Get AI summaries of any video or article — Sign up free
PhD Essentials: Don't Start Your PhD Without Watching This! thumbnail

PhD Essentials: Don't Start Your PhD Without Watching This!

Andy Stapleton·
5 min read

Based on Andy Stapleton's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

Choose a reference manager that fits your workflow; Zotero is favored here for stronger support with AI-assisted research, while Mendeley remains viable for traditional citation management.

Briefing

A modern PhD setup hinges on two practical systems: a reference manager that keeps citations organized and an AI toolkit that accelerates literature work, mapping, and writing. For citations, the choice comes down to Zotero versus Mendeley: after experimenting with AI-based research workflows, the recommendation leans toward Zotero because it has stronger support for AI toolchains. Mendeley remains a solid option if the priority is straightforward, reliable reference management rather than AI integration.

Once papers start coming in, the workflow shifts from collecting to navigating. The recommended AI toolkit is built around distinct jobs—searching, mapping, reading papers, multi-document chat, and drafting/editing thesis text. For mapping, tools like “L i p Maps” (with alternatives such as Research Rabbit or Connected Papers) help visualize how a seed paper connects to surrounding literature. For reading without time for full back-to-back coverage, “S i a c e” is used to scan a library and surface what’s relevant. When multiple documents need to be discussed together, “Doc Analyzer” is highlighted as a strong option for uploading PDFs and chatting across them. For writing quality, “right.com” is suggested as an editor to improve clarity and presentation.

Beyond software, the essentials get physical and social. A notebook remains central, but the advice is to use two types: a large “lab/idea” notebook for ongoing scientific work and a pocket-sized system for capturing ideas immediately. Pocket Mod is presented as a simple, wallet-friendly method—print, cut, carry, and jot down flashes of insight on the spot—so memory doesn’t become the bottleneck for creativity. Those notes can later be transferred into a more structured digital workflow such as Notion.

Mentorship is treated as a separate pillar. The guidance is to find a mentor you can talk to regularly, and ideally one outside the exact research field. That distance helps avoid getting stuck in narrow technical details and instead supports broader academic problem-solving. Universities often run mentoring programs for early career researchers, and the recommendation is to push the research office for one if it doesn’t exist.

Academic communication and visibility also matter. The “Academic Phrasebank” from Manchester University is recommended as a phrase library for structuring writing—introducing work, describing methods, reporting results, discussing findings, and writing conclusions—along with a workflow of mixing phrase options into drafts (including using ChatGPT to refine result descriptions).

Finally, the transcript stresses resilience and focus: back up thesis data in multiple places (computer, laptop, hard drive, and cloud) because losing everything to a drive failure is a real risk. For productivity in open-plan offices, it argues for noise control—using headphones and white noise from MyNoise.net—to protect deep work. Social platforms are framed as practical support networks too, with ResearchGate and Academia.edu suggested for discovering papers, connecting with researchers, and getting help when work feels isolating.

Cornell Notes

The core message is that a successful PhD setup combines organization, speed, and support systems. A reference manager (leaning toward Zotero for stronger AI workflow support) keeps citations manageable, while an AI toolkit should be assembled around specific tasks: searching, mapping, reading, multi-document chat, and writing/editing. The workflow is complemented by a two-tier note system—an ongoing notebook plus wallet-ready Pocket Mod slips to capture ideas immediately. Mentorship is recommended, ideally from outside the narrow research area to get help with broader academic challenges. The transcript also emphasizes backups, academic phrase resources, and noise control for sustained focus in open-plan environments.

Why does the transcript favor Zotero over Mendeley for a modern, AI-assisted PhD workflow?

The preference is tied to AI tool support. After experimenting with AI tools during research, the guidance is that Zotero has more robust support for integrating with AI-based workflows than Mendeley. Mendeley is still described as perfectly fine if AI integration isn’t a priority and the goal is reliable, traditional reference management.

How should an AI toolkit be structured so it actually speeds up PhD work?

The toolkit is organized by function rather than by one “all-in-one” app. It includes tools for searching, mapping literature, reading papers, multi-document chat, and writing/editing thesis text. Mapping tools like Li p Maps (with Research Rabbit or Connected Papers as alternatives) visualize connections from a seed paper; reading tools like Si a c e help review what’s been found without reading everything end-to-end; Doc Analyzer supports uploading many PDFs and chatting across them; and right.com is used as an editor to improve academic writing presentation.

What’s the practical purpose of using two notebook types (a big notebook and Pocket Mod)?

The big notebook is for carrying and maintaining the full record of experiments, results, diagrams, and evolving understanding. Pocket Mod is for immediate capture: print and cut pocket-sized sheets, carry them in a wallet, and write down ideas on the spot so memory doesn’t fail. Later, those notes can be transferred into a digital system like Notion for organization.

What makes a mentor “effective” in this advice—field alignment or something else?

The transcript argues for a mentor outside the exact research field. The reason is that cross-field mentors are less likely to dive into narrow technical nitty-gritty and more likely to help with broader academic issues—what’s normal in academia, what’s unusual, and how to respond. A concrete example is given where a biomedical-sciences mentor supported a chemistry PhD by validating whether problems were typical or “completely outrageous.”

What backup strategy is recommended to prevent catastrophic thesis loss?

The guidance is to maintain multiple copies in multiple locations: at least one local backup on a computer and another on a laptop or hard drive, plus cloud storage. The transcript stresses regular backups and warns against single-point failure, including a horror story where a hard drive containing an entire PhD was formatted by an IT staff member.

Review Questions

  1. Which parts of the literature workflow are matched to specific AI tools (mapping, reading, multi-document chat, writing), and what does each tool type do?
  2. What are the two notebook formats recommended, and how does each one support different kinds of PhD work?
  3. Why does the advice prefer a mentor outside the student’s exact research field, and what problem does that help solve?

Key Points

  1. 1

    Choose a reference manager that fits your workflow; Zotero is favored here for stronger support with AI-assisted research, while Mendeley remains viable for traditional citation management.

  2. 2

    Build an AI toolkit around distinct tasks: searching, mapping, reading, multi-document chat, and drafting/editing rather than relying on a single tool.

  3. 3

    Use literature mapping to visualize how a seed paper connects to broader research, with Li p Maps as the preferred option and Research Rabbit or Connected Papers as alternatives.

  4. 4

    Capture ideas immediately with a pocket-sized note system (Pocket Mod) so creativity isn’t lost to memory, then transfer notes into a structured digital workspace like Notion.

  5. 5

    Find a mentor you can talk to regularly, ideally outside the narrow research field to get help with broader academic judgment and problem-solving.

  6. 6

    Back up thesis materials in multiple places (computer, laptop/hard drive, and cloud) and do regular backups to avoid irreversible data loss.

  7. 7

    Protect deep work in open-plan offices with noise control—headphones and white noise via MyNoise.net are recommended.

Highlights

Zotero is recommended over Mendeley specifically because AI-based research workflows get better support.
An AI toolkit should be modular: mapping, reading, multi-document chat, and writing/editing each have their own job and tools.
Pocket Mod turns quick, wallet-carried notes into a reliable idea-capture system that doesn’t depend on internet access.
Mentorship works best when it’s outside the exact research field, helping with broader academic decisions rather than only technical details.
Multiple backups across computer, drives, and cloud are treated as non-negotiable after real stories of thesis loss.

Mentioned