, some moderate conservatives start to feel uneasy working at Google. Being inclusive to this extent is being exclusive.

Instead, Google needs to worry about its innovation culture. It needs to make sure that inclusiveness applies to its employee’s careers. If James Damore was not valuing any woman engineer, maybe his job performance would show it. This is when communication does not replace action. James Damore may have been fired to quickly close the topic, it may actually have opened a pandora box and create a deadly fight internally to the company.

Democracy and capitalism

In order not to do evil, Google needs to learn to actually be inclusive: by paying taxes, by participating in society at large without taking part in the debate, and by enforcing performance metrics that are affected by inclusiveness. In short, accept not to be in control of everything, as long as it allowed the debate to take place. Let’s call that … democracy?

Book references, Skills, Society

[Book review] Fooled by Randomness

January 26, 2016 harrouet

I don’t usually read books about “how to see life differently”, but I liked the idea that an original-minded trader (by background) wrote about the role of randomness in life. “Fooled by randomness” is just that: a (sometimes cynical) view of our lives from the eyes of a statistician.

The Author

Nassim Nicholas Taleb is most famous for his previous work about unlikely events — “The Black Swan”. A former trader, he nows teaches at NYU and feeds trolls on Twitter. Of Mediterranean culture, seemingly extensive thereof, he excels in seeing things in their wholesomeness, not just events by events. This was the philosophy of the Black Swan, and it is still present in this book.

The Takeaways

The book is a constant reminder to check real facts against their probability of occurence, before jumping to conclusion about their cause. Several phenomenons are at work in building well-accepted-but-false theories out of observation:

Randomness: while people might use complicated theories to explain visible successes, they might just be the result of the number of agents at play. Then, there is always a slight probability that you are observing a small-but-visible sample of that large number set. Imagine that 10 000 traders play the casino on foreign-exchange rates. Out of this quantity, a marginal number will almost surely come up with impressive results. Their big bonuses and acclaim will make them more visible that the losers. Add cognitive biais to this, and you get a false theory that might seem to explain their success.
Cognitive biais: several biaises are at play, but few of them seem of particular importance:
- survivor biais: we usually don’t talk about the losers, just about the winners. When you see a traders who “made it”, you may not see all those who got fired because of their poor performance;
- risk/loss aversion: we may make the wrong choices by fear of losing out. As a trader specialised on unlikely events with big impact (the “black swans”), the author shares the pressure he gets while investors see minor loses on his portfolio while others (the “random fools”) are in the green. This may be the case most of the time. Until that time when he makes up in a few weeks several times what the “random fools” made in the previous years. The example that comes to mind is the one where traders make $5M every year during several years, enjoy promotions and big bonuses, until their theories turn out to malfunction and they lose $30-50M in a few weeks. Over time they clearly lost their investors’ money in hot-air theories, but investors do prefer this way, rather than to accept losing little during several years until a disproportionate reversal of fortune. Think of it next time your friend shows off how much s/he made on the stock market;
- induction: we tend to expect the future to be a continuation of the past. We all heard that “real-estate market will never go down”. Until it does. This may be intuitively a particular case of the fact that our brains like linearity — while many things are non-linear;
- environmental biais: we have greater chances to live in environments where people are similar to us. Therefore, our house may never seem big enough (“the neighbour has a new car”…), our fortunes never seem to be as good or bad as they actually are.
Laziness : the medias are particularly held responsible to spread biaises. The striking example is how they have to explain the direction of the stock market with whatever news of the day. They can’t justify their existence by just saying that we are observing noise. This makes reading news much less important. Obviously this applies to many managers in larges companies, or politicians, etc. trying to take credit for random events.

Personal opinion

If you have a background in statistics, and have read recent books about cognitive biases, you may not feel like you will learn a lot in this book. Let’s just say that it is a kind reminder. I found the writing style not particularly pleasant, but I will nevertheless keep this book as a reference when confronted to random fools. It is not often that these things are dealt with altogether in one convenient place.

The book is available here.

Innovation, Law, New business, Society

On self-driving cars and side mirrors

December 1, 2015 harrouet

According to car manufacturers or large software companies, the self-driving cars are around the corner. What was still a science-fiction research field 15 years ago — my university had such project, led by AI guys walking bare feet and called “the crazy guys” — is now meeting massive capital influx, mindshare from top scientists and consumer momentum. Such cars are being tested live in real roads in California, and now even Switzerland.

But as always in innovation, there are adoption barriers, and they do exist for the self-driving car. Assuming the technology will ultimately work, the main remaining barrier is probably the law: in most countries, you are supposed by law to remain in control of your vehicle. This is a key principle for insurance companies to assess responsibilities in a crash. Easy to overcome, you would think ? not quite. For even technology-mature solutions that over-perform established ones may not make it because of regulations.

Side mirrors are an artefact from the past

As an example, every car has a artefact from the past which should no longer be here: side mirrors. If they had to be invented today, side mirrors would probably not make it to market.

Imagine that side mirrors don’t exist. Drivers need to turn their heads to look at what is coming up behind. A marketing manager identifies it as a customer pain, and asks engineers to find a solution. One comes up with the side mirror — but it comes with its own collection of issues:

it will decrease the performance of the car, as it increases the air-penetration ratio. So, the gas consumption will have to increase, say, by 0.5L per 100km.
it will impact manufacturing, which will find it more difficult to hold a mirror on the side of the car, rather than having a flat surface.
Mirrors will break regularly, generating other sets of customer pains.
oh, and by the way, there will still be blind spots, so you will have to put a disclaimer about how poorly it is addressing the issue.

Obviously the marketing manager will ask the engineering team to reconsider. Needless to say, small webcams would be much more efficient to solve this issue — as is the case for going backward in some cars. Some concept cars already support this solution.

So, why are side mirrors still very present in cars ? Because they are compulsory. The law says that cars must have side mirrors. And law has the single biggest kind of inertia that can be. From such regulation, e.g. insurance companies built their processes, defined ways of dealing with others and, ultimately, to cover the risks.

Side mirrors show us what could happen to self-driving cars

The side mirror gives us a glimpse of what it means to have a technology-mature solution that don’t make it to market because of the law. And this could very well be the case with self-driving cars. Actually, you can read more about what happens with side mirrors and regulations at this page.

With that said, we are not even accounting for the difficulties for regulators to certify self-driving algorithms (imagine Apple with a great algorithm, Google with another one, but when put together the cars end up making incompatible choices that end up in crashes). And once you find out how to certify them, which will probably a lengthy process, how do you deal with software updates every once in a while to fix a security issue?

So, it won’t be before long that you see self-driving cars : regulation need to change for this to happen and this typically takes a lot of time. Actually, we don’t even know how to even regulate self-driving cars. This requires to make choices about people’s safety, and this is the last thing that regulators and politicians like to do.

Society, Uberization

My problem with Uber — and its detractors.

July 20, 2015 harrouet

It might not be politically correct to say this in the innovation community, but I have two problems with Uber:

it is no longer an innovative company, but it became an unethical corporation with deep pockets.
its detractors do nothing to be constructive against Uber, just opposing the old, established world against the digital tide-wave.

Uber is no longer driving innovation — but chose to adopt predatory behavior.

Now let me get more precise: Uber is no longer innovative. I mean, really. Truth be told, it takes just around $2M of founding to start a Uber look-alike. Granted, Uber set the way, but most cities now have many Uber competitors beyond taxis. So no, it is not rocket-science anymore.

With that said, how can Uber be the most valued (i.e. expensive) unicorn ? It raised a total of $5.9 Billions, the last round reportedly valuing the company at over $40B. That’s a lot of money, for a company that seems to lose as much as it has revenues. Uber’s promise may be about self-driving cars and (probably) fully-robotized last-mile parcel logistics. But nothing to be happening for the next 10 years or so, given regulatory and social barriers to launch such services. Meanwhile, the company’s main asset is… its brand: therefore it does everything to install it, including by extensive use of unethical, and sometimes illegal, business techniques:

In some countries, where UberPoP was declared illegal, it keeps the service running no matter what, even paying for the fines of drivers being caught (which is also illegal).
In other countries, it traps the drivers of competing services to either recruit them at higher rates (remember Uber is non profitable already?), or just to generate a no-show.
It goes as far as making battles personal, where mayors bar the service from their city.

Examples like this are plenty. There is even a wikipedia page about it. So let me repeat it: Uber is not about innovation anymore, but about predatory behavior to preempt markets at a loss. I would not want to be a shareholder, unless relying on the greater fool theory.

The debate is about a new way to organize labor.

Now let’s not be mistaken, even blinded by all the reproaches that Uber gets or deserves. For it has one great merit, which is to open a new, necessary debate about how the digital revolution changes our labor organization.

In most developed economies, the industrial revolution introduced a new relationship between the those running the business (“the employer“), and the person actually getting things done (“the employee“). This relationship is generally established with a work contract and, depending on the country, union agreements and/or labor law. It materializes with institutions running this system.

With Uber, as with many examples of the collaborative economy (e.g. AirBnB…), the market built towards filling a need is shifted out of social institutions’ control. There is a price effect to that phenomenon: social institutions have a cost and, by short-cutting those, these services acquire a price advantage which is not about creating value, but about refusing to pay for social institutions. Beyond Uber’s great UX, there is social arbitrage at play.

As a result, it is logical that Uber, arguably being the most aggressive new entrant, is becoming a part of the next US presidential campaign: can society accept a company which provides a great service, in the consumer’s interest, but shows no single sign of Social Responsibility? Indeed, how can companies pretend to do good, when they all optimize their tax scheme so that they don’t pay for schools or hospitals? Answering this question is far beyond the scope of this blog, but this leads to the second problem stated in this article: Uber’s detractors are no more constructive.

Uber and its detractors need to co-design the new world

Detractors of Uber are generally those directly threatened by its activities: taxi drivers (in many countries a rent situation), but also social institution financed by the existing work contract momentum — health insurances, retirement administrations, etc. All these have financial liabilities only viable if workers pay their taxes. Governments sometimes clearly sided with them.

Problem is, these detractors have sometimes fought against Uber to defend the existing momentum. This is doomed to failure, as institutions never win against mass adoption by the people. This is all the more surprising, that in the past institutions have evolved against new services:

Copyright holders found a way to deal with Youtube, making the service fully legal (remember when this was not the case?) in most western countries.
AirBnB agreed to pay the same taxes as hotels in France, and no one pretends that it should be illegal.
BlablaCar, also in the transportation industry, made sure that its drivers could not turn a profit with the service.

Now, maybe Uber likes this situation, where paying fines is seen as a cost of doing business, and a convenient barrier-to-entry for its competitors. But it is time to sit at the negotiation table and find a sane and sustainable way to conduct business. Unless Uber can introduce driverless cars fast enough.

HOPSYS – Your future is digital