Reading up on ai-image-generation

Manus Design View adds granular visual editing via a Mark Tool and mobile support, enabling multi-step AI design workflows without re-prompting.

Older AI image generation models produce more aesthetically pleasing, artistically coarse results than newer, precise models like Nano Banana Pro, according to the author.

OpenAI releases GPT Image 1.5, a cheaper image generation model that improves on its predecessor but still trails competitors like Nano Banana Pro in quality and consistency.

OpenAI releases GPT-Image-1.5, a new image generation model for ChatGPT and the API, claiming 4× faster generation with better editing and consistency.

AI image generators improve realism by mimicking the imperfections of phone camera photos, making fakes harder to detect.

Graph theory constraints in a prompt for six multi-pointed stars with non-intersecting connections cause all major image-generation LLMs to fail, while humans solve it in a minute.

Experiments with Google's Nano Banana Pro visual reasoning model show that structured, constraint-rich prompts produce better comic book adaptations than vague requests.

Google's Nano Banana Pro uses tool integration for factually accurate infographics and maps, but still has blind spots like circuit diagrams.

Hacker News commenters critique Google's Nano Banana Pro image model for vague rollout, SynthID watermark removability, and unreliable text and diagram generation.