Ever since OpenAI started including Deep Research in with the normal $20/month paid plan (albeit with a limit of 10 uses), I’ve been crunching the numbers internally about whether it’s worth it to keep my upgraded $200/month plan that comes with 120 uses. That’s a lot of Deep Research queries, and the other day I thought, “I really haven’t been using it much the past week or so.” Granted, part of that is because I’m still recovering from my trip to Singapore, but still.
Then, funny enough, yesterday I ended up using it five times in a day. Guess I still need it after all! A couple of those were traditional Deep Research usage, but two of them were in a way that I hadn’t thought about using it before.
I used it to find something that I remembered part of, but not the full story. There’s a story I often tell when I’m teaching, and while I remember the general narrative, I couldn’t recall the exact specifics. If someone had asked for a link or the precise details, I wouldn’t have been able to provide it. I was working on a PowerPoint and wanted the exact information to be able to cite properly.
I first tried GPT-4o with ChatGPT, and it came up basically empty. So then I turned to Deep Research, which took about 8-9 minutes to complete the search, and it came back with exactly what I was looking for—links, pictures, everything! I thought, “Man, that is really cool.”
It’s not a way I had considered using Deep Research before—for something seemingly as trivial as “I remember a few details, now go find the exact specifics and references to it.” This tool continues to surprise me with its versatility and usefulness in unexpected scenarios.
