I used Claude to generate the prompt that i entered into Dal-e, learned in webinar Steve Little. My husband paints images on boards from mid-century modern furniture which is then sold in my son’s business. this is the prompt from Claude. The next prompt included different architects. Came out good.
Create a Photorealistic architectural rendering of a classic mid-century modern home nestled in a secluded woodland or desert landscape. The house features clean horizontal lines, floor-to-ceiling glass walls, and a low-pitched butterfly or flat roof. Exterior materials include warm wood paneling, concrete, and large glass expanses. Color palette inspired by 1950s and 1960s design: sage green, burnt orange, mustard yellow, warm teak brown, and slate gray. Integrate the home seamlessly with the surrounding natural environment - either positioned among pine trees with dappled sunlight or set against a stark desert backdrop with minimal landscaping. Emphasize architectural details like cantilevered sections, exposed structural elements, and minimal ornamentation. Lighting should be soft and natural, highlighting the home's geometric precision and harmonious relationship with its environment. Perspective from a slight angle to showcase the home's integration with the landscape, with mid-century modern furniture visible through the glass walls to reinforce the design aesthetic.
Sometimes I think ChatGPT plays games with me. When I was generating an image of two men playing cards in a tavern recently for one of my posts it returned a Pic that was perfect in every detail except the cards in one man's hand were back to front. He was looking at the back of them. I asked it to turn them round but leave everything else the same. It did but one man had cards with a blue back like the stack of cards on the table and the other (the one it turned around) had a red back. I asked it to fix that and leave everything else the same but in addition to doing that it added a ridiculous hat to what of the men. Feeling slightly exasperated I gave up for the day. I finally got something acceptable the next day.
Thanks @DearMyrtle. I am not sure what version I have on my laptop but I will check. Generally, I have been happy with what it has produced but on the particularly occasion that I mentioned it was being unusually mischievous.
I must try some of my images for restoration. I'll be sure to mark the AI generated, but it has been interesting to watch the prompt developments on the Facebook group.
Interesting! I like the last one, which really looks like a photograph. I've tried a few times but didn't like my results. I think I'll go play a bit with it now.
I used Claude to generate the prompt that i entered into Dal-e, learned in webinar Steve Little. My husband paints images on boards from mid-century modern furniture which is then sold in my son’s business. this is the prompt from Claude. The next prompt included different architects. Came out good.
Create a Photorealistic architectural rendering of a classic mid-century modern home nestled in a secluded woodland or desert landscape. The house features clean horizontal lines, floor-to-ceiling glass walls, and a low-pitched butterfly or flat roof. Exterior materials include warm wood paneling, concrete, and large glass expanses. Color palette inspired by 1950s and 1960s design: sage green, burnt orange, mustard yellow, warm teak brown, and slate gray. Integrate the home seamlessly with the surrounding natural environment - either positioned among pine trees with dappled sunlight or set against a stark desert backdrop with minimal landscaping. Emphasize architectural details like cantilevered sections, exposed structural elements, and minimal ornamentation. Lighting should be soft and natural, highlighting the home's geometric precision and harmonious relationship with its environment. Perspective from a slight angle to showcase the home's integration with the landscape, with mid-century modern furniture visible through the glass walls to reinforce the design aesthetic.
That prompt conjures a beautiful image in my mind. My prompts tend to be less detailed.
Interesting that he gave AI the choice of environmental settings. I’d pick the pine trees and dappled sunlight over the desert scene.
I've had a lot of fun generating images. I don't use all of them, but it's interesting and sometimes scary to see what AI can produce.
Part of why I keep taking AI courses is so I can distinguish AI from reality. Now with some AI we can dictate rather than type our prompts.
We are nearing the “Open the pod bay door Hal” era in reality.
Sometimes I think ChatGPT plays games with me. When I was generating an image of two men playing cards in a tavern recently for one of my posts it returned a Pic that was perfect in every detail except the cards in one man's hand were back to front. He was looking at the back of them. I asked it to turn them round but leave everything else the same. It did but one man had cards with a blue back like the stack of cards on the table and the other (the one it turned around) had a red back. I asked it to fix that and leave everything else the same but in addition to doing that it added a ridiculous hat to what of the men. Feeling slightly exasperated I gave up for the day. I finally got something acceptable the next day.
Did you try ChatGPT4o ? It’s pretty good with images.
Thanks @DearMyrtle. I am not sure what version I have on my laptop but I will check. Generally, I have been happy with what it has produced but on the particularly occasion that I mentioned it was being unusually mischievous.
The photo came out great!
Thank you for your detailed prompt. I borrowed it, made some changes, and this is what I got.
An image I could use in my blog post. https://mytrailsintothepast.blogspot.com/2025/06/my-favorite-names-to-test-out-online.html
Your AI-created image is beautiful! BRAVA! 🤗
I must try some of my images for restoration. I'll be sure to mark the AI generated, but it has been interesting to watch the prompt developments on the Facebook group.
Interesting! I like the last one, which really looks like a photograph. I've tried a few times but didn't like my results. I think I'll go play a bit with it now.