ChatGPT Became So Obsessed With Goblins That OpenAI Had to Intervene - WSJ
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI shifted ten Prism engineers to Codex to curb the goblin‑heavy output.
- ChatGPT began spewing goblin references across user prompts, prompting complaints on multiple platforms.
- Prism’s “nerdy personality” was identified as the root cause, leading to a model‑tuning effort.
Why it matters: OpenAI’s redeployment of ten Prism engineers to Codex will curb goblin‑laden output, improving enterprise user experience, reducing support tickets, and protecting OpenAI’s brand reputation in the short term.
