Schweickart Media
← How I use AI

Technology · Working practice

Connecting photographs, recordings and stories

The work between finding an image and using it to tell a particular story.

A photograph may preserve something I did not stop to describe. A recording may preserve something the photograph cannot show: why I was there, what I was considering or what happened before and after. I use AI to help bring those records back together.

My archive includes food, travel, home projects, family material and equipment. The goal is to make that whole collection useful over time, including material that has no immediate place on a website.

Describe what is actually in the frame

A useful image record goes beyond a filename. It describes visible subjects, objects, setting, activity and readable text, with tags that make the image easier to find again. Image quality also matters: the clearest view of a component may be a detail photograph rather than the widest shot of the machine.

Dates and filenames help narrow a search. Looking at the photograph establishes what it shows. Those are separate parts of the work.

Connect the records without inventing the event

A voice note about cooking and a food photograph taken nearby in time are candidates for the same account. They still need to agree. A photograph of pasta cannot establish that I made a particular saved recipe, and a picture outside a museum cannot establish which exhibits I saw inside.

I want captions to describe the supported connection. If an image provides general equipment context, the writing should say so. If it documents the actual meal or visit, that connection should be clear from the records.

Build a sequence around the experience

For travel, I want to see the trip: the setting, movement between places, distinctive details, meals and the stops that made it memorable. A single attractive view can leave most of that experience out.

The existing New Orleans, Alabama and Wrocław accounts show the direction I want to develop. Photographs belong beside the relevant part of the account, with enough variety to help a reader follow it. More images help when they add another scene or perspective.

Carry the evidence forward

The original photograph, its description, the proposed story connection and the finished caption each serve a different purpose. Keeping them connected makes a later correction easier. An improved caption should not require rediscovering the original file.

Finding and reviewing an image also does not make every use appropriate. Private family information and background details need attention before a story becomes public. The test version is where I can review the selection and ask for a better one.

More about how I use AI →

A good sequence respects the boundaries of an experience

The same archive can support several stories without turning them into one event. A photographed plate, a recorded recipe idea and another day at the stove may each be useful, but their value depends on keeping the experiences distinct. A food account becomes stronger when its pictures explain that particular preparation rather than supplying an attractive but unrelated finish.

Travel needs a similar discipline at a different scale. An establishing view helps place the reader, while a close detail can explain what caught attention at a stop. Repeating a wide view cannot do both jobs indefinitely. The edit should ask what each image adds and which scene still lacks a picture. That is where visual review and the spoken account become complementary: one shows the frame, and the other may explain why it mattered.