Deepseek for Research Writing | Better than Gemini Ai & Chatgpt
Based on Dr Rizwana Mustafa's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.
DeepSeek is presented as a research-first AI tool that retrieves topic-specific papers with references and clickable links for verification.
Briefing
DeepSeek is positioned as a research-focused AI assistant that helps academics move faster from literature discovery to writing—especially when citations and source-grounded summaries matter. The workflow starts with using DeepSeek via a web interface (or a free mobile app) and prompting it to retrieve a targeted set of papers with references. In the example, a request for literature on “application Bas St liquids in Madison” returns a list of research papers with clickable links, enabling direct access to the sources for verification and deeper reading.
A key advantage highlighted is how DeepSeek handles citation-grounded writing. When asked to draft a literature review chapter (about 1,000 words) using the previously gathered papers, DeepSeek produces structured sections—such as introductions, drug synthesis and solubilization, antimicrobial applications, protein stabilization, biosensing/diagnostics, challenges, and future prospects—while tying the content to the provided references. The transcript contrasts this with Gemini AI, where the writing is described as less complete on the “links of all the papers,” prompting the user to manually search by title and author to authenticate details. The comparison frames DeepSeek as more reliable for researchers who need both a bibliography and a draft that stays anchored to specific sources.
DeepSeek is also presented as stronger for extracting information from documents and files, not just text prompts. The transcript describes attaching up to 50 images at once and asking DeepSeek to extract and summarize the technical content into a coherent “story,” including categories like crystal design and materials, polymer materials and fiber fabrication, and other scientific systems and processes. It further claims compatibility with PDFs and Microsoft Word files, where users can ask for key findings, methodologies, and step-by-step procedures from a specific paper.
In the methodology example, DeepSeek is said to break down a research workflow into stages—such as initial synthesis steps, mechanistic studies, synthetic modifications, green chemistry approaches, DNA-encoded libraries for high-throughput screening and drug discovery, followed by biological evaluations and computational/analytical techniques to finalize results. The practical takeaway is that researchers can “chat with” documents—images, PDFs, and Word files—to generate summaries, results-focused overviews, and method breakdowns that can be reused in academic writing.
Overall, the transcript argues that DeepSeek’s combination of reference-backed literature retrieval, more targeted chapter drafting, and document/image extraction makes it a more research-compatible tool than alternatives like Gemini AI and ChatGPT for academic writing tasks—particularly for users who want citations, verification, and structured synthesis without starting from scratch.
Cornell Notes
DeepSeek is presented as a research-writing assistant that helps academics find relevant papers with references, then draft literature review sections grounded in those sources. In a demonstration, it returns a set of clickable papers for a targeted topic and produces a structured ~1,000-word literature review chapter using the provided references. The transcript contrasts this with Gemini AI, which is described as less consistent in providing links for all papers, requiring manual verification. DeepSeek is also portrayed as capable of extracting information from up to 50 attached images and from PDFs/Word documents, generating summaries, key findings, and step-by-step methodology breakdowns that can support academic writing.
How does DeepSeek support literature review writing beyond generating text?
What specific difference is highlighted between DeepSeek and Gemini AI for research workflows?
What document formats and attachment limits does the transcript claim DeepSeek can handle?
How does DeepSeek handle extracting information from images and turning it into usable writing?
What does the methodology example suggest about DeepSeek’s ability to summarize research papers?
Review Questions
- When building a literature review with DeepSeek, what two-step workflow is used to ensure the draft stays tied to sources?
- What kinds of outputs does the transcript claim DeepSeek can generate from attached images and from PDFs/Word documents?
- How does the transcript describe the need for manual verification when using Gemini AI compared with DeepSeek?
Key Points
- 1
DeepSeek is presented as a research-first AI tool that retrieves topic-specific papers with references and clickable links for verification.
- 2
A literature review can be drafted by instructing DeepSeek to write using only the provided references, producing structured sections (e.g., drug synthesis, antimicrobial applications, biosensing, challenges, future prospects).
- 3
The transcript contrasts DeepSeek with Gemini AI by claiming DeepSeek provides more complete paper links, reducing the need for manual authentication.
- 4
DeepSeek can extract information from up to 50 attached images and summarize the extracted technical content into narrative form.
- 5
DeepSeek is described as compatible with PDFs and Microsoft Word files, enabling users to ask for key findings and results summaries from specific papers.
- 6
For methodology-focused requests, DeepSeek is described as capable of producing step-by-step breakdowns of research procedures, including synthesis, mechanistic work, green chemistry, DNA-encoded libraries, and downstream evaluations.