T
T
After submitting the following prompt: “Humorously render the image into a simplified cartoon character,” this is the resulting illustration.
To be sure, as I did not ask for a caricature, I was not surprised that the face did not resemble me. I also assumed that ChatGPT would zero in on my furrowed brow and knit eyebrows to produce an all too serious expression, which it did. So I prompted it to remove the stern glare and go a little lighter on the overall expression “Unknit the furrowed brow a smidge and make the face a tad more friendly and welcoming.”
The A.I. came back with this.
Okay, I suppose, but it lost the five-o’clock-shadow and white goatee in the process. On a side note I was intrigued by ChatGPT’s assumption that the sample was wearing a tee, which I was, but I just as easily could have been wearing a long sleeved sweatshirt. The cropping doesn’t permit one to know one way or the other. Ah, machine learning.
I then asked it to return the shadow.
Instead, next went and did this.
Uh-oh, I didn’t ask for a more realistic depiction. Notice the five fingers, instead of the cartoon’s standard four. Unfortunately, when I asked ChatGpt to simplify once again, the successive picture (not shown) only got further away from the original concept, that is to say the original intent. So, I ended the session. Maybe I should have prompted the A.I. to keep the style of the of the first two frames, but was already too late. Oh, well., on
T
T
T
T
T
No comments:
Post a Comment