Short YouTube talk by Mollie Amkraut Mueller giving three practical tips for how to provide useful context to a large language model (LLM). The tips: (1) explicitly tell the model what it is and what its goal is (give specific roles and objectives), (2) feed the model as much relevant information as you can (documents, text, videos) while minimizing your own cleanup work, and (3) ask the LLM what additional information it needs and either upload documents or provide a quick voice-note transcript to fill gaps. The piece emphasizes practicality and saving time.
Questions this video answers
Mollie Amkraut Mueller.
July 23, 2025.
Chapters
Hi. Today I want to talk about context. Context is the information that you feed into your LLM so that it can do a really good job on the task that you're giving it. And I'm going to give you three tips for how to give it the right context. Number one, tell it what it is and what its goal is. For example, you are my executive coach and your goal is to help me get promoted in the next six months. Or you are my startup founder who specializes in seed to Series A and your goal is to help me get to product market fit or get my fundraise done, etc. Be specific. Number two, feed it as much information as you can in terms of text and documents, even videos, Whatever is easy for you to upload or copy paste. Try and make this not a big lift for you, but give it whatever is readily available. And number three, once you've done those two things, ask your LLM so ChatGPT, Claude, etc. Say what additional information do you need from me to do a good job on this task? And and it'll come back and it'll say what the gaps are and you can evaluate if those gaps are worth your time to fill or if you think it's going to do a good job without them. If you do want to fill in those gaps for it, then either find a document you can easily copy, paste in or upload. Again, this is all about saving you time. Or if it really doesn't exist anywhere, then my hack is to have it ask me a few questions and and then I'll either speak to it in voice mode or I'll open up a voice note on my phone and just do a brain dump speaking in all sorts of messy language. And then I'll send that transcript or even upload that audio file into the LLM and it can synthesize that way faster than it would take for me to clean up my own Speaking with those three tips, you should have a really great head start on giving your LLM the context that it needs to help you with Pro. Pretty much any task. Let me know how it goes.
