Category: Uncategorized

Smoking is deadly. I see this on every cigarette pack in Germany, and in every part of the world.

I was with a friend of mine, he is a smoker, he was telling me these photos of sick people they put on the packs have no influence on him. I told him if they didn’t work, they wouldn’t be there. Maybe I am wrong, maybe they are there because a lobbyist convinced governments to put it without real evidence to support that this works.

This reminded me of the identifiable victim effect. People sympathize more with a cause if they see someone hurt, more than if they see an anonymous person, and more than if they just see the number of people hurt.

Saying donate to save “name” led to more donations than saying donate to save people suffering from “cause”.

Maybe the photo on the cigarette pack doesn’t work because it is anonymous. Maybe instead of saying “Smoking leads to xx” they should say “xx suffered from yy because he was a heavy smoker, he died in 20xx”.

One of the common misconceptions people make is confusing rain probability with rain amount.

The higher the chances it is going to rain doesn’t necessarily mean it will rain more water, and vise versa.

I was planning a trip today, and I used the wheelchair accessibility filter on AirBnB, I got this popup.

This text and the call to action “I understand” sounded to me like “we know your anxiety when you book something, but we are not responsible for it”.

Probably a big problem happened with a wheelchair user, or AirBnB customer service reported wheelchair users are making many complaints, which led product to do this.

Finding wheelchair accessible accommodation is a big pain, even with hotels. When I was working for Booking, I worked on a project to define the features of wheelchair accessible accommodations. Unfortunately this project never got to the company priorities because of the very small niche it targets*, and the logistics needed to communicate and serve what’s accessible.

I hope AirBnB dedicate more resources to make hosts clarify their house accessibility options. AirBnB positions itself as an inclusive community, and they can do better.

* 0.40% of Europeans use wheelchairs, out of those how many travel?

Few weeks ago I finished Chaos Monkeys. It is the story of a physicist turned Goldman Sachs analyst, who quit wall street to join a Silicon Valley startup, then went to found his own startup, got into YC, got sued by his previous employer, sold the company to twitter, and landed as one of Facebook’s early Ads product managers, before leaving Facebook and becoming a twitter advisor. As crazy as it sounds, it is a real story with real names and real companies.

As an outsider the valley, and so as most of people, it takes long time to create a mental image of what happens there.

What we see in tech media is full of noise, and it takes time to be able to read between the lines. Personally for me this came through a combination of heavy twitter usage, podcasts, medium, and different media outlets consumption. That’s why when I read this book, I felt like it is a wrap up to everything in a single man’s story.

Here is a quote from the book, about silicon valley rules

Investors are people with more money than time.

Employees are people with more time than money.

Entrepreneurs are simply the seductive go-betweens.

Startups are business experiments performed with other people’s money.

Marketing is like sex: only losers pay for it.

Company culture is what goes without saying.

There are no real rules, only laws.

Success forgives all sins.

People who leak to you, leak about you.

Meritocracy is the propaganda we use to bless the charade.

Greed and vanity are the twin engines of bourgeois society.

Most managers are incompetent and maintain their jobs via inertia and politics.

Lawsuits are merely expensive feints in a well-scripted conflict narrative between corporate entities.

Capitalism is an amoral farce in which every player – investor, employee, entrepreneur, consumer – is complicit.

If you want to understand these lines in detail, read the book.

Since I moved to Berlin, I feel like I am on the most wanted list for Christian evangelists.

Every time they see me, they stop me to give me a brochure. It started to extend further than those in subway stations. Sometimes I get stopped by random people on the street, I think they need help go somewhere to find them giving me one of the brochures. Today a lady stopped me, opened her wallet, and kept searching until she found a brochure to give me.

I am not annoyed by them. They are nice people. It is just an observation.

I was talking to a friend yesterday and telling her why I love Berlin. I found myself spontaneously saying: it is organized so that it is not annoying, and it is chaotic so that it is not boring.

This says it all.

As summer comes, error rates go up, soldouts are more, and provider outages become a norm in a travel company like ours.

I was thinking what would solve the problem of providers being down?

I think the root cause is the centralization of the system. GoEuro, and whomever wants to allow users to search and book tickets have to connect to the rail/bus company’s system. If this system is down, all the partners are also down*.

On the other side, train companies are public companies, their sales is well known – most of the time – and even if the numbers are not public, through the APIs/scraping and if you have the right tracking, you can know when are they going to be soldout, or how many tickets they sell. The worst case you will get a fairly close, yet inaccurate estimation of these numbers.

This makes me think the solution to all of this is a decentralized train availability blockchain. A public ledger of who is on each train, where user identity is anonymous/hashed and can only be unhashed with their private key (for when the conductors are checking them on the train). Miners who verify the transactions are rewarded in ticketcoins, which they can use to hop on trains or sell them to other users.

This same blockchain can have different coin types for different providers, for example a DuetscheBahnCoin for Deutsche Bahn, a FlixBusCoin for FlixBus**.

The beauty of this – although I am not sure about its feasibility – is it is decentralized. You no longer have  provider bottleneck suffocating everyone from booking their tickets, it also offloads big part of the effort from the provider, since the computing, storage, and technical effort is mostly offloaded to the public.

* Companies can still cache and show results, but users would still be unable to book.

** This will be a bigger challenge with private companies since they don’t want to reveal their capacity and sales.

Teams meet, discuss, and decide to have a microservices strategy.

They start migrating the monolithic system to microservices, along the way they decide to build new features, for the sake of speed, those new features have to be built inside the monolithic system, and later be migrated to microservice.

A divide happens in teams, with half migrating to microservice and half adding more to the monolithic system. At one point, the team migrating to microservice is rushed to support more features into the microservices they just built without consideration of the bigger picture, after huge number of resources spent, they are back to square one, with a legacy monolithic system, and half cooked microservices that kept growing until they became other monolithic systems.

Then teams meet, discuss, and decide to have a microservices strategy.

As I am still reading Chaos Monkeys, I came through this brilliant paragraph.

Even Facebook, whose ability to maintain a fast-moving, always-be-shipping culture well into corporate middle age was admirable and unique, was simply a German-style autobahn, not a racetrack. The days of a few engineers going rogue and launching Facebook Video despite Zuck’s wishes were long gone. Most vehicles moved at the speed limit, lots of trucks hogged the right lane, and a select few drivers traveled at full speed in the left lane (with no-speed-limit passes given out by Zuck alone, and anyone else who dared race ahead did so at peril to his or her career).

When people ask me what was the culture at booking.com, I used to say it is not a startup as they claim, but it is also not a corporate à la IBM or Oracle. It is something in the middle I couldn’t describe, until I read that paragraph. That’s it, “it was simply a German-style autobahn”. Teams are autonomous, but they can only move within the speed limits of their tracks, few teams had the green light to pursue new and bold ideas, as long as they are not far from the core (read about the stack fallacy).

This post is not to criticize but rather to draw a mental model of how to think about it. I am very grateful for the time I spent there. I learned so much and met many great people. I believe there are areas like SEM, conversion optimization, and pragmatic usage of data/machine learning, Booking.com is one of the best places on the planet to learn and work on these things (if you pass their tough interviews).

 

There is a lot going around the rhetorical of who is going to lead the next tech era, is it the US or China? (there is no Europe in this discussion). One of the balanced opinions on the matter is this post. It was tweeted earlier by Andrew Ng.

The tl;dr version is that AI research in China is mostly driven by a pressuring consumer need for AI powered solutions. For example typing Chinese on mobile is very hard, which drove companies to do research on alternative input methods such as voice. Another example mentions Chinese need to consume non Chinese content, which also drives innovation in machine translation.

China excels in certain aspects of AI in part because there’s a confluence of cultural, technological, economic and geographic market forces that make those aspects a higher priority in China.

….

I take that not as fear that China will come to dominate the AI arena and thus the world, but rather that Chinese companies are serving their consumers with cool products, and U.S. companies need to do the same—only with products that U.S. consumers actually want.

One thought I had recently was to learn Chinese*. There is no doubt in a not so far future, there will be huge interest in that part of the world and what it provides for tech. People who are able to understand what’s going on there will be of great value, since capital follows opportunities.

However, a counter-thought I also have is that China is a closed system, they have over a billion people, which gives them the ability to be self sufficient and not needing outsiders as much as US and Europe do. They are also a communist country, which while we always think of as a bad thing, but it has some benefits. They can get as many people as they want into computer science**, they can make programming mandatory in schools, and they can do it faster than their US counterparts***.

I believe China will be a big power player in the tech world, I am curious to see how this transformation will impact the west. Will everyone learn Chinese and try to go there, or the west to China will be like the middle east for tech giants, only used to market and sell whatever the dragon builds.

* Says the man who is 10 months in Germany and can’t finish a full sentence in German.

** Trust me on this one, I am from a country that decides how many people can get into engineering, medicine, and every other field. Communist/socialist/authoritarian systems can control the supply, and hence the demand for universities.

*** This argument isn’t new and doesn’t only apply to tech, for example think of the benefits of a social democracy (Europe) or a communist government on providing more inclusive healthcare compared to the US.