[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74164":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":13,"stars7d":15,"stars30d":16,"stars90d":14,"forks30d":14,"starsTrendScore":17,"compositeScore":18,"rankGlobal":9,"rankLanguage":9,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":14,"starSnapshotCount":14,"syncStatus":13,"lastSyncTime":27,"discoverSource":28},74164,"AI-Crash-Course","henrythe9th\u002FAI-Crash-Course","henrythe9th","AI Crash Course to help busy builders catch up to the public frontier of AI research in 2 weeks",null,6089,872,112,2,0,3,17,6,72.02,"MIT License",false,"main",true,[],"2026-06-12 04:01:13","# AI-Crash-Course\nAI Crash Course to help busy builders catch up to the public frontier of AI research in 2 weeks\n\n**Intro:** I’m [Henry Shi](https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fhenrythe9th\u002F) and I started Super.com in 2016 and grew it to $150MM+ in annual revenues and recently exited. As a traditional software founder, I needed to quickly catch up to the frontier of AI research to figure out where the next opportunities and gaps were. I compiled a list of resources that were essential for me and should get you caught up within 2 weeks.\n\nFor more context, checkout the [original twitter thread](https:\u002F\u002Fx.com\u002Fhenrythe9ths\u002Fstatus\u002F1877056425454719336)\n\n**Start Here:**  \n[Neural Network \\-\\> LLM Series](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=aircAruvnKk&list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi)\n\n**Then get up to speed via Survey papers:**\n\n- Follow the ideas in the survey paper that interest you and dig deeper\n\n[LLM Survey](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2402.06196v2) \\- 2024  \n[Agent Survey](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2308.11432) \\- 2023  \n[Prompt Engineering Survey](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2406.06608) \\- 2024  \n[Context Engineering Survey](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.13334) \\- 2025\n\n**AI Papers:** (prioritize ones with star \\*)\n\n**Foundational Modelling:**  \n[**Transformers**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1706.03762) (foundation, self-attention) \\- 2017  \n[Scaling Laws](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2001.08361)\u002F[**GPT3**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2005.14165) (conviction to scale up GPT2\u002F3\u002F4) \\- 2020  \n[LoRA](https:\u002F\u002Farxiv.org\u002Fabs\u002F2106.09685) (Fine tuning) \\- 2021  \n[Training Compute-Optimal LLMs](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2203.15556) \\- 2022  \n[**RLHF**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2203.02155) (InstructGPT-\\>ChatGPT) \\- 2022  \n[DPO](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2305.18290) (No need for RL\u002FReward model) \\- 2023  \n[LLM-as-Judge](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2306.05685) (On par with human evaluations) \\- 2023  \n[MoE](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2401.04088) (MIxture of Experts) \\- 2024  \n\n**Planning\u002FReasoning:**  \n[AlphaZero](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1712.01815)\u002F[**MuZero**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1911.08265) (RL without prior knowledge of game or rules) \\- 2017\u002F2019  \n[**CoT**\\* (Chain of Thought)](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2201.11903)\u002F[ToT (Tree of Thoughts)](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2305.10601)\u002F[GoT (Graph of Thoughts)](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2308.09687)\u002F[Meta CoT](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2501.04682) \\- 2022\u002F2023\u002F2023\u002F2025  \n[ReACT](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2210.03629) (Generate reasoning traces and task-specific actions in interleaved manner) \\- 2022  \n[Let’s Verify Step by Step](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2305.20050) (Process \\> Outcome) \\- 2023  \n[**ARC-Prize**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2412.04604) (Latest methods for solving ARC-AGI problems) \\- 2024  \n[**DeepSeek R1**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2501.12948v1) (Building OSS o1-level reasoning model with pure RL, no SFT, no RM) \\- 2025  \n[Recursive Language Models](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2512.24601v1) (Simple REPL + basic tools → models emergently learn adaptive strategies without explicit prompting) \\- 2026  \n\n**Applications:**  \n[Toolformer](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2302.04761) (LLMs to use tools) \\- 2023  \n[GPT4](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2303.08774) (Overview of GPT4, but fairly high level) \\- 2023  \n[**Llama3**\\*](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2407.21783) (In depth details of how Meta built Llama3 and the various configurations and hyperparameters) \\- 2024  \n[Gemini1.5](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2403.05530) (Multimodal across 10MM context window) \\- 2024  \n[Deepseekv3](https:\u002F\u002Fgithub.com\u002Fdeepseek-ai\u002FDeepSeek-V3\u002Fblob\u002Fmain\u002FDeepSeek_V3.pdf) (Building a frontier OSS model at a fraction of the cost of everyone else) \\- 2024  \n[SWE-Agent](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2405.15793)\u002F[OpenHands](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2407.16741) (OpenSource software development agents) \\- 2024\n\n**Benchmarks:**  \n[BIG-Bench](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2206.04615) (First broad & diverse collaborative OSS benchmark) \\- 2022  \n[SWE-Bench](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2310.06770) (Real world software development) \\- 2023  \n[Chatbot Arena](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2403.04132) (Live human preference Elo ratings) \\- 2024\n\n\u003Chr \u002F>\n\n**Videos\u002FLectures:**  \n[3Blue1Brown on Foundational Math\u002FConcepts](https:\u002F\u002Fwww.youtube.com\u002F@3blue1brown)  \n[Build a Large Language Model (from Scratch) \\#1 Bestseller](https:\u002F\u002Fwww.amazon.com\u002FBuild-Large-Language-Model-Scratch\u002Fdp\u002F1633437167)\n[Build a Reasoning Model (From Scratch)](https:\u002F\u002Fwww.manning.com\u002Fbooks\u002Fbuild-a-reasoning-model-from-scratch)\n[Andrej Kaparthy: Zero to Hero Series](https:\u002F\u002Fwww.youtube.com\u002Fplaylist?list=PLAqhIrjkxbuWI23v9cThsA9GvCAUhRvKZ)  \n[Yannic Kilcher Paper Explanations](https:\u002F\u002Fwww.youtube.com\u002F@YannicKilcher)  \n[Noam Brown (o1 founder) on Planning in AI](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=eaAonE58sLU)  \n[Stanford: Building LLMs](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=9vM4p9NN0Ts)  \n[Foundations of LLMs](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2501.09223)  \n[Why You’re Not Too Old to Pivot Into AI](https:\u002F\u002Fwww.latent.space\u002Fp\u002Fnot-old) (motivation)\n\n**Helpful Websites:**  \n[History of Deep Learning](https:\u002F\u002Fgithub.com\u002Fadam-maj\u002Fdeep-learning?tab=readme-ov-file) \\- summary timeline of deeplearning with major breakthroughs and key concepts  \n[Full Stack Deep Learning](https:\u002F\u002Ffullstackdeeplearning.com\u002F) \\- courses for building AI products  \n[Prompting Guide](https:\u002F\u002Fwww.promptingguide.ai\u002F) \\- extensive list of prompting techniques and examples  \n[a16z AI Cannon](https:\u002F\u002Fa16z.com\u002Fai-canon\u002F) \\- similar list of resources, but longer (slightly dated)  \n[2025 AI Engineer Reading List](https:\u002F\u002Fwww.latent.space\u002Fp\u002F2025-papers) \\- longer reading list, broken out by focus area  \n[State of Generative Models 2024](https:\u002F\u002Fnrehiew.github.io\u002Fblog\u002F2024\u002F) \\- good simple summary of current state\n\n**Others (non LLMs):**  \n[Vision Transformer](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2010.11929) (no need for CNNs) \\- 2021  \n[Latent Diffusion](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2112.10752) (Text-to-Image) \\- 2021\n\n**Obvious\u002Feasy papers (to get your feet wet if you're new to papers):**  \n[CoT (Chain of Thought)](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2201.11903) \\- 2022  \n[SELF-REFINE: Iterative Refinement with Self-Feedback](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2303.17651) \\- 2023  \n","AI-Crash-Course 是一个为期两周的快速入门课程，旨在帮助忙碌的技术人员迅速掌握当前人工智能研究的前沿知识。该项目通过精选一系列高质量的学习资源，包括视频教程、综述论文及关键学术文章，覆盖了从神经网络到大型语言模型（LLM）、强化学习等多个核心领域。它特别适合那些希望在短时间内了解并跟上AI技术最新进展的企业家、开发者以及研究人员。此外，项目采用MIT许可协议开放源代码，鼓励社区贡献与分享。","2026-06-11 03:49:06","high_star"]