ProfGrizzlyJon
Cosmic Mongoose
This is fascinating, because the discussion is taking extremes. Yes, we largely all agree that purely generative AI is soulless and not good for the industry (cue any of the thousands of Youtube videos on how to create a low-effort AI side-hustles for details).
I use a specially-trained LLM-enhanced tool as a "second-brain" that I have built over the last three years, while using Obsidian as an editor and a research assistant. I painstakingly fed it my paper notes and drafts, my equivalent of a showrunners' bible. When I ask it to help me edit copy, especially with connecting disparate items, how is that different from asking an editor for assistance? The work is still my own.
I used AI to write python code that created the Jumpspace Weather phenomenon heatmap that I then underlaid on TravellerMap; this took many additions to get it to look right, and weeks of late-night effort.
After watching Bakshi movies as a kid, I used to use a light table and dabble in "rotoscope" animation. The end result was horrible. Now, have been able to take screen captures of designs that I made using 3D models, and then run them through an AI-assist to add texture and background in a manner that takes my boring grey Meshmixer kitbash into something colorful and exciting (see attached for my version of the Type-A+ Free Trader Plus).
Yes, for some character images, I have fed an image-generator paragraphs of description, and then run through multiple edits and revisions until I get the look right. I have also used pictures that I have taken (especially from my time in the military) and had them retextured/reimagined into something suitable for sci-fi.
Shucks, my profile image was a selfie of me diving, that I asked an AI to reimagine as an astronaut.
These have a place in the market. I don't mind letting someone know that I used the tools, but don't ban people like me from participating.
I use a specially-trained LLM-enhanced tool as a "second-brain" that I have built over the last three years, while using Obsidian as an editor and a research assistant. I painstakingly fed it my paper notes and drafts, my equivalent of a showrunners' bible. When I ask it to help me edit copy, especially with connecting disparate items, how is that different from asking an editor for assistance? The work is still my own.
I used AI to write python code that created the Jumpspace Weather phenomenon heatmap that I then underlaid on TravellerMap; this took many additions to get it to look right, and weeks of late-night effort.
After watching Bakshi movies as a kid, I used to use a light table and dabble in "rotoscope" animation. The end result was horrible. Now, have been able to take screen captures of designs that I made using 3D models, and then run them through an AI-assist to add texture and background in a manner that takes my boring grey Meshmixer kitbash into something colorful and exciting (see attached for my version of the Type-A+ Free Trader Plus).
Yes, for some character images, I have fed an image-generator paragraphs of description, and then run through multiple edits and revisions until I get the look right. I have also used pictures that I have taken (especially from my time in the military) and had them retextured/reimagined into something suitable for sci-fi.
Shucks, my profile image was a selfie of me diving, that I asked an AI to reimagine as an astronaut.
These have a place in the market. I don't mind letting someone know that I used the tools, but don't ban people like me from participating.