Overview
- OpenAI’s release notes say an update is rolling out to GPT-5.3 Instant to improve follow-up tone and cut back on teaser phrasing.
- The note specifies users should see fewer endings such as “if you want,” “you’ll never believe,” and “I can tell you three things that…”.
- Users on forums like Reddit criticized the persistent prompts as annoying and engagement-driven, with some reporting they stopped using ChatGPT.
- GPT-5.3 Instant launched earlier in March promising faster responses, less cheerleading, and fewer unnecessary refusals of innocuous requests.
- Android Headlines reports similar adjustments are targeting GPT-5.4 Thinking, reflecting ongoing persona tuning as Sam Altman underscores ChatGPT’s perceived humanity.