Morfternight #101: Photo gear, GPT agents, and storytelling.

Morfternight is your weekly dose of photography, leadership, and AI.

Welcome to Morfternight, your digital postcard about Photo, Leadership, and AI.

“Good Morfternight” was born out of working with teams distributed across time zones. It’s how we unite Good Morning, Good Afternoon, and Good Night in a unique greeting, including everyone in the same smile. :)

---

👇 tl;dr

Today, we dive into my catalog to analyze the EXIF data in my photos and understand why I am not happy with my photos anymore. We’ll briefly consider that the author is not getting any younger, which had been so far hidden by autofocus. We introduce the Morfternight community to my new team of GPT Agents, and we discuss briefly the importance of storytelling for humanity and the impact LLMs can have on it.

👋 Greetings from Vienna!

From Malaga to Hamburg, back to Malaga, then Costa Rica, and finally Munich, my travels have been non-stop since October. Now, with my Rimowa finally at rest, I'm home, catching my breath in the rapidly advancing world of AI. No more excuses for delaying Morfternight — it's time to dive in.

Despite a recent lull in photography, I'm not about to let that hinder our discussions. Instead, I'm turning this into an opportunity to delve into the reasons why I have no new photos to show. There's a lot to unpack, and I'm here to break it down for you.

Let’s explore these developments together in this edition of Morfternight, where we cut through the noise to focus on what truly matters.

📷 A Postcard from… Vienna

Volkstheater U3, Vienna — More photos

As I said, I haven’t made any good photos lately, so here’s one I took on April 1st, 2019.

I remember it particularly as it was the first one shot with a brand-new lens, the Fujinon XF 16mm f/2.8 that I had just purchased and mounted on the Fuji X-H1 I had back then. Like many photographers, one of the reasons I make photos is that I like cameras and lenses, incredibly sophisticated tools.

The problem with the passion for the gear is that the camera and lens must become almost an extension of the photographer’s hand and eye. That’s how I repeatedly end up in the same routine: the new camera I just picked up is perfect, and I’ll never change again. Until, eventually, enough time passes for me to rationalize the next upgrade. Rinse, and repeat.

So, today, I decided to dive into my photo catalog. Using the EXIF data (which stands for Exchangeable Image File Format, basically the information stored with each digital photograph about the camera, settings, and more), aggregate the information I have available about the last seventeen years or so of photography. The EXIF data aggregates the information I have available about the last seventeen years or so of photography, since digital cameras became serious tools.

Breakdown by camera

The table below shows the cameras I owned since 2006. The Start and End dates come from the first and last photo in my catalog that was shot on each camera.

The number of photos is not the number of shots (that would be much bigger), but the number of photos I considered good enough to keep.

The first thing that is immediately visible is that there are three primary periods:

• 2006-2014: The first eight years account for 78% of the photos. Looking at them, the explanation is simple, the quality bar for keeping them was much lower back then.

• 2015-2020: Very few photos during these five years, only 8% of the total. Not randomly, this is the time when the iPhone camera started getting good, and for a while, I focused primarily on using it.

• 2021-2023: On January 1st, 2021 I decided I was going to take photography seriously again. The numbers may seem lower, but they don’t tell the full story. The reality is that I raised the bar much higher in terms of quality, nowadays, I only keep the photos I publish.

Focusing only on the recent years, I find some interesting, and surprising, insights. What's your experience with different camera models? Do you find yourself gravitating towards certain types over others? I'd love to hear your thoughts and experiences in the comments or in your replies.

First, I would have expected these numbers to tell the opposite story. The Q2 and the M11 are by far easier to carry around. The former doesn’t have an interchangeable lens, the latter does, but I only possess one, the Summilux 50mm f/1.4. They are both compact and light.

That’s the thing, though, I do carry the M11 almost everywhere I go, as I did for each of the three other cameras, so that’s not a differentiating criterion justifying the results.

If anything, I would say that paradoxically, the size and weight played in favor of the much heavier SL2-S. I find carrying a large, heavy, camera strapped around my neck or my shoulder very uncomfortable, which is why I had a tendency to always have the SL2-S in my hand.

Breakdown by focal length

To dig deeper, I checked the distribution of images by focal length.

It’s important to note that the Fuji cameras have an APS-C sensor with a crop factor of 1.5, so the focal lenses should be multiplied by 1.5 when used on such cameras. This affects the 35mm and 50mm photos: with the former all made on Fuji APS-C cameras, and the latter all on Full Frame sensors, they are de facto one and the same, with a total of 1749 photos.

The 85mm and 90mm are very close to each other on Full Frame sensors, but the 90mm on the Fuji APS-C sensors is equivalent to a 135mm in Full Frame. Of the images shot with a 90mm, 108 were made on the SL2-S (Full Frame), and 472 on APS-C sensors.

Looking at these numbers, the first obvious conclusion is that I enjoy using lenses with a focal length of 50mm or more, as they represent the vast majority of the images I keep.

This is the problem with the Leica Q series and the Fuji X100 series. All incredibly good cameras, pleasant to use and very high quality, but stuck with respectively a 28mm and a 23mm (equivalent 35mm) lens.

So, what’s the issue with the M11?

Simply put, it’s manual focusing. I am not 25 anymore, and my reflexes (and let’s be honest, also my eyesight) are not what they used to be. Not only, but I recently realized that I tend to give up without even trying. If an interesting subject in the street is moving, I won’t even take the shot, knowing I’ll miss the focus anyway.

Of course, I could use zone focus.

Zone focus is a technique based on depth of field, where you set the lens to focus at a certain distance and adjust the aperture to get a range of focus. For instance, by setting the lens to focus at 3 meters, then closing the diaphragm to 8, anything between 2.36 and 4.13 meters from the camera is in focus. The issue is that the whole reason I like longer lenses is the bokeh I can obtain by opening them wide. In the same configuration, open at 1.4, the depth of field only goes from 2.87 to 3.15 meters. This now requires being a lot more precise.

I have been trying to get used to measuring distances at a glance, but so far, I haven’t reached great results.

Conclusion

A part of me, that just wants to make good photos, wishes I didn’t listen to the other part that likes trying new cameras. The X-Pro3 was already the one I was getting the best results with, and had I kept it since 2021, my familiarity with it would only have increased, making it easier and easier to achieve the results I want. However, the sheer numbers are not telling the whole story. When I look at the images I prefer, the distribution is almost even across the X-Pro3, the Q2 Mono, and the SL2-S. Once more, the M11 is far behind.

As I continue to explore and grow in my photographic journey, I look forward to sharing new perspectives and breakthroughs with you. The path of creativity is ever-evolving, and each step brings new insights and joys. Stay tuned for what's next on Morfternight, and let's keep the conversation going!

Meanwhile, if you are facing similar questions, I can only encourage you to also look at the EXIF records of the photos you prefer, draw your own conclusions. If you feel like it, share them with me, I’d love to hear other points of view and opinions!

🤖 Much ado about nothing

A couple of weeks ago, OpenAI announced a new ChatGPT capability, allowing anyone to create customized GPTs to perform specific tasks without having to instruct them each time.

In the short time since the announcement, OpenAI’s board of directors fired the company’s CEO, named an interim CEO, fired the interim CEO, hired a new CEO, got fired along with the new CEO. The old board was replaced by an (almost) new board that re-hired the original CEO.

🦾 Here come the GPTs

During the same time, I started experimenting with the new GPTs, and I now have a small team of AI agents to help me:

• Agent Schedule — Has access to my calendars via the Google Calendar API, and can list the events of the day or the week, summarize my week, identify the busiest day, things like that.

• Agent Notes — I can upload a photo of a page of my paper notepad, it will transcribe it, summarize it, and provide both to me.

• Agent Tana — Works like ChatGPT, but I can ask it to send to Tana, my Personal Knowledge Management (PKM), any part of my chats by using the Tana API.

• Agent Feature — I give it the address of one of my posts on the blog, and it creates a featured image to go along with the post, based on its content.

• Agent Morfternight — It has access to the archives of Morfternight, so I can ask whether or when I already mentioned a subject, what I said about it, and use its help to make sure I don’t ramble inconsistently :)

Creating GPTs is quite straightforward, but if you think it would be useful for me to describe step by step how I created these, please let me know, and I’ll gladly share in the next issue of Morfternight.

📚 Storytelling

Once upon a time…

This (or it’s many variations in other languages), is how stories start when we are kids.

As children, stories are how we learn about the world, about good and evil, about morals, ethics, honesty, and all other values we need to understand so we can live in our society.

We consider that our civilization started 5,000 years ago when we invented writing.

We consider the printing press one of the most important inventions of our society.

Telegraph, telephone, radio, television, computers, internet, smartphones: all modern means of communications, all about telling stories.

We know that the past doesn’t exist, and neither does the future, but stories tell us about both.

History is written by the victors. The future by writers, politicians, marketers.

Teaching, selling, fundraising, building products, campaigning are all based on storytelling.

Politics, the economy, peace, and war all rely on someone telling a story and whether others believe in it or not.

Artificial Intelligence, Machine Learning, and more specifically Large Language Models, will play a crucial role given their ability to generate, transform, and summarize text.

Even if they didn’t represent a step towards Artificial General Intelligence, or Superintelligence, LLMs translate our stories, connect them, make sense of them, repeat them, improve them, and help us write more of them.

In this context, they are the most important tools of our generation.