| Some thoughts: - Reminds me of a GitHub commentor saying they left tech to build furniture out of wood ("I no longer build software" [0]). - AI art might eat your lunch, especially in the concept art world. The main guy on Twitter who put out the thread about Stable Diffusion was a concept artist [1], and I think concept artists, graphic designers and generally non-fine-art artists will be hit the hardest. - The part about a more diverse player roster was interesting given that ATVI use a tool for it, although not for Overwatch apparently [2]. Some people were outraged online (on both sides of the political spectrum) but personally I don't see a problem with it. People may not necessarily remember the various axes on which others operate (race, class, gender, sexuality, etc), so there's nothing wrong with a tool that can help people analyze that and create characters for that. It reminds me of the outrage doctors first had over needing checklists for tasks, but they actually do work when studied. They're just tools, nothing more, nothing less. [0] https://news.ycombinator.com/item?id=24541964 [1] https://news.ycombinator.com/item?id=32461138 [2] https://www.pcgamer.com/overwatch-creators-explain-they-didn... |
A marketing person can't sit in front of Stable Diffusion and work it towards the campaign they want. They still need an art department that can iterate and make it match the campaign. The artist may very well be sitting in front of some iteration of stable diffusion/imagen/dalle but they still need to discern between AI generated outputs that work and that don't, and to iterate on it until it's right.