How to Easily Get More Information on Your Next Professors' Research
This is an extremely effective and efficient way of keeping abreast with new information and it’s nicely general, so you can use it for many other purposes
This post is an introduction to how I prepare myself for university lecture where the professor and the central questions are different each week. This is both for satisfying my personal curiosity and for getting inspiration for new services that my company can provide.
As a trainer in collaboration between humans and generative AI, I use a lot of AI tools. And the tools keep getting better. This use case should be seen as inspiration for figuring out how you get better prepared more quickly for meeting new people discussing new topics with the current tools.
Expect this description and your own skills to be slightly outdated already next week.
Background
As part of my consulting business, Frimer-Rasmussen Consulting, I build new services when I’m between clients. I try to implement a complete business idea each month, so I need a backlog of interesting ideas. The university is a great source of ideas that has a hard time leaving the research labs - and I believe I can help get the research out into the hands of the public via simple services.
Also the services serve as a catalogue of “what can be built” which makes it a lot easier for clients to discuss and adapt to the power of generative AI in their own companies.
My first public service, OctaNews - The Democratic Newsletter - is already described in previous posts Building Software with AI parts 1, 2, and 3.
My Problem: Preparing for Something Completely New Each Week
I’m following a course at the University of Aarhus in Denmark. The central theme is how national economy models need to be updated to be able to support important decisions on spending on climate change.
I’m a linguist by training, so national economy is far from my comfort zone, but I firmly believe that more people should know more about one of the central governing tools.
The complete description that I have for today’s lecture is:
“"Sustainable Development – Between Vision and Action"
Lars Josephsen, MSc, Master of Public Policy“
I usually start out with putting the lecture description into a short prompt that broadly describes the context and the purpose.
In this case I start out using Google Gemini 2.0 Flash Thinking Experimental as it generally produces some reasonable starting points.
The output is high-level and that’s what I need. The summary at the end goes:
In summary, your preparation strategy should be:
Investigate Lars Josephsen's background (education, current affiliation, profession).Search for his research and publications related to sustainable development.Analyze the lecture title "Sustainable Development – Between Vision and Action" and consider potential themes.Formulate questions and expectations for the lecture based on your research and interests.
Even if you can only find limited information about Lars Josephsen, the act of searching and thinking about these points will make you much better prepared to engage with his lecture. Good luck! Let me know if you find some information and want to discuss it further.
I copy the entire output from Gemini into ChatGPT o3-mini-high with Search functionality enabled. That combines the overall ideas and needs with the actual details filled in and anchored in reality by searching the web for the information.
The output is rather long. This is just the top quarter of the output. This serves as a quick check ( less than 10 seconds of “thinking”) to make sure we’re moving in the right direction.
I take the output Background and Research paragraphs and input them into Google Gemini 1.5 Pro with Deep Research.
The Deep Research “agent” then defines this plan:
Find Lars Josephsen's academic publications on sustainable development.
Find Lars Josephsen's outreach activities related to sustainable development, such as articles, interviews, and lectures.
Find information about Lars Josephsen's work on the transition from vision to action in sustainable development.
Find reviews and discussions of the book "Global Goals: The Vision of Sustainable Development" (2019).
Find information about Steen Hildebrandt and his work on sustainable development.
Find articles or chapters from the book "Global Goals: The Vision of Sustainable Development" (2019) that specifically address the challenges and opportunities of implementing sustainable solutions.
Find information on how Lars Josephsen's work has contributed to the understanding of sustainable development.
And I let it loose on the internet. The Gemini agent notifies me when it’s done and I can open the output report in Google Drive.
The report ends with a listing of used (8) references out of 83 found references.
The references are copied into ChatGPT-4o. Note: no “thinking” required as we’re just interested in a simple description of each reference.
Each reference now gets copied into a Google Doc. We have two simple reports: one about Lars and his research, and one about the main references to Lars and his research. But we need to think about the most interesting questions based on all this.
So I switch to ChatGPT o3-mini-high in the same chat to keep context and ask for the three most important questions:
The prompt is simple as the context is nicely established above: “What are the three most interesting questions with an impact on Danish policy on climate change? Keep it short (less than 500 characters)“. This length limit is due to NotebookLM.
The output is very good, I think:
How can Denmark ensure a fair transition to a climate-neutral economy?
How can investments in renewable energy and green technology drive both climate goals and economic growth?
How can Denmark strengthen cross-sector coordination of climate and energy policies?
Certainly not trivial questions to answer!
Now we’re ready for NotebookLM Plus. I add the two documents from Google Drive (left column below). The abstract is automatically created in the middle chat window. In the right Studio column, I add the three questions (there’s a limit of 500 characters here) and create the “podcast”. That takes a while, but you don’t have to wait for it.
Now I have a sound file, I can play on my phone NotebookLM app and listen to while I bike to the university and I feel well prepared.
During the lecture I keep adding the most interesting references to the notebook. I can also ask questions to the new references directly in the NotebookLM chat, so I don’t disturb the progress of the lecture. Some of the professors are very keen on not getting interrupted.
The End Game
When we’re done, I update the research questions for the podcast based on what I’ve heard and the discussions during the lecture. I refresh the podcast and have an updated notebook with better references, better questions and an improved podcast to listen to on my way home.
Overall this proces give me a way better understanding than just going there with the ultra basic course description. And the podcasts deepen my understanding both times.
This is an extremely effective and efficient way of keeping abreast with new information and it’s nicely general, so you can use it for many other purposes.