1. The average food quality in Japan is quite high, be it Tokyo or one of the smaller towns and cities. The average sushi in Japan seemed superior to the best sushi I have had in Germany. The best sushi I had in Japan was also much cheaper than the ones in Germany. When key ingredients are sourced nearby, the prices drop.

2. Sit-down restaurants are not the only places you can find delicious food in Japan. Hole in the wall shops and food stalls (yatai) are great options. You can find rows of yatai in Fukuoka or stand-alone ones in different places. Some food stalls excel in one dish. Tamagoyaki (Japanese omelet) from a food-stall run by an elderly Japanese near Ōji station in Tokyo was my favourite.

3. Japanese restaurants specialise in specific genres of food. Prefer a sushi-only place or an Okonomiyaki place over one that offers many different kinds of food.

There are times when we make a decision and move on. Then there are others when we go back to a decision and regret the road not taken. To decide, what kind of choices do we have?

AI is not restricted to data-hungry techniques. There are several alternative techniques that rely on less data, require less computation and less memory. They are more in line with the principle of data minimisation in the GDPR. The Commission should take a smarter approach and incentivise AI techniques that require less data, that benefit society and that assists with climate change mitigation.

More data is not always useful for AI

We often wait. But, not all waits are made the same. Waiting for the train that we take every weekday morning is not the same as waiting for a friend to arrive. We know how long we have to wait for the train to arrive. At least approximately. We may not know when the friend will arrive if they are already late.

There are other times when the wait is longer. When you are waiting to receive the reviews for the first academic article you submitted to a conference or journal. When you wait for the approval of a visa application and don’t know whether it will be approved. We know how long we need to wait and yet, we are anxious.

In Agnès Varda’s film Cléo from 5 to 7 (1962), we see Cléo wait 90 minutes[^1] for her medical test results. She suspects that she might have cancer. In the beginning, we see her wait anxiously. She asks the opinion of a tarot card reader, shops with her assistant, practices a song with a composer who comes to her apartment and has a superficial afternoon chat with her lover.

On 2 November 2021, Facebook announced that they will delete the data and shut down the facial recognition system on Facebook.

Which data is being deleted? Facebook’s blogpost does not say that they will delete the models that were generated using the data. It also does not say that they will not use people’s image data to train models. It only says “we will delete more than a billion people’s individual facial recognition templates.” Is deleting templates enough?

What are these templates? Templates are not images. Templates are generated using images. To understand how templates fit into a facial recognition system, it can be useful to understand the different steps involved in a template-based facial recognition system [^1]. Here is a simplified version of the steps involved:

I have many friends who place the sunlight on a pedestal. They value the sun much more than many other natural elements. They have a much more favourable view of the sunlight than the rain, for instance. This is in part because they have grown up in parts of the world where sunlight and the warmth it offers is not prevalent throughout the year. They make the most of the opportunities they get to bask under the sunlight.

Would they feel the same if the sun was beating down, day after day, without any possibility to find shade? Many who live in hotter parts of the world, especially closer to the equator experience the sun in this way. Many others, even those further away from the equator, have tasted days when the heat was too much to bear.

The sun can also be used as a metaphor as in the film A Sun (2019). The elder son, Hao, of A-Wen is brilliant. He is the sun of the family. His parents are proud of him. His brother hates him for being brilliant. He is shy, helpful and caring. People around him know that he is thoughtful. But, they have no idea what he is thinking.

A-Wen in the 2019 film A Sun is one such person. His identity as a father is questioned when the younger of his two sons is sent to juvenile detention for accompanying a friend who chopped-off the hand of a similar-aged boy. A-Wen’s approach to continue living his life involves a mental tweak. Whenever someone asks him how many children he has, he responds confidently that he has one son. This mental tweak helps him to avoid thinking about his role as a father in the development of his younger son.

This mental tweak helps him for a few days until he faces another tragedy. One evening, the older son jumps to his death from the roof of their apartment building. His older son had been the spark in the family. Excellent student at school and high-achieving. He was shy and thoughtful. No one really knew what he was thinking about. Just that he was often deep in thought. He was the son the father was proud of.

How many sons did the father have now? He still had one son, the one he did not want to acknowledge. No mental tweak needed now. As the younger one tells one of the officers in the detention center, the older brother was brilliant. He was just brilliant. But, it does not matter anymore. Only the younger one is alive now.

Many others have written about the system, how it works and the potential privacy and surveillance issues. As part of the system, Apple uses a variant of private set intersection (PSI) that when properly used can provide cryptographic privacy. I want to focus on this part of the system and whether its use provides even cryptographic privacy in this context.

PSI is a cryptographic technique where only the intersection of two data sets is revealed and nothing else about the data sets is revealed. Lets say that you and I are huge fans of Marcel Proust’s In Search of Lost Time. But, neither of us has the entire collection of seven volumes. We want to avoid embarrassing each other by mentioning Proust’s writing from the volumes that one of us has not read. We want to identify the books that both of us have read. We also don’t want to leak which other volumes we have read. For this purpose, we could run a PSI protocol to identify which volumes both of us have while not revealing anything else.

Apple PSI does something similar. In fact, it does even better. It does not even reveal the intersection, only the associated data when there is an intersection. However, I think that the assumptions required for the protocol to provide cryptographic privacy in this particular setting of Apple do not hold.

Sending and receiving SMSes was a habit. Some days a few tens of messages and on other days a little more than hundred messages. SMSes were not free. We were able to pay a small lump-sum to have the privilege to send a few thousand SMSes over a month.

As useful as text messages were, there were times when we had to call each other. Calls were much more expensive. For someone still in school, with no income, saving costs was the priority. So we had to figure out a way to use the calling function of our mobile phones without having to pay for them. This was possible, as long as we did not need to speak over the phone. You might wonder, what might be the use of calling without speaking. Read on.

Lets say I was to travel to a book exhibition by train with my friend Sam. Sam lives a few kilometres away from my place and we board the train at different train stations. Sam boards the train at station A and I board the train at station B. Sam wants to make sure that we board the same train so that we can have a pleasant chat during the journey. We had decided that we will board the first carriage of the train in the direction of travel. We still needed a way to make sure we do not board different trains, especially as there was one train every 5 minutes on that particular route.