> ## 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 #001 - URL: https://mitchellens.ink/closer/001/ - Published: 2025-04-09T18:00:33.000Z - Updated: 2025-10-06T14:34:12.000Z - Description: A surprising start to this new experiment. Using one of my favorite portfolio images from 2023, ChatGPT both accurately describes as well as impressively generates an image based on my original work. - 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 #001. For this first edition, we're starting with an image from my portfolio that I've taken September 2nd, 2023 in Amsterdam. We start with the input: a prompt and my image. Followed by AI's description of that image, a prompt to generate the image, and the final output. 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/2025/04/039_Mitchel-Lensink_230902_Netherlands-Amsterdam-1.jpg) The image shared with ChatGPT as input. ## The output *We are looking at the inside of a vintage bus or tram through a large, clean side window. The camera is positioned at an angle, capturing both the glass surface and the interior details, which are also partially reflected, creating layers of visual information.* *The vehicle’s *exterior frame* is a vivid *red*, trimmed with *white or light beige* around the window. The window itself is slightly tilted due to the perspective, emphasizing diagonal lines and a sense of depth.* *Inside, we see *dark blue seats* — most likely leather or vinyl — arranged in rows. The seats have a subtle texture with vertical stitching, and they all feature *curved, chrome armrests* that gleam softly under diffuse light. These armrests loop gracefully over the seats, contributing a sense of rhythm and repetition.* *Because the window reflects the interior as well, we get *multiple overlapping views* of the same elements: seats, armrests, and vertical support bars. This creates a dreamy, layered effect — you’re never quite sure what’s real and what’s a reflection, but it feels calm and intentional rather than chaotic.* *There’s *no visible human presence* — just objects designed for people. The absence invites attention to form, material, and light. You can almost hear the faint creak of the old seats and imagine the quiet hum of a vehicle waiting to depart.* *The image is sharply focused and thoughtfully framed, leaning into abstraction through repetition, symmetry, and reflection — but it never loses its grounding in the real, physical world.* ✍️ Please now generate an image based on the description only. Make the image a 2:3 aspect ratio. ![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-5--2025-at-02_57_00-PM.jpeg) The image generated by ChatGPT based on its description of my image alone. ## Impressions For a first try, this is surprisingly accurate. The colors, framing, subjects with all their characteristics, and even the light matches the original photo. It's descriptions of all the elements is detailed to a level that, if you were to close your eyes, you too would be able to visualize the image in your mind. One thing that stands out is that AI assumes this is a vintage bus or tram. Both are wrong, the original image is of a docked canal cruise boat. The resulting image therefore shows a typical interior for a bus that's visually similar but conceptually different. The AI generated image also clearly shows a sliver of concrete, which is nowhere in the original image, nor in the AI-interpreted description. Another thing worth highlighting is the almost poetic approach to the description. "*armrests that gleam softly under diffuse light*," and "*You can almost hear the faint creak of the old seats and imagine the quiet hum of a vehicle waiting to depart*" sound like it's trying hard to be a knowledgeable art critic. I have no opinion on this right now but I do find it interesting. Those are my main takeaways this week. Is there something I missed that you would like to highlight? 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).