Interview with the founder of Instagram: “Being ahead of your time is also a mistake”

Mike Krieger is the co-founder of Instagram and head of the application development team. It was he who created the first version several years ago, and now he is called the soul of the company, although he is less media than his partner Kevin Systrom. Before Instagram, he was a designer and developer on the Meebo team, which created a browser-based messaging app.

Krieger (pictured left) spoke to The Verge about the origins of Instagram filters, the principles of human-computer interaction, and some of the service's crazy users. We publish our translation.

How have you been doing since Facebook bought you ?

We add new code to Instagram about 20 times a day. We think this way: if something is ready, then we implement it. The fact that Facebook is adding new code twice a day is also impressive considering how many engineers they have. Now it’s like, “We would add this, but Mike needs to look at it in case something goes wrong, but he’s in a meeting.” I actually slow down the team if I try to create more stuff, so I need to step back a little and make sure we're focusing on the right things.

Obviously, you don't implement new features 20 times a day. What do you actually do?

Basically it's optimization. We've gone from 30 million to 100 million active users in a year or less. Users get essentially the same (or better) experience, but the infrastructure that supports it all has changed quite a bit.

In San Francisco there is such a thing called Developer Boot Camp (training for developers - editor's note). I gave a talk there, which I ended with a slide saying: “At any given time, you will be unhappy with at least half of your infrastructure,” which is undoubtedly true.

tweet

There are many things that constantly need to be improved - it’s a kind of Sisyphean work.

You optimize the entire system from start to finish, and then you have to start over again because conditions have changed.

The news feed has become faster, so?

We were obsessed with it. From day one, we wanted to make the system fast, both on the client and on the server. Lauren Brichter, a former Apple and Twitter employee, once wrote about this: “Okay, you guys are doing this all wrong. You're trying to have 30 different layers on the screen, but you want it to stay flat—make it monolithic." We spent many hours on this. In the first version we cut out a lot of features just to make everything work very, very quickly.

People are now getting away with doing more but optimizing less thanks to the existence of the iPhone 4, 4S and 5. They just have better processors. I love reading Jordan Mechner , creator of Prince of Persia. He posted online his diaries, which he kept while writing the game - then he was 18 or 19 years old. There is a moment in The Prince where the main character's shadow appears. Jordan writes: "We didn't have the spare memory to make a second character's shadow, so we just mirrored the Prince himself bit by bit - so it's his actual shadow." This is an example of the ingenuity you have to be if all you have is 8 kilobytes of memory.

What do you think about music on the Internet, and how do you think music streaming should work?

I remember when I was in college, an anonymous donor gave Stanford students an entire year of subscriptions to Yahoo Music Engine. This is a service like Rdio or Spotify , but it worked 10 years ago. I wrote to Ian Rogers, the creator of all this, and asked: “Dude, you had such a cool thing, what happened?”, to which he replied: “Being ahead of your time is also a mistake.”

I remember being drawn to this thing, and then it disappeared. Lala did something similar, but then also disappeared. And then when Rdio finally came out, I thought it was incredible. As for me, this is complete harmony. But I was a little surprised when I found out how it works in economic terms - it turns out that this is not the best way to support artists. Now, if I like the way someone sings, I just buy the album on iTunes, even if it's just a symbolic gesture - I still listen to those songs on Rdio later. It's my way of saying, "I liked your music, here's my money."

You studied human-computer interaction in college. Why did you choose this area?

My favorite thing in high school was electives, where you sit and learn all kinds of computer stuff with the other kids. I remember us sitting and mastering concepts that seemed completely obvious. For example, when you click on a link in your browser, how do you know that something has started loading? Then the answer was simple: that little globe started spinning in Internet Explorer.

The hardest thing for people who didn't grow up with computers was the click. They will move the cursor to an object, hit the mouse with their whole hand, which, of course, will move the cursor - and nothing will happen. It's really annoying.

tweet

And why do you need to click once in the browser, and twice in the OS itself?

Why? I knew that I was interested in making all this more convenient for people, so I went to study interface development.

How have you applied this knowledge to Instagram? The bottom navigation bar inspired a generation of apps.

Before we started writing code, we took a notepad in which we drew the entire application. We came up with the idea of ​​​​accompanying the photographs with a description, but after sketching them, we decided to throw everything away - because it wouldn’t work. We spent four hours on this, which may seem like a waste of time, but otherwise we would have spent over a week on development.

The most important moment in the development of early Instagram was when Kevin said, “I haven’t taken a vacation in two years, so I’m going to Mexico. There I will think about all these filters.” I said, “Okay, okay, let's finish all the app sketches, draw the interface, and design the user experience before you leave.” I had a very interesting time that week. I woke up around 2pm and programmed until 6am. I think this is my most natural state when I need to finish something.

A week later we already had a prototype of the application. We were able to do this without going astray because we already had ready-made interface layouts.

How did you invent filters?

Most of the first filters were made by Kevin. He simply looked through photo collections on Flickr and saw what happened as those photos aged. Or he even looked at our childhood photographs - they were only 20 years old, but they had already somehow changed, they had some kind of different mood. We tried to reproduce these effects in Photoshop and then translate them into code.

Over time we got better and better. I go back to my first photographs - and there were frames everywhere. And we had too much color grading. The current photographs are not very well processed, but they have some kind of mood that is not in the original picture. And they are not so heavily altered that they are unrecognizable.

So why did people stop using frames?

I think it's because at the end of the day you still want a photograph. Frames were a largely unintentionally amazing marketing tool in the early days. What kind of photo is this? Oh, this is a photo from Instagram, it has a frame. The New York Times once had a photo with an Earlybird filter on the cover. It was the coolest moment of all our appearances in the press. It's not even media coverage of us - it's just using Instagram photos. It was a Yankees player. You knew right away that it was Instagram by the Earlybird frame. But I think people are slowly moving towards a simpler rather than super editable version of the filter. Additionally, we turned off frames by default, which had a huge impact on users.

Can you give me an example of some crazy use of Instagram?

I love watching people use and adapt the product. Adrian Grenier took advantage of the fact that the profile photos are tiled: he made a mosaic out of them, laying out one photo at a time. It looks pretty cool, and then you realize that with the next photo you'll break everything. Previously, we had four photos in a row, but we wanted to make them more because it was impossible to see anything in such thumbnails. We enlarged them and left only three in one row, and then we started receiving all these letters: “Well, I did a mosaic!”

has a wonderful comic on this topic xkcd . When you have 100 million users, someone is bound to use your product in an interesting way. If you change the product... you'll break several use cases.

Why did it take so long to develop the web version? Did it meet your expectations?

There were already third-party sites that implemented this functionality, but each of them was special in its own way. There was definitely a need that these sites filled. We already know. It's funny to look at the logs and observe the behavior of users who received 404 errors as a result of a set of addresses like instagram.com/username. We are very pleased to have filled this gap. I'm really glad we made a mobile version initially, but I don't think it's possible to survive on just the mobile version in the long term.

All your breakfasts went through this man,
Konstantin Panfilov,
TJournal

#Article