data availability | The Fourth Revolution Blog

How Everyone you Will Ever Meet Knows Something that You Don’t

“Everyone you Will Ever Meet Knows Something that You Don’t” is a quote by Bill Nye, and american science educator. It is a very powerful statement that I find by experience to be actually quite true. However we can only find out provided we take the time to establish the right connection to figure it out.

The reason for this situation is of course the variety of individual experience and interests.

We often tend to dismiss the knowledge that is available around us, while daily experience shows how fruitful it can be. For example on the workplace, leveraging on the interest and knowledge of the people constituting the team or the extended team is a very effective way to increase effectiveness. It is too often forgotten in particular with the race to efficiency.

Let’s never forget that anyone around us, however menial their occupation be, have something to teach us.

Hat tip to Valeria Maltoni post ‘Conversations are Not a Promotional Opportunity‘

How AI Powered Chatbots Start to Be Everywhere

The new trend seems to be Artificial-Intelligence powered chatbots. In an interesting experiment, a university professor replaced one of his assistants by such a chatbot. The students did not detect it and actually wanted to nominate it as the best assistant in the class! (see ‘Innovations
What happened when a professor built a chatbot to be his teaching assistant‘)

chatbot While we are probably still miles away from a chatbot that can emulate a real speaking conversation, instant messaging chatbots seem to exist quite successfully in a number of instances. This removes clerical work for sure and we need to get used to it. In all instances more complex requests need to be handled by humans who are unburdened by the menial requests.

Actually it would be interesting to have a test or some kind of identification that would enable us to determine if we are interacting with a chatbot or a human. But we need to be aware that it will be increasingly the rule that our interface will be with robots. In 5 year time it might even be the majority of the customer service we will be facing. For sure that will produce sometimes astonishing results.

How Most People Sharing Links or Commenting Don’t Read the Posts

Since I am publishing a lot, I am aware that many times when people comment on my posts, they seem not to have read it. They react on the title, or on some idea more or less closely linked to the post topic.

This is even confirmed when sharing links- according to studies, ‘6 in 10 of you will share this link without reading it, a new, depressing study says‘. Moreover, the study shows that viral effects can be created without anyone having read the original post! So much on the attractiveness of the picture or the video enclosed with the text to help through the click rate…

It is difficult in the Collaborative Age to read everything that is thrown to us. But we can maybe check first that what we are sharing makes sense. Another basic behavior rule that will slowly emerge and be taught to future generations maybe?

Why Big Data Prediction Capability will Remain Limited

Big Data is trendy, and the graal of Big Data is to be able to predict behaviors and ultimately influence them. But the world is complex and whatever power we put behind Big Data, there will be a close limit to what can be inferred.

The most well known complex system is weather. In spite of the tremendous increase in computing power in the last decades, our prediction capacity remains limited to a week or so. That is because it is inherent to a complex system that prediction capability is limited by the system, our understanding of the initial conditions, and not by its equations or by the computing power we put behind.

So the graal of Big Data is in fact elusive – it will never possible to predict behaviors beyond a certain limit which is still to be determined practically.

Big Data will never allow the long term prediction we hope for. It will be a disappointment for many. It is also another sign of our freedom.

How Software Default Values Impact Real Life

When you write software, to avoid bugs you assign to each variable some default value, that is afterwards supposed to be updated by the program.

The farm at the default location for IP addresses in the US

What happens when the default value does not get updated? Something like the nightmare of happening to be located at the default value of a mapping application, like what happened to a quiet farm in Kansas. The story told in this Fusion article ‘How an internet mapping glitch turned a random Kansas farm into a digital hell‘ is really though-provoking. Just because it happens to be at the center of the country, it is mapped as the location by default of IP addresses. The article contains many other similar stories of misplaced geographical locations of IP addresses.

It happens all the time also on our favorite online maps when they show the center of a long avenue when searching for an address – this center could be far remote from the actual location we are looking for!

I am not speaking of people driving in entirely wrong locations by their GPS because they did not check that they had selected the adequate town or checked there was actually a road!

Even in this century of overwhelming information some checks are required before believing what the machine says. Stay vigilant!

What We Should Trust From the “Man on Site”

There are two schools of thought regarding how truthful the information from the man on site can be. One school follows Winston Churchill: “Never trust the man on the spot“. Another school believes that local knowledge offers sometimes a better insight than what is available in headquarters.

What’s the right way about this? It’s all about what information we want to have.

Information about the actual progress and the actual situation on the ground is best retrieved from site. Far-away management does not work and leads to unrealistic assessments of the situation. I observe this effect all too often in large projects.

On the other hand, do not expect the site people to have a very worthwhile assessment of the whole strategic or even tactical picture. They can only have a limited view of the whole due to their position. The breadth of the subjects they can apprehend depends on their scope. Local representatives in a particular country will often have a much better assessment of the political situation of that entire country and what can or cannot be done than the global headquarters. A local representative on a site can only apprehend very local issues. In general I have observed that often the local representative can be trusted on a scope slightly larger than his assignment.

In general, I tend to trust more the people on site except if the topic is clearly beyond their observation range.

Churchill quote from H. R. McMaster Dereliction of Duty (a recommended read about how the US politicians and top military got embroiled in the Vietnam war)

Why there is Debate Between Content Moderation and Free Speech

Moderation on social networks is an essential feature. In an excellent essay on the Verge ‘The secret rules of the internet – The murky history of moderation, and how it’s shaping the future of free speech‘, the relation to free speech is discussed. “As law professor Jeffrey Rosen first said many years ago of Facebook, these platforms have “more power in determining who can speak and who can be heard around the globe than any Supreme Court justice, any king or any president.'”

The testimonies about content moderation are quite breathtaking, and the decisions whether to keep some videos that have shocking content but are important from the political perspective (like the murder of people during demonstrations) an example of tough decisions to make.

And because “The stakes of moderation can be immense. As of last summer, social media platforms — predominantly Facebook — accounted for 43 percent of all traffic to major news sites. Nearly two-thirds of Facebook and Twitter users access their news through their feeds“, this determines what people will ultimately see from the world.

Of course before there was journalism, a limited number of sources and effective censorship by governments. What has changed is that it is now privately handled and not susceptible to democratic control. I would anticipate that at some stage, guidelines might be defined by governments (e.g. related to anti terror campaigns) but at the moment it is an issue to be kept in mind.

How Mobile is Eating the World

Mobile is eating the world, and the proof is in a quite famous presentation by Benedict Evans from the Venture Capital firm Andreessen Horowitz which has been updated in 2016.

I share here some highlights which have particularly struck me.

First, mobile will represent a roughly 5-10x increase in the number of users and devices compared to the previous ecosystems, and is quite comparable to the move to personal PCs, as shown in the figure on the right. And this means everyone has a super computer in his/her pocket. The presentation goes on to argue that this will also lead to a significant increase of productivity. That might be true in theory, but I think this can be controversial as mobile devices are also a great source of lost time! (ref our post on How Mobile Phones Distract Us – A Real Life Example).

Mobile apps lead to a concentration of valuable online properties

Mobile is an ecosystem and actually most people use apps to access the data of the internet. What happens is that actually, people don’t use a lot of apps on their mobile devices, hence there are much less valuable online properties, but they are quite more valuable. Actually one third of the users access the internet through Facebook!!

Finally, we are just at the beginning of the disruption brought by mobile devices in the business models that pervaded the world before, so hold on!

The full presentation is available below:

Mobile Is Eating the World (2016) from a16z

How We Now Have the Opportunity to Overcome the Limits of Averaging

The Fourth Revolution and the availability of data and data processing allows us to go one step beyond looking at averages. We can now observe data distributions and figure out finer conclusions than those based simply on averages.

This seems only a small step, but right now our institutions and large companies still have not figured it out. It is is the same in projects – we only consider average performance. Average makes sense only when aggregating a large number of instances, but we lose so much information in doing so!

In particular we miss critical information on what is working better and where we can seek to learn what is the impediment to better performance. We miss information on the variability which is so important. Read Seth Godin’s post On average, averages are stupid to have a great illustration of this effect.

Let’s get beyond averages and take benefit of the wealth of available data to produce better informed decisions!

Why We Should be Particularly Wary of Unanimous Situations

Unanimous opinions and decisions should be looked upon suspiciously, because they might reveal common cause of mistake. “Unanimity is often assumed to be reliable. However, it turns out that the probability of a large number of people all agreeing is small, so our confidence in unanimity is ill-founded. This ‘paradox of unanimity’ shows that often we are far less certain than we think.” The idea is developed in this excellent post on phys.org ‘Why too much evidence can be a bad thing‘.

“Under ancient Jewish law, if a suspect on trial was unanimously found guilty by all judges, then the suspect was acquitted. This reasoning sounds counterintuitive, but the legislators of the time had noticed that unanimous agreement often indicates the presence of systemic error in the judicial process, even if the exact nature of the error is yet to be discovered. They intuitively reasoned that when something seems too good to be true, most likely a mistake was made.”

In any case the paper shows that when results of a process or experiment are too consistent to be true we should search for a common cause that might explain this consistency. An example in the paper is particularly vivid: “Police in Europe found the same female DNA in about 15 crime scenes across France, Germany, and Austria. This mysterious killer was dubbed the Phantom of Heilbronn and the police never found her. The DNA evidence was consistent and overwhelming, yet it was wrong. It turned out to be a systemic error. The cotton swabs used to collect the DNA samples were accidentally contaminated, by the same lady, in the factory that made the swabs.”

So the next time that you are faced with an overwhelming unanimity, look further for a possible cause that have nothing to do with what is being decided or sought. It might save the day!

How Powerful Algorithms That Shape Our Lives Still Rely on Human Creativity

The great Quartz post ‘The magic that makes Spotify’s Discover Weekly playlists so damn good‘ gives a wonderful insight of the positive brought to us by advanced algorithms and basic Artifical Intelligence.

It goes into the details of how Spotify proposes new playlists based on one’s own preference, the playlists of other members with close preferences, and advanced algorithmic to bring all together into a wonderful proposal of new music tracks.

What I find extremely interesting is how the basis for the value that is created is actually human-produced: the playlists of other people. The algorithm does not find new tracks or discover new musicians by itself. It relies on the curiosity, the knowledge of its members. The algorithm exploits the community effect to create value for all members, leveraging the efforts and chance encounters of all subscribers.

“Spotify is also using deep learning—a technique for recognizing patterns in enormous amounts of data, with powerful computers that are “trained” by humans—to improve its Discover Weekly picks“. That’s where AI comes in to further improve the algorithm. But still at the core are the lists of others and how they interact with them to fine-tune their preferences.

All those algorithms enhance the power of the community but can’t replace it. All original creative data is still created by humans. Algorithms are still only powerful crutches to create value in our lives putting together all these individual contributions.

How Collaborative Networks Always Rely on Few People

In collaborative networks, forums and wikis, actual production only relies on a small percentage of users. This is confirmed in a business environment in a post from the Harvard Business Review ‘Collaborative Overload‘: “In most cases, 20% to 35% of value-added collaborations come from only 3% to 5% of employees“.

The reasons are multiple:

Collaborative systems act as complex systems and hence, contribution follow a ‘long tail’ curve: major contributors really produce a large part of the value (however the aggregated value of the contributions of all the others should not be neglected)
Most users generate interactions of low value to the community
Most users are swamped by daily urgencies and do not have the time to do longer term contributions.

This small percentage has an interesting implication when it comes to organizations’ internal collaborative networks – they can only work if there is a sufficient number of potential users so that the core group of 3-5% of users generating most of the value is large enough. That is why a minimum of a few hundred to a few thousand potential users is necessary for successful internal collaborative networks.

The entire HBR’s paper is quite an interesting read as it focuses on the emotional drain for the key collaboration contributors and the fact that their contribution is often not recognized enough.

M	T	W	T	F	S	S
						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30