OpenAI Targets Goblin, Gremlin Bugs in ChatGPT
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI rolled out a model update aimed at suppressing the “goblin”, “gremlin” and “troll” hallucinations surfacing in ChatGPT responses.
- ChatGPT users reported a surge of goblin references, prompting the company to prioritize the fix across its deployment pipeline.
- Codex will absorb 10 engineers from the Prism research group to focus on cleaning up the goblin/gremlin bug pipeline.
- Prism’s team is being trimmed as those 10 engineers are reassigned, shifting resources from experimental work to core product stability.
Why it matters: OpenAI users gain a cleaner ChatGPT experience while Prism engineers lose their current projects; the 10‑engineer shift accelerates bug fixing, reducing support tickets and preserving enterprise contracts.
