In 2011, the famous investor Marc Andreessen had published an article that had made great noise: Why software is eating the world . If this article had generated many debates, 8 years later, it is clear that he was right, especially with the advent of artificial intelligences. Of course, the software will not replace the hardware components, but can optimize them dramatically , thus concentrating all the added value. While some have incorporated this fact (the valuation of Uber is 40 times higher than that of Hertz), the public remains very dubitative and suspicious vis-à-vis these software and algorithms become powerful. Like what, the emotional, irrational dimension of digital transformation is still largely undervalued.
Upgrades that improve performance and extend the life of hardware
Last week, Google updated the camera app on its smartphones. An update of the software that drives the photo sensor, but also manages the processing of the photos themselves. This new version of the application incorporates a feature called Night Sight that allows you to make beautiful night photos through the use of a machine learning algorithm that composes a perfectly illuminated photo from many shots: Night Sight: Seeing the Dark on Pixel Phones . Being the owner of a Pixel 2, and inveterate geek, I was frustrated last week to not own the latest Google smartphone that offers this feature Night Sight.
Obviously I’m not the only one to appreciate this new feature: Google deploys its “night vision” mode on smartphones Pixel: a real photographic miracle? Admittedly, making beautiful night photos is not the most noble achievement in the history of mankind, but this update is for me emblematic of an era: that of the optimization of the performances ( De the contribution of artificial intelligence for the fourth industrial revolution ).
We are talking here about “only” smartphones, but there are more striking examples, such as the update of the software that manages the ABS Tesla and allows electric cars to gain nearly 6 meters of braking distance to 100km / h: Tesla starts rolling out Model 3 update to reduce distance by up to 20 ft . I am not an automotive expert, but for me, 6 meters, that’s what differentiates between a living family and a shriveled family in a carcass. More recently, Tesla cars in a hurricane-prone area have been updated to increase battery life: Tesla remotely extends to Hurricane Michael .
At the launch of Tesla, the whole industry made fun of Elon Musk when he predicted that in the future we would not change cars anymore, we would simply update them. This prophecy seems now realized, because the regular updates of the different models of Tesla make it possible to improve their performances (autonomy, braking distance …). A principle that could be extrapolated to other equipment: lower gas consumption for Nest thermostats, shorter cooking times for Amazon’s microwave ovens, a more pronounced aroma for coffee machines. Espresso…
According to this approach, the value of a product no longer resides in its hardware components, but in the different software bricks that it integrates. This is more or less the case with smartphones from Apple and Google that are sold at a high price, not because they incorporate the best chips, but because they are equipped with the best software (iOS for one , Android “stock” for others). Software versions that do not have access to Chinese manufacturers who nevertheless offer cheaper and better equipped smartphones. Like what, the value of a product resides in the experience or the service that it renders, not in its material components .
Beyond the pure mercantile aspect (increasing the perceived value, therefore the selling price, therefore the margins), the improvement of the material performances is one of the main motivations of the use of the artificial intelligence. This improvement makes it possible in particular to optimize the consumption of resources: less electricity for cryptocurrency mining farms, less fuel for car journeys, less fuel for the heating of collective dwellings … Is this an utopia of to think that software updates can benefit the environment? Not at all, even on your small scale, for example using the Ecofont font that consumes 40% less ink than a traditional font. When you know that printers ink is more expensive than champagne, it’s quite useful, right?
What about optimizing human resources?
So on paper, artificial intelligences help us optimize the consumption of energy resources or raw materials. Great, it relieves our ecological conscience to say that the software that recommends a trip by car or a trip saves the resource (ex: TheTreep ). The intention is laudable insofar as raw materials (and especially fossil fuels) begin to be scarce. But when is it another resource that is expensive and highly unstable (the human being)?
If we have already experienced several waves of optimization of the human resource with the automation of the production chain, ie the replacement of workers by machines during the first three industrial revolutions. With the fourth industrial revolution, we are just at the beginning of the optimization of the human resource that we do not know (yet) to replace. Yes it’s digital platforms I’m talking about, and Uber is probably the best example of this search optimization: the self-employed are replacing the employees (who tend to go on strike or get sick ) and the algorithm rewards drivers who are the most invested, ie those who work at peak hours and in areas where demand is strongest.
The platform principle is to rely on digital tools to facilitate linking and lowering transaction costs. Doctolib is another emblematic example: by automating the making of appointments via a mobile application, health professionals do not need a medical secretary, and therefore a salary to be paid every month. While they replace a salary by intermediation fees charged by the platform, but it remains a profitable operation. In this case, we do not really destroy jobs, but we limit hiring . Viewed from this angle, platforms like Doctolib are harmful economic agents because they penalize employment; but on the other hand, they reduce the running costs of a medical office, thus limiting health expenses. The question is not simple, because what should be facilitated: the hiring of medical secretaries or access to care? This dilemma is certainly much more complex than it seems, because it brings many questions: should we focus on employment or purchasing power? What status should be given to platform providers (who are not truly independent)? How to solve the problems of personnel management (hired or independent) from the moment it is an algorithm that divides the tasks ( If your boss is an app, what, exactly, are you? ).
There are other examples of the use of software that is chilling in the back, especially those that are exploited to optimize the human resource as citizens (the face / visual recognition systems used in China: China can actually now identify citizens based on the way they walk ) or as voters ( How ZTE helps Venezuela create China-style social control ). The Venezuelan case is particularly worrying, because while the initial intention is commendable (to identify the population and its needs to optimize the distribution of food aid), the implementation of this digital identity system has been transformed into a new one. nightmare that futurists did not dare to consider: those who vote for the party in power are the first to benefit from state aid. In absolute terms, ration tickets and tactics of intimidation to guide the votes are not new (we have known many in the twentieth century, even in France), but here it is done on an industrial scale and systematic way.
A change of civilization that implies an update of mentalities
I admit that the link between the beautiful night photos of Google smartphones and the unfortunate fate of Venezuelans is thin, but there is one thing in common: the software used to optimize resources (material or human). The debate is not so much technical (smartphones vs. traditional cameras) or political (the must of interference in a food dictatorship), as cultural. This is not the first time that I evoke a change of civilization: the passage of the twentieth century analog digital twenty-first century.
From a computer point of view, everything works pretty well. But from a societal point of view, it gets stuck, especially when it comes to the propagation of alternative truths that influence the elections, or the anxiety generated by the rise of artificial intelligence and technologies. automating.
The harsh reality faced by citizens is that we are in a human race where technological innovation is forced, but public opinion has remained stuck in the twentieth century . Hence tensions (revolt of taxis) and a digital divide that grows, not at the level of equipment, but the acceptance of digital in everyday life. For simplicity: the people are happy to benefit from low prices on Amazon, original series on Netflix or friendly games on their smartphone, but when asked to change their habits or question their skills So, their jobs, they take offense, yell at the plot and point to the wicked GAFA ( Advocacy against digital populism ).
In general, we can note a strong mistrust of algorithms and digital: Public Attitudes Toward Computer Algorithms . A widespread mistrust resulting from a lack of knowledge of technological innovations, their origin, what they are capable of doing or not doing . For example, one wonders about the place of the autonomous vehicles, whereas these will not be visibly not ready for many years ( It’s 2018, where are the self-driving cars?) And Autonomous Vehicles Have Some Regulatory Roadblocks to Navigate ). Similarly, we fantasize about the advent of a generalist artificial intelligence that will enslave humanity while specialists do not even know if it will be possible, even in 50 years ( How far are we from artificial “general” intelligence, really? , Google ponders Shortcomings of the machine learning and Facebook’s head of AI wants us to stop using the Terminator to talk about AI ).
These fears are very clearly irrational. I am always surprised by the ability of citizens to fear the rise of automation or machine learning when they consume products that have been manufactured by robots for decades or that they benefit from services, equipment or offers that are based on artificial intelligences also used since the last century (ex: voice discs, automatic boxes, DAB …).
There is therefore a genuine psychological and emotional challenge to be met to ensure that the population accepts the idea of cohabiting with machines and algorithms . Again, this mental process is rather surprising, as I doubt that they agree to give up their car for the benefit of a horse, or even give up features like ABS or traction control.
All this inevitably leads us to the digital culture deficit. A problem that gets worse every year as the digital debt grows and the digital giants nibble the traditional economy ( Digital acceleration is a reality, and it benefits especially GAFA ). Maybe one day, public authorities and companies will recognize the benefits of digital awareness and the interest of changing mentalities (update our cultural reference) to better live in a digital daily and be in phase with this new civilization.