I've been spending a lot of time gathering detailed research for a McKinsey-style report using ChatGPT. I crafted a thorough prompt based on a template I found online, and ChatGPT initially estimated it would take about three days to produce the report. I was fine with that timeline and didn't rush it. However, after waiting three days, all I got was a one-page PDF containing a brief five-sentence summary, despite asking several times for updates. ChatGPT kept insisting it needed 'six more hours' or 'three more hours' before it would be ready. A week later, and still no complete report! I tried the same prompt in a new chat and got the research in text format within half an hour. But again, when I requested it in a downloadable PDF, I had the same frustrating result – just a few sentences. Am I missing something in the prompting, or is there something else going on? It's not an issue with the prompt since I got it from a deep research expert on Reddit.
3 Answers
Honestly, it sounds like whatever issue you're having isn’t just about the prompt. If you’re getting decent output in one chat and a lousy summary in another, maybe it's a glitch or just how the system is responding at the time. When dealing with AI, sometimes you just have to roll with the punches and try again. Wouldn't hurt to reach out to support too if it keeps happening, just to see if there’s anything they can do about it.
I feel for you! It can be super annoying when you’re waiting for something that never seems to arrive. You’ve got to remember that if it’s consistently producing the short summaries, chatbots can sometimes just improve on the fly with different prompts. Maybe try tweaking your prompt a bit or asking it to elaborate on specific sections instead of asking for a full report all at once? That way, it might start to piece together a bigger picture for you.
From what I understand, ChatGPT’s time estimates can be pretty random. It doesn’t actually know how much server time it has, so when it says it’ll take three days, that’s just based on a statistical guess. The timing doesn't really mean much since the processing power can vary greatly. It's frustrating, I get that, but maybe it’s just overestimating its capacity? You might want to just keep prompting for the report until it gives you the solid details you need.
Yeah, that makes sense, but the issue is more about it promising a complete report and just delivering a vague summary every time. That's what's really getting to me.