Ultra deep dive
Winkletter • 25 Jan 2026 •
Yesterday I was assembling a flat pack filing cabinet while running research queries in between steps. There’s something deeply satisfying putting together a well-designed, solid kit. The pieces weren’t flimsy. The holes were in the right place. Everything I needed was in the box. The people who made the kit knew what they were doing even though the company has an alphabet-soup name.
A month ago I wrote about the Transmission of Buddhist teachings through Thailand and into the UK, AUS, and US. To research that I ran about five deep research queries and used the reports to learn the history of the 2009 schism that led to the expulsion of the Buddhist Society of Western Australia.
This time I assembled my furniture and ran 90 deep research queries on the side which generated about 1000 pages of cited research about the transmission of Buddhist dhamma across the world. And that’s just 45% of my daily research quota!
I collected more information than I can parse in a week. But I was learning the history of Buddhism throughout the process and I now have 95 different reports that can be used to generate videos for my daily YouTube posts. Here’s an overview of the process which I’m calling the Ultra Deep Dive.
The Ultra Deep Dive Process
1. Use the Gemini Thinking model with Deep Research to explore a large well-defined topic. Clearly define borders: temporal, spatial, and topical.
Starting from the Buddha’s decision to teach the dhamma until the moment of his parinibbana, research how he built the sangha into an organization that could last over 2500 years.
2. Use NotebookLM’s Reports to generate more specific research prompts from the research generated in step 1.
Brainstorm 12–15 high-quality research prompt ideas a researcher could run to expand the source material. Each prompt should be optimized for deep research and able to stand alone without referring to other prompts or the source material. Cover a broad range of goals including: understanding core concepts, expanding into applicable domains, and identifying active or under-developed research areas. Write each item with a clear title and a 100–150 word description in one paragraph. Output only the prompt ideas and descriptions, not analysis of the text itself. Prioritize originality, research value, and real-world investigatory depth.
3. Feed the resulting prompts into Deep Research with the Fast model, reading and editing the prompts as you go, and skipping anything that doesn’t interest you. You might need to add a note about the overall research project.
I’m researching the transmission of Buddhism globally. Deconstruct the linguistic technique of geyi (matching concepts) used by early Chinese translators. Evaluate the ontological risks of equating Sunyata (emptiness) with the Daoist Wu (non-being) and Nirvana with Wu-wei (non-action). Analyze whether this was a necessary upaya for early acceptance or a fundamental categorical error, and explain how the later, more accurate translations of figures like Kumarajiva and Xuanzang attempted to rectify these initial conceptual mappings to define a distinct Chinese Buddhist identity.
4. Export the research reports to Google Docs (numbering as you go) and collect the docs into a NotebookLM notebook. Repeat this process to build out knowledge in other related areas.
At the end of the process I had an AI notebook with almost 100 custom reports with citations that I can read, interrogate, summarize, and reshape to make texts, images, videos and other study materials that I can share. Thankfully, I don’t have to store them in my new filing cabinet.
- Researching Buddhism and the Thai Forest Tradition << Link to the notebook if you want to see or even use it.
Comments
That’s a good question to ask. I’m going to set aside my real answer which is research and knowledge are my main sources of joy. I’m broken that way.
I have seen improvements recently from using AI to study Buddhism and specific psychological paradigms (Method of Levels). This has helped me respond better to anxiety. I’ve also been studying social systems theory and how it relates to media. I’m more aware of media manipulation and spend less time watching channels that rage-bait.
Here’s a grab bag of recent wins.
- Learning about gardening to put together a garden plan.
- Learning about farming conditions near my brother’s farm so we have things to talk about.
- Learning to cook better with an air fryer.
- Learning about other Substack newsletters to use as models. Not just URLs, but in-depth reports and profiles.
If you want I could run some research into biohackers on Substack, or something similar to see if you find it useful. I’d be interested to know if someone more on top of things can make better use of this kind of research.
Thanks for your detailed response. Your first answer shouldn’t be discarded. Joy is fleeting, and we should do all we can to experience it often. If it’s not too much effort, I’d be happy to take a look at what you find with biohackers on Substack.
Here’s a link to the notebook I created. Biohacking Newsletters. I ran two research reports to gather names of popular newsletters (00a and 00b). I then ran research on 20 newsletters, and finished off with a rhetorical analysis. I also generated a bunch of audio summaries, some slides, and a video. I might generate more media.
You should be able to view the sources (left) and the various media I created (right), although I don’t think you can create anything. You should be able to chat (middle) with the research reports in the Sources column (left). Let me know if you want more research in a different direction, or you want anything generated in the Studio. I have ridiculous levels of excess capacity.
Wow. I opened that up, and honestly, it’s overwhelming. No idea where I would start.
I have a question after seeing many of your deep dives on various topics. I’m genuinely curious and not asking in a judgmental way because you are the only person I know who has done such a deep dive into AI. How are all these AI tools, the research, knowledge, and findings making your life better?