Web Summit Lisbonne 2017 – some insights

The Web Summit held place in Lisbonne, 5th to 9th of November 2017. It was a great opportunity to meet actors on digital market, but also to hear about 5 topics that attract my attention at this conference:


  1. Do tech companies pay enough tax?

This debate during the Web Summit showed 2 interesting perspectives on tax avoidance and innovations’ monopoly.

First, let’s say that legislations have allowed tax avoidance.

If we reached that situation, it is because tech companies have grown so rapidly, there is a huge inequality as corporate tax rate has decreased.

To the question: should we raise tax? The answer is we should not penalised small tech companies.

The difficulty also comes to the fact that tech companies are hard to define. The problem with tech companies is not the “tech” part of it, but the “companies”. The risk is that increasing tax may kill innovation.

“No, the world is not worse with tech company. It became a better place because of tech companies over 1000 years.”

There are no other issues like hiring, immigrations, bigger than tax.

Should small and big companies pay the same tax? Why do country let big companies allow not to pay their taxes?

Globalisation has done good things to us. People can travel now easily. This human creation are great and we can destroy the planet too. Tech can help fix the mess.

GAFA and others will help make a better world.


  1. Europe Great again and Google

Matt Brittin (Google) that used to work in publishing came to the Web Summit to talk about how Google sees Europe and how they encourage everyone to become an entrepreneur.

“Our objective is to be a good citizen. We want Europe to be successful with the digital world.”

They want to engage as a constructive partner to succeed.

When it comes to privacy, Google position is that they created a “my account” Google, created by German, to allow anyone to adjust and personalize its privacy.

Going back to the subject of tax, Google said they pay taxes everywhere.

Google is an Open source operating system, that allows competitors such as Androïd to use its services. They also offer free trainings such as in the Google Digital Garage

Google Digital Garage


The subject of Google responsibility when it comes to post-truth was raised. 18 Youtube channels link to Russian activist were out, does Google feel responsible? What actions?

Matt Brittin agreed to say misrepresentation and spamming content wass an issue. Here are some actions, Google took to takle post-truth: 11 billions $ were allocated to this subject. We can’t imagine a magical solution to solve this.


5 billions people are connected and they are all publishers now.

The chance with Google is that you ask for a specific subject and we bring you the most accurate content. Google see itself as a navigator, 10 billions clics on Google news are going to publishers’ website. Their role is to help publisher to publish their content.

To conclude the discussion, Matt B said we need to think big when we start a business in Europe. Our tendancy is to think local.


  1. Amazon : protect and servce


Amazon speaker pointed the protection of datas as key for success in the digital business.

Security is the first thing to do when you create a startup : protect your customers’ data!

Any information we have about customers have to be well protected. There is no easy solution for this.

Right softwares and development are required for this to keep hackers out.


5 years ago, we did lots of releases a day, how did we make sure security was integrated?

Everybody needs to be careful about this.

Security team is made of all of this, as partners for any product development:

– Operations

– Engineering

– Application Security

– Compliance

The security team cannot arrive after the product is developed, it is too late.

Security automation : Humans can’t face rapid changes that are happening now. That is why continuous development needs to include security in the pipeline (continuous scan, login…)

Tools are available for this like in Amazon with AWS.

We need clear processes that describe how our infrastructure is going to look like.

To automate security processes=

– create guidance

– assess your vulnerability

– create rule packages and choose wichone is best

– AWS cloudtrail : get all information to need to provide security teams

– AMAZON MACIE : identify sensitive datas that were not protected the right way at the beginning…This tool helps us to discover quality of information we are collected.

Usine Machine Learning to help notify dangers that we may not be aware of such as subnormal patterns of our users.

There is a strong need to encrypt like everyone is. You should encrypt everything all the time because that is the way we are sure to be the only one to access to data.

Amazon created a certificate manager training with a free certificate and presented in during the Web Summit.


  1. Dark trace : face Attacks from the web

Dark Trace

Trust attacks can be within cooperations , they are engineered by skilled attackers, very discreet. There is a new kind of attack from AI . If the attack is using AI, he would learn how to work around your defenses. As an answer to this,  we can look at the way our imute system works.

We can analyse of the traffic is normal, what is unusual. I liked the comparison with a body that tells you where it hurts. The probability that the network is under attack can be measured.


How can we use AI to response to the attack?

It is difficult, 2 types of set up :

  1. Networks with rules may have to face lots of different possibilities, it is impossible to figure all of them out.
  2. Signatures attacks : autonomous response. They areautomatically slow down, you can take action or shut down attack.

Dark Trace did some work with hedgefunds on recommandation mode confirmed by human mode. That helps the AI to learn, and if the AI keeps doing right , it becomes a « trust mode » … Human interferes only at the beginning of the control process.

Machine learning 

A team of 80 threat analysts are dedicated to look for attacks. Their methods were analyzed and transferred to Machine learning.

Deep learning

Deep learning will help by telling you: “you can you go back and watch if I have to take action or not” to human.

Only deep learning car watch out attacks that stay in a system for years.

IOT is the biggest danger and was highly mentioned in the Web Summit. In our culture today, noone is bringing IOT to their IT department to check for security. One good example of a danger is this aquarium with IOT in a casino managed to uploaded all datas from the Casino. http://money.cnn.com/2017/07/19/technology/fish-tank-hack-darktrace/index.html

  1. Fake news (NY Times, Outbrain, European Commission talk together)

To be a publisher or not to be, that is the question for Facebook and Google? The content they provide is produced by others, but do they have a responsibility with it?

Ann Mettler from European Commission says : « we have to stop lying to ourselves about fake news ». Even if no “ministry of truth” is planned to be built, they are going to take actions “because there is too much as stake”.


Facebook & google don’t want to be considered as a publisher or media company. Otherwise they would have responsbility of what they say. They are trying to decide weather or not they have responsibility on the quality of content and it requires a lot of work to to go down that road.

Their business model is to be neutral and available for any kind of publisher. Governements have responsibilities on educating people on the right usage of the web.


I hope those insigths gave you an overview of how passioning subjects are in this great events.

If you want to know more on this event, please click here.

%d blogueurs aiment cette page :