> ## Content Index > Fetch the complete content index at: https://mitchellens.ink/llms.txt > Use this file to discover other available public pages before exploring further. # Closer #004 - URL: https://mitchellens.ink/closer/004/ - Published: 2025-04-30T18:00:12.000Z - Updated: 2025-10-06T14:34:00.000Z - Description: On the surface, things seem to be looking good. But look more closely, and you'll find that fiction is, thankfully, still stranger than reality. - Author: Mitchel Lensink - Tags: Closer, Field SPONSORED This weekly experiment is on an indefinite hold, until significant AI-advancements have been made. Curious about the conclusions? Read [Photography in the Age of AI (2025): Lessons From 25 Weeks of Experimentation](https://mitchellens.ink/closer/photography-in-the-age-of-ai/) [Read now ](https://mitchellens.ink/closer/photography-in-the-age-of-ai/) Welcome to Closer, a weekly newsletter where I take one of my photos, ask AI to describe it, then see what kind of image it can generate from its own description—an ongoing experiment in the capabilities of AI, and what that means for humans. This is edition #004. During a trip to London, back in 2022, I created an image with a well-defined foreground, middle, and background, where many elements come together to create one chaotic cityscape. As always it goes: input image & prompt to describe → AI's description of that image → AI generated image → some of my first impressions. Prompts are in the grey callout boxes, the AI description is *cursive.* ## LLM models used **GPT-4-turbo** for image interpretation. **DALL·E 3** for image generation. ## The input ✍️ Please describe the attached image as well as you can. Don't skimp out on details and make sure you will be able to understand what's in the image when you read the description. You need to be able to form a mental picture of the image, without actually seeing it. ![The image shared with ChatGPT as input.](https://storage.ghost.io/c/8a/15/8a1564aa-a008-4e21-8398-12d04de07fcb/content/images/2025/04/011_Mitchel-Lensink_220805_United-Kingdom-London-1.jpg) The image shared with ChatGPT as input. ## The output *This photograph, taken in daylight under partly cloudy skies, captures a lively urban scene in London, centered around a stepped concrete plaza or amphitheater-style seating area near the River Thames. The image is filled with a sense of summer relaxation, socializing, and urban vibrancy.* ### *Foreground & Middle Ground* *The lower half of the frame is dominated by the large, gray stone seating steps that rise gradually from the bottom of the image. These steps are broad and serve as informal seating rather than formal stairs. Scattered across them are many people of all ages and backgrounds—some sitting alone, others chatting in pairs or small groups. Many appear relaxed and casual, dressed in light summer clothing like shorts, skirts, T-shirts, and sunglasses. Some are eating, talking, or simply lounging in the sun. A few are looking at their phones, and one or two appear to be deep in conversation.* *The next level up from the steps holds rows of white picnic tables beneath red Coca-Cola-branded umbrellas. These tables are more densely populated, and the crowd there is a bit more active—standing, walking, or queuing at a nearby vendor.* ### *Background* *Behind the seating area is a small, pastel-colored refreshment stand or kiosk. It’s decorated in soft blues and pinks with simple geometric patterns, and there’s a stylized smiley face on one side. The booth has an open window where a few people are being served. Flanking the kiosk are artificial palm trees, adding a playful, almost beachy vibe to the urban setting.* *Towering behind this festive lower level is the dramatic skyline of London’s financial district. The cluster of high-rise buildings is immediately recognizable:* - *To the left is the iconic Walkie-Talkie building (20 Fenchurch Street), known for its top-heavy, curved design.* - *To the right of that, you can see the Leadenhall Building (aka “The Cheesegrater”) with its diagonal cuts and sharp lines.* - *Cranes are scattered among the skyscrapers, suggesting ongoing construction and a city still expanding vertically.* - *A more modern, glassy residential or office building curves horizontally in front of the towers, creating a layered architectural effect.* ### *Sky* *The top third of the image shows a partly cloudy sky with soft, puffy cumulus clouds scattered across a bright blue canvas. The lighting is even and bright, casting soft shadows but no harsh contrast, suggesting it was taken around midday or early afternoon.* ### *Overall Mood & Composition* *The photo juxtaposes urban structure with human leisure, blending hard concrete, steel, and glass with warm, social human presence. The layers—from the people on the steps, to the colorful kiosk, to the distant skyscrapers—create a sense of depth and liveliness. It’s a celebration of public space: casual, bustling, and distinctly summery.* *In summary, this image portrays a relaxed summer day in a metropolitan setting—probably along London’s South Bank—with people enjoying the city while surrounded by iconic architecture and soft sunlight.* ✍️ Generate an image with an aspect ratio of 3:2 based on this description: {**the description previously generated*} ![The image generated by ChatGPT based on its description of my image alone.](https://storage.ghost.io/c/8a/15/8a1564aa-a008-4e21-8398-12d04de07fcb/content/images/2025/04/ChatGPT-Image-Apr-12--2025-at-10_56_29-AM-Large.jpeg) The image generated by ChatGPT based on its description of my image alone. ## Impressions Wow, that's an epistle of a description. It's much longer and more detailed than previous entries. This one even has headers and lists in it. Not sure where those come from. I've been asking for image descriptions in the same chat though. Do you think previous prompts in the same chat are influencing new prompts? I've also been generating my images in the same chat, albeit a separate one from where I share the input images (to ensure it definitely doesn't peek at the original when generating a new image.) Should I open a new chat for each image description and image generation? I did ask for another image description in a new chat and it presented me with a result much more akin to what we've seen in previous weeks. But you know the rules: one try, no tweaks. This is what it spat out. Some things I noticed: - It correctly recognized London's skyline. Judging by the list it produced, it knows *a lot* about that skyline, too. So much, in fact, that it has taken some liberties and fully depicts *The Gherkin*, even though that's almost completely obstructed by a different building in the original image. - There's no smiley face on the pastel-colored building (but I can see what element it believes to be one) - The umbrellas have a similar red as Coca-Cola but they are of a different brand. Then again, the output image does not show Coca-Cola-branded umbrellas either. Overall, the balance of the layers in the images are properly depicted but it took some significant liberties with the elements and their placements. I also think the detail on the benches and the buildings in the back are... not great. Very AI-like. Are we keeping score? Because in this case I'd give the point to the real life image. Anything else that caught your eye? Feel free to send a reply. Please also let me know if you have questions, suggestions or remarks about my approach. This is just as much a journey for me as it is for you. See you next week. Mitch 👤 Tip: you can easily [manage your newsletter subscriptions](https://mitchellens.ink/#/portal/account/newsletters) and [membership status](https://mitchellens.ink/#/portal/account/plans) from [your account page](https://mitchellens.ink/#/portal/account).