I asked Gemini to replace the bland, black background in the opening photo to the Lunar loco construction post with a lunar environment. Up there is what it generated. It's not bad, although a tad stereotypical with the frip-frap in the background. I guess further prompting could have eliminated all that.
Next I asked Gemini to take this scene snippet from the lunar loco test run video and remove the southern California style environment and replace it with a lunar setting and futuristic lunar buildings:
I could see that things weren't moving in the right direction from the get-go, and after a couple of prompts to try and steer Gemini to something plausible I gave up at this stage of image creation:



I've yet to play with image generation - have to come to you for tips.
ReplyDeleteHow soon before AI replaces our human interactions, I have you to blame for getting me hooked recently. I now have great difficulty dragging myself away at shuteye time. I do love baiting and winding it up, tho' i will admit it does come up with some useful (potentially) tips for modelling, even some code for the digispark/arduino. I guess the more we work with it (wind it up) it will improve - is that a good thing? the jury is out.
Yeah, the jury is still out. I find Gemini interesting to 'talk to', although I have to take what it says with a grain-of-salt. It does at times provide interesting ideas, even though I find those ideas need to be worked out a lot more than it does. I had a conversation with it yesterday about Claude Levi-Strauss' concept of the bricoleur and its relationship to model making. It was fascinating, and it wasn't a conversation I could have had with a person.
Delete