{"course":{"slug":"prompting","title":"Getting Real Work out of a Model","url":"https://addaly.com/learn/prompting","level":"beginner","track":"start"},"lesson":{"slug":"context-is-most-of-the-job","title":"Context is most of the job","url":"https://addaly.com/learn/prompting/context-is-most-of-the-job","position":10,"of":72,"minutes":7,"answer":"The model only knows what is in front of it, so paste the thing itself and paste the part that matters.","keyIdea":"The model only knows what is in front of it, so paste the thing itself and paste the part that matters."},"body":{"format":"markdown","text":"## What context means here\n\nContext is everything sitting in front of the model when it answers: your question, the text you pasted, the earlier turns of this conversation, and anything the product pulled in for you. Nothing else exists. Not the files on your laptop, not last week's chat, not what you meant.\n\nSo the working question is simple. What does it need that it could not possibly work out on its own?\n\n## 1. The thing itself, not a description of the thing\n\nThis is the most common mistake by a wide margin.\n\n> My Python script keeps crashing when it loads the sales file. What is wrong?\n\nYou have described a crash. The model now guesses at the ten most common causes of crashes, and one of them might be yours.\n\nNow paste four things: the twelve-line function, the full traceback, one row of the CSV, and one sentence of background.\n\n```\nTraceback (most recent call last):\n  File \"daily.py\", line 31, in totals\n    return row[\"total_amount\"]\nKeyError: 'total_amount'\n```\n\n> The file is exported from our point-of-sale system every night. The export changed last week.\n\nThe CSV row shows a column named `totalAmount`. Answered in one line. Nothing else was needed, and no amount of clever phrasing would have substituted for it.\n\nThe same rule holds outside code. Replying to an email: paste the email. Editing a draft: paste the draft. Asking whether a contract is fair: paste the contract.\n\n## 2. The constraints that make it hard\n\n> Suggest a birthday plan for my mother.\n\nversus\n\n> Suggest a birthday plan for my mother. She turns 70 in Bogotá in March. Total budget 400,000 pesos. She uses a walking frame, so nothing that means standing for long. Eleven people, four of them under six. She hates surprises and hates being sung to in restaurants.\n\nThe second one is a real problem, and real problems have interesting answers. Budget, deadline, word limit, the person who will read it and how they will react, what you already tried and rejected, the thing you are not allowed to say. These are not extra. They are the task.\n\n## 3. Things you think are too obvious to mention\n\nToday's date. Your country. Your currency. That your school year starts in January. That the shops near you close at six. That \"college\" means something different where you live. That the reader is your father-in-law.\n\nA model has a rough sense of when it was trained and no sense of where you are. If a detail would change the answer, it goes in the prompt.\n\n## More is not automatically better\n\nHere is where a lot of advice goes wrong. Models now accept enormous amounts of text, and people conclude that pasting everything is the safe move.\n\nIt is not. As the pile grows, a model gets less reliable at finding and using one specific fact inside it. Researchers call this being lost in the middle, and the size of the effect is argued over, but the direction is not. A relevant chapter beats an irrelevant book.\n\nSo: paste chapter four, not the whole textbook. If you need the whole document in there, say where to look. \"The payment terms are in section 9; the termination clause is somewhere near the end.\"\n\nOld context sticks too. If turn three of your conversation contained a wrong assumption, turns four to thirty were all built on it, and it is still in the room. That is a real reason to start again rather than keep correcting, and we come back to it later.\n\n## The habit\n\nBefore you send anything, ask: if a competent stranger had only this message and no way to reach me, could they start?\n\nIf the answer is no, you have found your next line."},"check":{"question":"A student pastes a 90-page textbook PDF and asks for the three formulas relevant to one homework question. One of the three comes back wrong. A classmate pastes only chapter 4 and gets all three right. What best explains the difference?","options":[{"text":"The relevant material was a small part of a large pile, and models become less reliable at locating one specific fact as the surrounding text grows.","correct":true},{"text":"Models only read the first few pages of a long document and ignore the rest, so the later formulas were never seen.","correct":false,"tempting":"People notice long-document failures and reasonably infer a hard cutoff near the top, rather than a reliability that thins out across the whole thing."},{"text":"PDFs cannot be read properly, so the text arrived corrupted for the first student.","correct":false,"tempting":"PDF text extraction really is sometimes mangled, which makes it a satisfying universal explanation for anything that goes wrong with a document."},{"text":"The second student was lucky. Run the first prompt again and it would probably get all three.","correct":false,"tempting":"Outputs really do vary between runs, so randomness becomes the catch-all; but a consistent difference in what was pasted is not luck."}]},"prev":{"slug":"the-product-around-the-model","title":"Why the same prompt behaves differently in different apps","url":"https://addaly.com/learn/prompting/the-product-around-the-model"},"next":{"slug":"paste-the-source-not-a-summary","title":"Paste the source, not your account of it","url":"https://addaly.com/learn/prompting/paste-the-source-not-a-summary"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Context is most of the job', in Getting Real Work out of a Model. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/prompting/context-is-most-of-the-job (Accessed: 31 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Context is most of the job","courseTitle":"Getting Real Work out of a Model","module":{"index":2,"slug":"the-material-you-give-it","title":"The material you give it"},"answer":null,"keyIdea":"The model only knows what is in front of it, so paste the thing itself and paste the part that matters.","hero":{"kind":"plate","image":"https://addaly.com/art/modules/prompting/the-material-you-give-it.webp","alt":"A tailor’s counter with a bolt of cloth, a paper pattern, chalk, shears and a phone lying screen down.","caption":"A tailor works from the cloth and the pattern rather than a description of them, and a model answers better for exactly the same reason."},"blocks":[{"t":"h2","id":"what-context-means-here","runs":[{"text":"What context means here"}]},{"t":"p","runs":[{"text":"Context is everything sitting in front of the model when it answers: your question, the text you pasted, the earlier turns of this conversation, and anything the product pulled in for you. Nothing else exists. Not the files on your laptop, not last week’s chat, not what you meant."}]},{"t":"p","runs":[{"text":"So the working question is simple. What does it need that it could not possibly work out on its own?"}]},{"t":"h2","id":"1-the-thing-itself-not-a-description-of-the-thing","runs":[{"text":"1. The thing itself, not a description of the thing"}]},{"t":"p","runs":[{"text":"This is the most common mistake by a wide margin."}]},{"t":"quote","runs":[{"text":"My Python script keeps crashing when it loads the sales file. What is wrong?"}]},{"t":"p","runs":[{"text":"You have described a crash. The model now guesses at the ten most common causes of crashes, and one of them might be yours."}]},{"t":"p","runs":[{"text":"Now paste four things: the twelve-line function, the full traceback, one row of the CSV, and one sentence of background."}]},{"t":"code","lang":"","label":"Text","text":"Traceback (most recent call last):\n  File \"daily.py\", line 31, in totals\n    return row[\"total_amount\"]\nKeyError: 'total_amount'"},{"t":"quote","runs":[{"text":"The file is exported from our point-of-sale system every night. The export changed last week."}]},{"t":"p","runs":[{"text":"The CSV row shows a column named "},{"text":"totalAmount","code":true},{"text":". Answered in one line. Nothing else was needed, and no amount of clever phrasing would have substituted for it."}]},{"t":"p","runs":[{"text":"The same rule holds outside code. Replying to an email: paste the email. Editing a draft: paste the draft. Asking whether a contract is fair: paste the contract."}]},{"t":"h2","id":"2-the-constraints-that-make-it-hard","runs":[{"text":"2. The constraints that make it hard"}]},{"t":"quote","runs":[{"text":"Suggest a birthday plan for my mother."}]},{"t":"p","runs":[{"text":"versus"}]},{"t":"quote","runs":[{"text":"Suggest a birthday plan for my mother. She turns 70 in Bogotá in March. Total budget 400,000 pesos. She uses a walking frame, so nothing that means standing for long. Eleven people, four of them under six. She hates surprises and hates being sung to in restaurants."}]},{"t":"p","runs":[{"text":"The second one is a real problem, and real problems have interesting answers. Budget, deadline, word limit, the person who will read it and how they will react, what you already tried and rejected, the thing you are not allowed to say. These are not extra. They are the task."}]},{"t":"h2","id":"3-things-you-think-are-too-obvious-to-mention","runs":[{"text":"3. Things you think are too obvious to mention"}]},{"t":"p","runs":[{"text":"Today’s date. Your country. Your currency. That your school year starts in January. That the shops near you close at six. That “college” means something different where you live. That the reader is your father-in-law."}]},{"t":"p","runs":[{"text":"A model has a rough sense of when it was trained and no sense of where you are. If a detail would change the answer, it goes in the prompt."}]},{"t":"h2","id":"more-is-not-automatically-better","runs":[{"text":"More is not automatically better"}]},{"t":"p","runs":[{"text":"Here is where a lot of advice goes wrong. Models now accept enormous amounts of text, and people conclude that pasting everything is the safe move."}]},{"t":"p","runs":[{"text":"It is not. As the pile grows, a model gets less reliable at finding and using one specific fact inside it. Researchers call this being lost in the middle, and the size of the effect is argued over, but the direction is not. A relevant chapter beats an irrelevant book."}]},{"t":"p","runs":[{"text":"So: paste chapter four, not the whole textbook. If you need the whole document in there, say where to look. “The payment terms are in section 9; the termination clause is somewhere near the end.”"}]},{"t":"p","runs":[{"text":"Old context sticks too. If turn three of your conversation contained a wrong assumption, turns four to thirty were all built on it, and it is still in the room. That is a real reason to start again rather than keep correcting, and we come back to it later."}]},{"t":"h2","id":"the-habit","runs":[{"text":"The habit"}]},{"t":"p","runs":[{"text":"Before you send anything, ask: if a competent stranger had only this message and no way to reach me, could they start?"}]},{"t":"p","runs":[{"text":"If the answer is no, you have found your next line."}]}],"check":{"question":[{"text":"A student pastes a 90-page textbook PDF and asks for the three formulas relevant to one homework question. One of the three comes back wrong. A classmate pastes only chapter 4 and gets all three right. What best explains the difference?"}],"options":[{"text":[{"text":"The relevant material was a small part of a large pile, and models become less reliable at locating one specific fact as the surrounding text grows."}],"correct":true,"tempting":null},{"text":[{"text":"Models only read the first few pages of a long document and ignore the rest, so the later formulas were never seen."}],"correct":false,"tempting":[{"text":"People notice long-document failures and reasonably infer a hard cutoff near the top, rather than a reliability that thins out across the whole thing."}]},{"text":[{"text":"PDFs cannot be read properly, so the text arrived corrupted for the first student."}],"correct":false,"tempting":[{"text":"PDF text extraction really is sometimes mangled, which makes it a satisfying universal explanation for anything that goes wrong with a document."}]},{"text":[{"text":"The second student was lucky. Run the first prompt again and it would probably get all three."}],"correct":false,"tempting":[{"text":"Outputs really do vary between runs, so randomness becomes the catch-all; but a consistent difference in what was pasted is not luck."}]}],"order":[3,1,2,0],"why":null},"prev":{"slug":"the-product-around-the-model","title":"Why the same prompt behaves differently in different apps","minutes":8},"next":{"slug":"paste-the-source-not-a-summary","title":"Paste the source, not your account of it","minutes":8}}}