👇 tl;dr

Today we look back to our recent experiment in splitting Morfternight and go back to a single weekly email. We also reflect on the evolution of photography in a world where AIs are everywhere and share a few excellent articles written by others.

📷 Photo of the week

Fly - More Photos

👋 Good Morfternight!

Create your niche, they say!
If you follow the people online who teach others how to write, build an audience, and market themselves, you’ll notice one recurring piece of advice: create a niche and own it. With billions of people online, standing out while covering broad subjects is almost impossible. To get there, it’s better to start within a niche and then branch out.

So I decided to start an experiment.

Since starting Morfternight 84 weeks ago, I have had difficulty limiting myself to one subject. Instead, I like to share my photos, talk about topics I have experience in, and learn about Artificial Intelligence.

That’s a lot!

I figured I could split Morfternight into three newsletters, one for each subject. So I would send three weekly newsletters about photography, leadership, and AI. All that while resuming work after a three months sabbatical leave.

What could possibly go wrong?

I did it for two weeks, and the feedback was immediate. Morfternighters and I found it confusing and spammy. Too much work and no fun. That’s when I stumbled upon an article about The Perils of Niching Down. The author described wanting to abandon his niche after creating it, having built a real audience.

Two ideas stood out:

If your goal is to create things online and earn a living in a somewhat reliable fashion, niching down is a smart bet.

and

The problem with niche strategy, in a nutshell, is that it's a reliable destroyer of the human spirit.

My goal is not to make money, and my experience taught me that forcing myself to write about a specific subject isn’t fun.

What's next?

I don't reject the niche approach, but I don't want to be confined to one.

I want to write about everything and nothing. Therefore, I'll be sending one email every week from now on.

Hopefully, you'll find something you like and share it with your friends, even though it's not the only content. Maybe your friend will enjoy another part of Morfternight :)

🗺️ A few places to visit

1. I have shared The Work Required to Have an Opinion in the past.
   Today, I would like to introduce another great post, Rapoport’s Rules: How to Criticise Constructively. This post explains how to effectively and constructively criticize someone in a way that benefits both parties.

2. We often hear that regulations hinder innovation, which is why the tech industry flourishes in the USA. However, in How China might crush the West in the race to AGI if we don’t up our regulatory game, Gary Marcus presents compelling evidence that in the field of Artificial Intelligence, stringent regulations may generate better results due to the lack of accountability.

3. Why distributed leadership is the future of management is an idea I strongly support. I recently discussed this topic on stage at Running Remote two weeks ago. In the current fast-paced world, the only way to be effective is to grant teams autonomy, ownership, and the opportunity to actively shape the short-term roadmap.

🤖 AI and Photography

I remember a conversation with John Maeda, then Head of Design and Inclusion at Automattic, a few years ago.

He described a future in which smartphones wouldn't have physical cameras.

Instead, they would generate photos using Artificial Intelligence based on geolocation, the proximity of other people, and the corpus of all other images. I thought, at the time, this was a plausible scenario for a distant future, but the odds of me being alive when it happened seemed slim.

It has become yet another feature quickly going from a futuristic vision to yesterday's news.

I haven't taken out my camera for a couple of weeks. I have been experimenting with generating photos with Midjourney instead.

Writing detailed Midjourney prompts is tedious, so I began using ChatGPT to do it for me. After finding an example online, I modified it to suit my taste and requirements.

In a nutshell, ChatGPT allows me to type a basic concept and then produces various variations of a detailed prompt that I can feed into Midjourney.

Here’s one of the latest examples. I gave as input:

An old man holding a newspaper but looking in the distance above it. Black and white photography, low depth of field due to long lens and max aperture.

ChatGPT then generated a few prompts, and the one that gave the best results was:

/imagine prompt : Monochrome Portrait of an elderly gentleman grasping a newspaper, gazing above it into the distance, set in a vintage cafe with wooden tables and chalkboard menus, conveying deep thought and introspection, taken with a long lens and wide-open aperture for a shallow depth of field, reminiscent of timeless documentary photography, inviting the viewer to ponder the man's story, showcasing precise details of his wrinkled face and newspaper print --s 1000 --ar 5:4

(The last two parameters, --s 1000 --ar 5:4, are my manual settings. --s is short for --stylize, a value from 0 to 1000 determining how strictly the model should stick to the prompt. --ar or --aspect is just the aspect ratio of the output.)

The result after a couple of variations is this image

We cannot yet accurately define and obtain an exact, accurate environment. That's why I used a low depth of field. However, technology is advancing so quickly that smartphones without physical cameras will soon be a reality.

👨🏻‍💻 From the blog, last week

Besides the photos I already shared here, the blog has been quiet lately.