How to research a person across podcast appearances

Build a focused question, compare interviews in context, and keep the evidence attached. A practical method with real source-linked podcast research examples.

Start with a question that has boundaries

To research a person across podcasts, define one question, find relevant appearances, and compare what each conversation actually supports. Record the episode, date, speaker, and timestamp for every important point. Then write an answer that separates repeated ideas, differences, and gaps in the evidence.

A request such as 'Tell me everything about Morgan Housel' leaves the useful outcome unclear. 'How does Morgan Housel think you should spend money to become happier?' gives the research a subject and a decision. It also makes it easier to recognise an episode that discusses investing but contributes little about spending.

Write down what you want to understand before collecting sources. If you care about a position changing over time, include that explicitly. If you need advice for a particular situation, state the situation without assuming the speaker addressed it directly.

Choose appearances for evidence, not name recognition

Look for original full episodes from the show or publisher. Check the guest name, episode title, publication date, and available description. A familiar name in a title is a lead, not proof that the discussion answers your question.

Watch for the same recording republished as clips, compilations, or another listing. Three URLs can be one conversation. Counting them as separate interviews makes repeated material look like independent support.

Keep a short source list and record why each episode belongs. For a question about mathematical AI, an interview where Grant Sanderson discusses what mathematical understanding means is more useful than an appearance that only mentions his work in passing.

Make a small evidence note for each claim

Use the same fields for each observation: the claim in your own words, the speaker, the episode date, the timestamp, and the surrounding qualification. This keeps a polished summary from hiding a weak source.

Suppose the subject says a tool helps with routine work but not with choosing important problems. Record both halves together. A note that says only 'supports the tool' loses the distinction the research should preserve.

Also distinguish a personal experience, a general recommendation, a prediction, and a claim about published evidence. They do different jobs. Repeating a prediction across interviews does not turn it into an observed result.

  • What exactly is the claim?
  • Who said it, and in which episode?
  • Where can someone hear it in context?
  • Which qualification would change its meaning if omitted?

Compare differences without forcing a contradiction

Lay the appearances out by date and group the evidence by idea. Repeated principles belong together. Changes in emphasis should remain visible. A difference is worth investigating, but it may reflect a different audience, time horizon, or question rather than a reversal.

For a position to count as changed, you need evidence of both the earlier and later position on the same point. If one interview is silent, say it does not address the point. Silence is not disagreement.

The public Morgan Housel answer below combines several appearances around spending and happiness. Read its source links alongside the synthesis to see how ideas such as autonomy, status, and relationships fit together without treating every remark as universal advice.

Write the answer, then challenge its scope

Open with the answer to your question. Follow it with the main supporting ideas, the differences, and the limitations. Attach citations where a reader needs them rather than putting a long, disconnected source list at the end.

Before sharing, ask whether every major conclusion is supported by the episodes you reviewed. Replace 'always says' with a narrower description when your evidence is a handful of interviews. Make the date range and source count visible.

A podcast can help you establish what someone said. It cannot by itself prove a medical, financial, or other consequential claim. When that distinction matters to your decision, check the underlying evidence separately.

Use TLDR Pods for the question you want to investigate

TLDR Pods turns a focused question into an original, source-linked Tldr drawn from relevant public podcast episodes. Include the person, topic, and any show or date clues you remember. Its coverage is bounded by the sources available and the plan allowance, so inspect what was reviewed before drawing broader conclusions.

Your own report and its follow-up chats remain private. The examples below are curated public answers you can inspect without an account. Use them to assess the source links, the level of detail, and the limits before deciding whether the paid workflow fits your research.

Further reading and source examples

This guide describes a practical research method. The examples below link to our published answers and their episode sources. Read the TLDR Pods methodology for our publication standards and the corrections process to report an error.

Read the finished research

These public answers are free to read. No account or card needed.

4 episodes reviewed / 64 timestamp citations

Morgan Housel on How to Spend Money for a Happier Life

Morgan Housel says money is most likely to improve happiness when it buys control over your time, protects your family, and enables attention-rich relationships. Save enough to gain independence, but use that freedom for worthwhile work, rest, health, and people you love. Spend on personally valuable comforts and shared experiences, not purchases mainly intended to impress strangers.

2 episodes reviewed / 65 timestamp citations

Grant Sanderson on AI and the Future of Mathematics

Grant Sanderson expects AI to become exceptional at proving and explaining mathematics, but says the deeper transformation begins when systems can choose valuable questions, invent productive definitions, build theories, and compress many results into intelligible ideas. Human mathematicians would increasingly judge significance, curate machine-generated work, and connect it to problems people value, unless machines eventually surpass them at those tasks too.

Bring a question of your own.

Creating a private Tldr requires a verified account and a paid plan or eligible introductory trial. Check the current allowance and billing terms before you choose.