{"slug": "google-colab", "name": "Google Colab", "category": "infra", "category_name": "Infrastructure and DevOps", "description": "", "recommenders": 4, "commenters": 8, "warners": 1, "score": 2.986, "rising": 0.02, "first": "2025-12-02", "latest": "2026-10-29", "why": "it cost me <$2 using Colab's pay as you go plan to train the model, which was surprisingly less than I expected.", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 4, "neutral": 4, "compare": 0, "warn": 1}, "spark": [0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1], "listed": true, "loved": [], "caveat": "", "short_quote": "it cost me <$2 using Colab's pay as you go plan to train the model, which was surprisingly less than I expected.", "kind": "", "icon": true, "domain": null, "github_repo": null, "official_url": "https://colab.research.google.com", "pricing": "", "aliases": ["Google Colab"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 804753570, "item_id": 50297098, "url": "https://news.ycombinator.com/item?id=50297098", "by": "areoform", "date": "2026-10-29", "version": null, "stance": "recommend", "quote": "Google Colab let's you just do it", "excerpt": "Fun fact. You can't use Claude to analyze this research paper. Because... bioweapons! > Opus 5.5's safeguards flagged this message. You may be seeing this for the first time on an Opus model: Opus 5.5 is more capable and has stronger safeguards as a result, which can sometimes flag biology-research-adjacent work. We're improving these safeguards to reduce the amount of incorrectly flagged messages. Send feedback or l", "context": "", "use_case": "Cloud GPU notebooks", "switched_from": "", "story": "IgE disruptor for food allergy with Claude-enabled antibody design [pdf]"}, {"id": 769528240, "item_id": 48095515, "url": "https://news.ycombinator.com/item?id=48095515", "by": "aidarbek", "date": "2026-05-11", "version": null, "stance": "recommend", "quote": "it cost me <$2 using Colab's pay as you go plan to train the model, which was surprisingly less than I expected.", "excerpt": "Hello everyone! I post-trained small LLM (Qwen Instruct 2.5 0.5B) to speak like GenZ using SFT+RL. Training was done in Google Colab using the cheapest GPU runtime. Honestly, you could probably get better results with simple prompting of frontier models, as I wasn\u2019t optimizing for results with the smallest model, cheapest GPU and synthetic data. But it was a good and fun learning exercise: how you can actually run RL", "context": "", "use_case": "cheap model training", "switched_from": "", "story": "Show HN: LLM post-training to speak like GenZ, costing less than a cup of coffee"}, {"id": 737937075, "item_id": 46121067, "url": "https://news.ycombinator.com/item?id=46121067", "by": "mark_l_watson", "date": "2025-12-02", "version": null, "stance": "recommend", "quote": "I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc.", "excerpt": "I used DeepSeek-v3.2 to solve two coding problems by pasting code and directions as one large prompt into a chat interface and it performed very well. VERY WELL! I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc. Google\u2019s moat for me is all the tooling and engineering around the models. That said, my one year Google AI subscription ends in four months and I mi", "context": "retired 'gentleman scientist'", "use_case": "", "switched_from": "", "story": "DeepSeek-v3.2: Pushing the frontier of open large language models [pdf]"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 739757266, "item_id": 46234829, "url": "https://news.ycombinator.com/item?id=46234829", "by": "drscotthawley", "date": "2025-12-11", "version": null, "stance": "warn", "quote": "Colab takes too long to setup and then kicks us off.", "excerpt": "Author here. Hey, My students and I could really use your help! Can anyone help shed light on why the MPS backend for PyTorch produces different numbers compared to the CUDA & CPU devices? I don't mean unsupported ops & CPU fallback, I mean fast, garbage numbers coming out of MPS. This PR references numerous other PyTorch issues related to MPS inaccuracy: https://github.com/Stability-AI/stable-audio-tools/pull/225 Th", "context": "", "use_case": "", "switched_from": "", "story": "Flow Where You Want \u2013 Guidance for Flow Models"}]}], "who": [{"by": "mark_l_watson", "context": "retired 'gentleman scientist'", "url": "https://news.ycombinator.com/item?id=46121067", "date": "2025-12-02"}], "tags": [], "use_cases": ["cloud gpu notebooks", "running whisper on cloud gpus", "cheap model training", "running quantum notebooks", "ssh access during shutdowns", "checking gpu hardware"], "compared": [{"slug": "gemini", "name": "Gemini", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "pytorch", "name": "PyTorch", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=50294486", "title": "IgE disruptor for food allergy with Claude-enabled antibody design [pdf]", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50018918", "title": "Ask HN: How do you turn videos into searchable text without uploading them?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48220105", "title": "Google officially announces that ads will be included in AI Mode search results", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48095446", "title": "Show HN: LLM post-training to speak like GenZ, costing less than a cup of coffee", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47317360", "title": "Is this the right approach to go from zero to designing quantum algorithms?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=46713444", "title": "From Stealth Blackout to Whitelisting: Inside the Iranian Shutdown", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {"recommend": 1, "neutral": 1, "warn": 1}, "2026-01": {"recommend": 1}, "2026-02": {}, "2026-03": {"neutral": 1}, "2026-04": {}, "2026-05": {"recommend": 1, "neutral": 1}, "2026-06": {}, "2026-07": {}, "2026-08": {}, "2026-09": {}, "2026-10": {"neutral": 1, "recommend": 1}}, "versions": [], "version_unnamed": 0, "quotes": [{"id": 804753570, "item_id": 50297098, "url": "https://news.ycombinator.com/item?id=50297098", "by": "areoform", "date": "2026-10-29", "version": null, "stance": "recommend", "quote": "Google Colab let's you just do it", "excerpt": "Fun fact. You can't use Claude to analyze this research paper. Because... bioweapons! > Opus 5.5's safeguards flagged this message. You may be seeing this for the first time on an Opus model: Opus 5.5 is more capable and has stronger safeguards as a result, which can sometimes flag biology-research-adjacent work. We're improving these safeguards to reduce the amount of incorrectly flagged messages. Send feedback or l", "context": "", "use_case": "Cloud GPU notebooks", "switched_from": "", "story": "IgE disruptor for food allergy with Claude-enabled antibody design [pdf]"}, {"id": 800302945, "item_id": 50018934, "url": "https://news.ycombinator.com/item?id=50018934", "by": "apresmoi", "date": "2026-10-09", "version": null, "stance": "neutral", "quote": "I have some Google Collab notebooks where I download/upload the videos", "excerpt": "I have some Google Collab notebooks where I download/upload the videos, load whisper there and use the cloud GPUs to process them! Here is an example: https://github.com/apresmoi/jianglens/blob/main/ops/notebooks/colab/Whisper_Transcription.ipynb", "context": "", "use_case": "running Whisper on cloud GPUs", "switched_from": "", "story": "Ask HN: How do you turn videos into searchable text without uploading them?"}, {"id": 771558067, "item_id": 48222379, "url": "https://news.ycombinator.com/item?id=48222379", "by": "mark_l_watson", "date": "2026-05-21", "version": null, "stance": "neutral", "quote": "occasionally pay for Colab", "excerpt": "I agree on search, really like DDG. I only rely on Google products that I pay for (Gemini APIs, YouTube Premium, occasionally pay for Colab, purchased Books on Play).", "context": "", "use_case": "", "switched_from": "", "story": "Google officially announces that ads will be included in AI Mode search results"}, {"id": 769528240, "item_id": 48095515, "url": "https://news.ycombinator.com/item?id=48095515", "by": "aidarbek", "date": "2026-05-11", "version": null, "stance": "recommend", "quote": "it cost me <$2 using Colab's pay as you go plan to train the model, which was surprisingly less than I expected.", "excerpt": "Hello everyone! I post-trained small LLM (Qwen Instruct 2.5 0.5B) to speak like GenZ using SFT+RL. Training was done in Google Colab using the cheapest GPU runtime. Honestly, you could probably get better results with simple prompting of frontier models, as I wasn\u2019t optimizing for results with the smallest model, cheapest GPU and synthetic data. But it was a good and fun learning exercise: how you can actually run RL", "context": "", "use_case": "cheap model training", "switched_from": "", "story": "Show HN: LLM post-training to speak like GenZ, costing less than a cup of coffee"}, {"id": 757077776, "item_id": 47317361, "url": "https://news.ycombinator.com/item?id=47317361", "by": "DARSHANFOFADIYA", "date": "2026-03-09", "version": null, "stance": "neutral", "quote": "The Colab notebook includes one-click access to IBM's real quantum hardware", "excerpt": "I started this series knowing nothing about quantum computing \u2014 no physics degree, no quantum mechanics background. Just curiosity, linear algebra, and a willingness to work through the math with AI as a thinking partner. The goal is ambitious: 14 parts, from \"what is a qubit\" to designing original quantum algorithms. Every equation derived from scratch. Every algorithm traced gate by gate. Working code throughout. P", "context": "", "use_case": "Running quantum notebooks", "switched_from": "", "story": "Is this the right approach to go from zero to designing quantum algorithms?"}, {"id": 747520528, "item_id": 46720033, "url": "https://news.ycombinator.com/item?id=46720033", "by": "RandyOrion", "date": "2026-01-22", "version": null, "stance": "recommend", "quote": "", "excerpt": "There is a dedicated thread of record, discussion, as well as some resources of the Iran Network Shutdown [1]. Hope it helps. [1] https://github.com/net4people/bbs/issues/561", "context": "", "use_case": "SSH access during shutdowns", "switched_from": "", "story": "From Stealth Blackout to Whitelisting: Inside the Iranian Shutdown"}, {"id": 739757266, "item_id": 46234829, "url": "https://news.ycombinator.com/item?id=46234829", "by": "drscotthawley", "date": "2025-12-11", "version": null, "stance": "warn", "quote": "Colab takes too long to setup and then kicks us off.", "excerpt": "Author here. Hey, My students and I could really use your help! Can anyone help shed light on why the MPS backend for PyTorch produces different numbers compared to the CUDA & CPU devices? I don't mean unsupported ops & CPU fallback, I mean fast, garbage numbers coming out of MPS. This PR references numerous other PyTorch issues related to MPS inaccuracy: https://github.com/Stability-AI/stable-audio-tools/pull/225 Th", "context": "", "use_case": "", "switched_from": "", "story": "Flow Where You Want \u2013 Guidance for Flow Models"}, {"id": 739360480, "item_id": 46210030, "url": "https://news.ycombinator.com/item?id=46210030", "by": "spwa4", "date": "2025-12-09", "version": null, "stance": "neutral", "quote": "go to colab.research.google.com", "excerpt": "Google is using nVidia GPUs. More than that, I'd expect Google to still be something like 90% on nVidia GPUs. You can't really check of course. Maybe I'm an idiot and it's 50%. But you can see how that works: go to colab.research.google.com. Type in some code ... \"!nvidia-smi\" for instance. Click on the down arrow next to \"connect\", and select change runtime type. 3 out of 5 GPU options are nVidia GPUs. Frankly, unle", "context": "", "use_case": "Checking GPU hardware", "switched_from": "", "story": "After the Bubble"}, {"id": 737937075, "item_id": 46121067, "url": "https://news.ycombinator.com/item?id=46121067", "by": "mark_l_watson", "date": "2025-12-02", "version": null, "stance": "recommend", "quote": "I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc.", "excerpt": "I used DeepSeek-v3.2 to solve two coding problems by pasting code and directions as one large prompt into a chat interface and it performed very well. VERY WELL! I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc. Google\u2019s moat for me is all the tooling and engineering around the models. That said, my one year Google AI subscription ends in four months and I mi", "context": "retired 'gentleman scientist'", "use_case": "", "switched_from": "", "story": "DeepSeek-v3.2: Pushing the frontier of open large language models [pdf]"}]}