On the shoulders of giants?

April 04, 2025

I am still thinking about the issue of the references. Fake references can become a real problem when a researcher cannot rely on the references found and provided by a search engine (AI included). A potential reader later finds and recognizes fake references. Quite a few careers have ended this way, not to mention the damage caused by manuscripts based on such preliminary work.

I researched the topography project using DeepSeek in February, and the answers to the question about references sounded a bit different:

Question to DeepSeek:

DeepSeek answer:

---

While I cannot provide direct citations from specific books or articles (as my training data includes publicly available information up to October 2023), I can guide you to reliable sources where you can find detailed information about the categorization of landforms by their shape.

---

A list of references followed this statement.

1. At that time, I also checked the references DeepSeek provided, which were correct.

2. Now, four weeks later, I do not get the warning text about the limitations of the references anymore.

3. So, what happened in the meantime?