I couldn't scroll past two posts yesterday without seeing a Nobel Prize post congratulating those who just won in physiology or medicine. Incredible!!!. I can only imagine the feeling. I'm hoping one day I will be able to make an impact in science just like these scientists.
I won’t advise anyone. On my first-ever trip, they gave me the worst experience. It’s been 9 months and I’m yet to see the hand luggage in these pictures. Made several reports yet nothing. Destabilized my life. I will never ever in my life even if my life is on the line.
My brother is considering flying Royal Air Maroc to the UK for Christmas. I've told him not to do it and he's wondering why
Please what are your experiences on this airline so he can understand why
Brasil tuvo esclavos negros
Portugal tuvo esclavos negros
España tuvo esclavos negros
Gran Bretaña traficaba negros
Francia tuvo un zoológico de negros
Los musulmanes matan judíos, cristianos, mujeres, gays, etc.
Quién es el racista?: Argentina
I asked Claude to teach me Excel like a $200/hour tutor.
31 days later I automated a report that ate 6 hours of my week.
Here are the 5 prompts I used (swap in any skill):
🚨BREAKING: Claude can now write your entire job application like a top recruiter.
Here are 10 prompts that turn a job description into a tailored CV, cover letter, and interview prep guide in under 5 minutes.
(Save this)
@TzuHanH@TheBorzoo Congratulations
I hope to achieve such a milestone. I have always dreamed of earning a PhD and contributing to the development of novel treatments for diseases that afflict humanity. Congratulations once again
Prompting is a research skill.
Like coding, writing, or presenting, mastering how to prompt AI will soon be a core academic asset.
The better you prompt, the better your research output becomes.
Quality in = quality out.
Unfortunately, most grad students and researchers don't know how to generate outputs effectively…
Here is a summary of how to engineer prompts that generate precise and reliable results.
—
➊ 𝗔𝗹𝘄𝗮𝘆𝘀 𝗦𝗲𝘁 𝗧𝗵𝗲 𝗖𝗼𝗻𝘁𝗲𝘅𝘁
LLMs respond better with clarity.
➤ Poor Prompt: “What’s a good research topic?”
➤ Better:
“I’m a PhD student in global health focusing on infectious diseases in Sub-Saharan Africa. Suggest 5 original research questions related to antibiotic resistance trends post-COVID.”
↳ Why this works: Context narrows the model’s scope. It helps AI act like a domain-specific assistant.
—
➋ 𝗨𝘀𝗲 𝗥𝗼𝗹𝗲-𝗕𝗮𝘀𝗲𝗱 𝗣𝗿���𝗺𝗽𝘁𝘀
Frame the AI as someone you want it to emulate.
➤ “You are a research mentor with expertise in social epidemiology. Help me refine this study objective…”
➤ “Act as a peer reviewer for an academic journal. Critique the clarity of this abstract…”
↳ Why this works: It tunes the tone, structure, and depth of the response to match expectations.
—
➌ 𝗔𝘀𝗸 𝗪𝗲𝗹𝗹-𝗦𝗰𝗼𝗽𝗲𝗱 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀
Break down complex queries into smaller parts.
✔ Instead of: “Help me write my entire literature review.”
✘ Do this:
➤ “Summarize 3 major themes in the literature on [topic].”
➤ “What are some unresolved debates in this field?”
➤ “Generate 3 counter-arguments to this commonly held assumption…”
—
➍ 𝗟𝗲𝘃𝗲𝗿𝗮𝗴𝗲 𝗣𝗿𝗼𝗺𝗽𝘁 𝗖𝗵𝗮𝗶𝗻𝗶𝗻𝗴 (Step-by-step logic)
Prompting is not one-and-done. Layer your prompts.
➤ Step 1: ���Generate a list of potential research gaps in [topic].”
➤ Step 2: “For #2 in that list, suggest 3 study designs to explore it.”
➤ Step 3: “Now draft a research question based on the best design.”
↳ Why this works: AI performs better iteratively. Don’t expect one prompt to do everything.
—
➎ 𝗘𝗱𝘂𝗰𝗮𝘁𝗲 𝗧𝗵𝗲 𝗠𝗼𝗱𝗲𝗹 (if needed)
LLMs are generalists. Feed them your working definitions.
➤ “In this context, X refers to [definition]. Use that to explain the limitations of…”
➤ “For this study, we define ‘access to care’ as…”
↳ Why this works: Eliminates ambiguity. Prevents the model from defaulting to vague or mismatched assumptions.
—
➏ 𝗥𝗲𝗾𝘂𝗲𝘀𝘁 𝗢𝘂𝘁𝗽𝘂𝘁 𝗙𝗼𝗿𝗺𝗮𝘁𝘀
Be specific about what format you want:
→ “Summarize in bullet points.”
→ “List as a table with columns for pros/cons.”
→ “Give examples in APA format.”
→ “Draft as a structured abstract with background, methods, results, and conclusion.”
—
➐ 𝗧𝗲𝘀𝘁 𝗳𝗼𝗿 𝗕𝗶𝗮𝘀, 𝗙𝗮𝗰𝘁𝘂𝗮𝗹 𝗘𝗿𝗿𝗼𝗿𝘀 & 𝗛𝗮𝗹𝗹𝘂𝗰𝗶𝗻𝗮𝘁𝗶𝗼𝗻
Always verify AI-generated facts.
➤ Cross-check references.
➤ Ask follow-up: “What is the source of that claim?”
➤ Use Google Scholar or PubMed to confirm validity.
—
➑ 𝗣𝗿𝗼𝗺𝗽𝘁 𝗘𝘅𝗮𝗺𝗽𝗹𝗲𝘀 𝗙𝗼𝗿 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀
• “Act as an NIH grant reviewer. Give feedback on this specific aim.”
• “Create 3 counter-hypotheses for this observed trend.”
• “Rephrase this abstract for clarity while maintaining scientific tone.”
• “Generate a one-paragraph summary of key themes in the literature on [topic] post-2019.”
• “List 5 journals that commonly publish work in [specific subfield].”
—
➒ Pro-Tip: Save Your Prompt Templates
Create prompt banks for recurring tasks:
→ Literature Review
→ Abstract Editing
→ Cold Emails to Researchers
→ Thesis Structuring
→ Study Design Comparisons
→ Scholarship Applications
This will dramatically save time and ensure consistency.
I watched a PhD student use NotebookLM to do in 3 hours what took me 3 months.
I asked him how. He showed me his setup. I have not studied the same way since.
Here is exactly what he did.
He did not use NotebookLM as a search engine.
He did not type questions like "explain this concept."
He uploaded an entire field.
12 textbooks. 40 research papers. Every review article published in the last 5 years on his
topic. All uploaded at once. Into a single notebook.
Then he asked one question that changed how I think about learning:
"If you could only teach this subject using 7 sentences — one for each core idea — what
would they be?"
Not a summary. Not an overview.
Seven sentences. For the entire field.
NotebookLM produced them. He read them.
He asked a follow-up:
"For each sentence, what is the single experiment or paper that proved it was true?"
Now he had the skeleton of an entire field — and the exact evidence underneath each bone.
Then he did something nobody told me about.
He uploaded his own notes alongside the source material.
Then asked:
"Where am I wrong? What have I misunderstood compared to what the literature actually says?"
The system found three misconceptions he had carried for two years. Two of them would
have cost him his PhD defense if they had gone uncorrected.
He called this the Misconception Audit. He runs it on every subject before any exam, any
paper submission, any conference presentation.
Then the final step.
"Generate 10 questions a hostile expert would ask someone who claims to understand this
field. Then answer each one using only the sources I uploaded."
He was not studying to pass.
He was studying to survive interrogation by the smartest people in the room.
He passed his qualifying exam on a subject he had started studying 11 days before.
I have been studying the wrong way my entire life.
Here is his exact setup if you want to replicate it:
1.Upload the entire field — not one source
2. Ask for the 7 core sentences
3.Find the evidence under each one
https://t.co/3gcg6x5dpi the Misconception Audit on your own notes
Survive the hostile expert interrogation
The information was always available.
The framing was the unlock.
He passed his qualifying exam on a subject he had started studying 11 days before.
I have been studying the wrong way my entire life.
Here is his exact setup if you want to replicate it:
1.Upload the entire field — not one source
2. Ask for the 7 core sentences
3.Find the evidence under each one
https://t.co/3gcg6x5dpi the Misconception Audit on your own notes
Survive the hostile expert interrogation
The information was always available.
The framing was the unlock.