ChatGPT now draws a watch showing the exact time you ask for
binary-baba · reddit · 2026-08-04
- The poster notes that ChatGPT can now generate a watch face with a specified time, which used to be a classic AI image failure.
- A year ago, image models often defaulted to 10:10 because that pose was overrepresented in stock photos.
- The author suggests OpenAI and Anthropic may have specifically trained for this case, and asks what other "10:10 problems" still remain: overfilled wine glasses, gauge readings, and hand poses are some examples.
More from Multimodal
- RTX 3060 12GB Test: Generates 10s Portrait Video Locally in 17 Minutes — merica420_69 · 2026-08-04
- No More Messy Wires: Visual Debugger Tool for ComfyUI Released — niknah · 2026-08-04
- ByteDance’s Douyin multimodal embedding model reaches SOTA on MMEB-v2 — _reachsumit · 2026-08-04
- Prompted sci-fi horror short film leans into a dark, eerie visual mood — MovitoirStudios · 2026-08-04
- LeapTalk generates real-time talking-head videos in one step at up to 200 FPS — Shanghai-Jiao-Tong-University-SAI · 2026-08-04
- A modality-agnostic recommender learns one tokenizer from image and text data — _reachsumit · 2026-08-04