I've been using O3 to help with constructing a literature review, particularly to include APA style in-text citations along with the hyperlinks required by OpenAI. However, I've experienced some major issues. O3 refused to follow my specific instructions to synthesize information, opting instead for a block-by-block approach, which is frustrating. I understand that there are baseline instructions from OpenAI, but it makes me wonder what the benefit is of these standards when they ignore personalized user needs. Additionally, I've noticed a drop in O3's performance; it seems to be making repetitive mistakes and not properly addressing queries. This change happened overnight—what's going on with it?
2 Answers
I totally get where you're coming from. While the earlier versions like O1 and Deep Research offered in-depth responses, O3 and the new iteration seem to churn out short, under-researched snippets. It’s disappointing and frankly, they’re becoming unusable for serious work.
It might be that O3's goals aren't aligning with what you're expecting in terms of research outputs, or maybe there's just a lack of reliable data for it to pull from. If you have specific websites that contain reliable data, feeding that info to the AI could help improve its responses.
The sites I'm using are definitely reliable; I've verified them myself. The real issue is with how it inserts the citations in APA style along with the hyperlinks. This was working fine just yesterday, so I’m really at a loss!