{"course":{"slug":"ai-for-students","title":"AI for Students","url":"https://addaly.com/learn/ai-for-students","level":"beginner","track":"work"},"lesson":{"slug":"getting-set-up-free","title":"Getting set up for nothing","url":"https://addaly.com/learn/ai-for-students/getting-set-up-free","position":8,"of":73,"minutes":8,"answer":"Free tiers, Hugging Face Spaces, Colab, and a small local model on Ollama cover almost everything in this course, so the real skill is spending a limited allowance on the questions that are actually hard.","keyIdea":"Free tiers, Hugging Face Spaces, Colab, and a small local model on Ollama cover almost everything in this course, so the real skill is spending a limited allowance on the questions that are actually hard."},"body":{"format":"markdown","text":"## The constraint most advice ignores\n\nA lot of study advice assumes a laptop with 16GB of RAM, a stable connection and a card on file. Plenty of students have a shared machine, an older phone, patchy data and no card. None of that stops you doing any of this course. It changes which route you take, and the routes are worth knowing before you need them.\n\n## Free access to capable models\n\nEvery major provider runs a free tier with a daily or hourly message cap. The caps move; the shape does not. Practical consequences:\n\n- **Spend your allowance on the hard questions.** Do the easy rewriting with a weaker free model or on your own.\n- **Batch.** One message containing five related questions costs one message from your allowance. Five messages cost five.\n- **Keep two accounts on different providers.** When one runs out, switch. This is not a trick, it is what the free tiers are for.\n\nBeyond the big names: **Hugging Face Spaces** hosts thousands of free demos, including chat interfaces to open-weight models, and needs no payment. **Google Colab** gives free GPU time in a notebook, enough to run a mid-sized open model for an afternoon. Some universities provide staff and students with a paid tier free: check before assuming; a surprising number do and never announce it well.\n\n## Running a model on your own machine\n\nThis is the option that removes quotas, works offline, and keeps everything private. It needs RAM more than anything else.\n\n- **8GB RAM:** a 3B or 4B parameter model, quantised to 4 bits, around 2-3GB on disk. Usable for rephrasing, practice questions from supplied text, vocabulary, simple code.\n- **16GB RAM:** a 7B or 8B model, 4-5GB. Noticeably better; still well short of the hosted ones on hard reasoning.\n- **Apple silicon** punches above its weight here because the memory is shared with the GPU.\n\nThe two free tools are **Ollama**, which is a single install and then `ollama run <model>` in a terminal, and **llama.cpp**, which is more work and runs on almost anything. **LM Studio** gives a graphical interface if terminals are unwelcome.\n\nBe honest about the gap. A 4B model run locally will hallucinate more, follow multi-step instructions worse, and give up on long documents. For a set of tasks (drilling, rewriting, quizzing you from a chapter you paste in) it is entirely adequate, and the fact that it works on a train with no signal is worth a lot.\n\n## A phone-only setup that actually works\n\n1. Chat tools: all major ones have web interfaces that work in a mobile browser; no app needed.\n2. Documents: Google Docs or the free Microsoft web apps; both save continuously, which doubles as your process record.\n3. Notes: Obsidian and Joplin both have mobile apps and sync free via a folder.\n4. Maths: GeoGebra and Desmos are excellent on a phone.\n5. Reading papers: get the PDF into a reader that lets you highlight; Zotero's mobile app is free.\n6. Voice: dictation is free on every phone, and talking through a problem out loud, then pasting the transcript, is a good way to think.\n\nThe thing to avoid on a phone is writing long prose directly into a chat window. It encourages accepting whatever comes back because editing is painful. Draft in a document, paste in, paste back.\n\n## Data, and working offline\n\nIf data is expensive or intermittent:\n\n- Download the material once. Lecture slides, past papers, PDFs. Read offline.\n- Batch your questions. Write them in a note during the day, ask them together when you have signal.\n- A local model needs no connection at all after the download.\n- Wikipedia has offline packages via **Kiwix**: the whole of English Wikipedia without images is a few tens of gigabytes, and a subject-specific selection is far smaller.\n\n## What not to spend money on\n\nThere is a large market in AI study apps charging monthly for a wrapper around a model you can use free. Before paying for anything, ask what it does that a free chat window plus a document does not. Sometimes the answer is real: a good spaced-repetition scheduler, for instance, though **Anki** is free on desktop and Android and is still the best one. Usually the answer is a nicer interface around the same free capability.\n\nThe honest exception is a paid tier of a major model when you are doing heavy work in a specific week. It is bought by the month and cancellable, and one month during a dissertation is a different decision from a standing subscription all year."},"check":{"question":"A student with an 8GB laptop and no budget installs a 4-bit 3B model with Ollama. Which task is it genuinely well suited to?","options":[{"text":"Checking whether a set of references from an essay actually exist.","correct":false,"tempting":"Existence questions need a database, and a small local model has weaker compressed knowledge than a hosted one, making fabrication more likely rather than less."},{"text":"Generating practice questions from a chapter the student pastes in, offline.","correct":true},{"text":"Answering questions about last month's changes to the exam syllabus.","correct":false,"tempting":"Any model without retrieval, local or hosted, has no access to recent information; the local one also has no search to fall back on."},{"text":"Explaining an unfamiliar advanced topic that is not in the student's materials, since local models are trained on the same data.","correct":false,"tempting":"Assumes parameter count does not matter for knowledge; a 3B model holds a far coarser compression of its training text than a large hosted one, and unfamiliar advanced material is exactly where that shows."}]},"prev":{"slug":"which-tool-for-which-job","title":"Which tool for which job","url":"https://addaly.com/learn/ai-for-students/which-tool-for-which-job"},"next":{"slug":"a-working-agreement-with-yourself","title":"A working agreement with yourself","url":"https://addaly.com/learn/ai-for-students/a-working-agreement-with-yourself"},"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) 'Getting set up for nothing', in AI for Students. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-for-students/getting-set-up-free (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Getting set up for nothing","courseTitle":"AI for Students","module":{"index":1,"slug":"what-the-machine-is-doing","title":"What the machine is doing when it helps you"},"answer":null,"keyIdea":"Free tiers, Hugging Face Spaces, Colab, and a small local model on Ollama cover almost everything in this course, so the real skill is spending a limited allowance on the questions that are actually hard.","hero":{"kind":"plate","image":null,"alt":null,"caption":null},"blocks":[{"t":"h2","id":"the-constraint-most-advice-ignores","runs":[{"text":"The constraint most advice ignores"}]},{"t":"p","runs":[{"text":"A lot of study advice assumes a laptop with 16GB of RAM, a stable connection and a card on file. Plenty of students have a shared machine, an older phone, patchy data and no card. None of that stops you doing any of this course. It changes which route you take, and the routes are worth knowing before you need them."}]},{"t":"h2","id":"free-access-to-capable-models","runs":[{"text":"Free access to capable models"}]},{"t":"p","runs":[{"text":"Every major provider runs a free tier with a daily or hourly message cap. The caps move; the shape does not. Practical consequences:"}]},{"t":"ul","items":[[{"text":"Spend your allowance on the hard questions.","b":true},{"text":" Do the easy rewriting with a weaker free model or on your own."}],[{"text":"Batch.","b":true},{"text":" One message containing five related questions costs one message from your allowance. Five messages cost five."}],[{"text":"Keep two accounts on different providers.","b":true},{"text":" When one runs out, switch. This is not a trick, it is what the free tiers are for."}]]},{"t":"p","runs":[{"text":"Beyond the big names: "},{"text":"Hugging Face Spaces","b":true},{"text":" hosts thousands of free demos, including chat interfaces to open-weight models, and needs no payment. "},{"text":"Google Colab","b":true},{"text":" gives free GPU time in a notebook, enough to run a mid-sized open model for an afternoon. Some universities provide staff and students with a paid tier free: check before assuming; a surprising number do and never announce it well."}]},{"t":"h2","id":"running-a-model-on-your-own-machine","runs":[{"text":"Running a model on your own machine"}]},{"t":"p","runs":[{"text":"This is the option that removes quotas, works offline, and keeps everything private. It needs RAM more than anything else."}]},{"t":"ul","items":[[{"text":"8GB RAM:","b":true},{"text":" a 3B or 4B parameter model, quantised to 4 bits, around 2-3GB on disk. Usable for rephrasing, practice questions from supplied text, vocabulary, simple code."}],[{"text":"16GB RAM:","b":true},{"text":" a 7B or 8B model, 4-5GB. Noticeably better; still well short of the hosted ones on hard reasoning."}],[{"text":"Apple silicon","b":true},{"text":" punches above its weight here because the memory is shared with the GPU."}]]},{"t":"p","runs":[{"text":"The two free tools are "},{"text":"Ollama","b":true},{"text":", which is a single install and then "},{"text":"ollama run <model>","code":true},{"text":" in a terminal, and "},{"text":"llama.cpp","b":true},{"text":", which is more work and runs on almost anything. "},{"text":"LM Studio","b":true},{"text":" gives a graphical interface if terminals are unwelcome."}]},{"t":"p","runs":[{"text":"Be honest about the gap. A 4B model run locally will hallucinate more, follow multi-step instructions worse, and give up on long documents. For a set of tasks (drilling, rewriting, quizzing you from a chapter you paste in) it is entirely adequate, and the fact that it works on a train with no signal is worth a lot."}]},{"t":"h2","id":"a-phone-only-setup-that-actually-works","runs":[{"text":"A phone-only setup that actually works"}]},{"t":"ol","start":1,"items":[[{"text":"Chat tools: all major ones have web interfaces that work in a mobile browser; no app needed."}],[{"text":"Documents: Google Docs or the free Microsoft web apps; both save continuously, which doubles as your process record."}],[{"text":"Notes: Obsidian and Joplin both have mobile apps and sync free via a folder."}],[{"text":"Maths: GeoGebra and Desmos are excellent on a phone."}],[{"text":"Reading papers: get the PDF into a reader that lets you highlight; Zotero’s mobile app is free."}],[{"text":"Voice: dictation is free on every phone, and talking through a problem out loud, then pasting the transcript, is a good way to think."}]]},{"t":"p","runs":[{"text":"The thing to avoid on a phone is writing long prose directly into a chat window. It encourages accepting whatever comes back because editing is painful. Draft in a document, paste in, paste back."}]},{"t":"h2","id":"data-and-working-offline","runs":[{"text":"Data, and working offline"}]},{"t":"p","runs":[{"text":"If data is expensive or intermittent:"}]},{"t":"ul","items":[[{"text":"Download the material once. Lecture slides, past papers, PDFs. Read offline."}],[{"text":"Batch your questions. Write them in a note during the day, ask them together when you have signal."}],[{"text":"A local model needs no connection at all after the download."}],[{"text":"Wikipedia has offline packages via "},{"text":"Kiwix","b":true},{"text":": the whole of English Wikipedia without images is a few tens of gigabytes, and a subject-specific selection is far smaller."}]]},{"t":"h2","id":"what-not-to-spend-money-on","runs":[{"text":"What not to spend money on"}]},{"t":"p","runs":[{"text":"There is a large market in AI study apps charging monthly for a wrapper around a model you can use free. Before paying for anything, ask what it does that a free chat window plus a document does not. Sometimes the answer is real: a good spaced-repetition scheduler, for instance, though "},{"text":"Anki","b":true},{"text":" is free on desktop and Android and is still the best one. Usually the answer is a nicer interface around the same free capability."}]},{"t":"p","runs":[{"text":"The honest exception is a paid tier of a major model when you are doing heavy work in a specific week. It is bought by the month and cancellable, and one month during a dissertation is a different decision from a standing subscription all year."}]}],"check":{"question":[{"text":"A student with an 8GB laptop and no budget installs a 4-bit 3B model with Ollama. Which task is it genuinely well suited to?"}],"options":[{"text":[{"text":"Checking whether a set of references from an essay actually exist."}],"correct":false,"tempting":[{"text":"Existence questions need a database, and a small local model has weaker compressed knowledge than a hosted one, making fabrication more likely rather than less."}]},{"text":[{"text":"Generating practice questions from a chapter the student pastes in, offline."}],"correct":true,"tempting":null},{"text":[{"text":"Answering questions about last month’s changes to the exam syllabus."}],"correct":false,"tempting":[{"text":"Any model without retrieval, local or hosted, has no access to recent information; the local one also has no search to fall back on."}]},{"text":[{"text":"Explaining an unfamiliar advanced topic that is not in the student’s materials, since local models are trained on the same data."}],"correct":false,"tempting":[{"text":"Assumes parameter count does not matter for knowledge; a 3B model holds a far coarser compression of its training text than a large hosted one, and unfamiliar advanced material is exactly where that shows."}]}],"order":[3,2,1,0],"why":null},"prev":{"slug":"which-tool-for-which-job","title":"Which tool for which job","minutes":9},"next":{"slug":"a-working-agreement-with-yourself","title":"A working agreement with yourself","minutes":7}}}