ChapterPal

Super proud to be able to say that the famous Attention is All You Need paper, the one that started the current unprecedented wave of interested towards AI, cited 237,078 times since 2017, is now available on @ChapterPal. Enjoy reading a great paper with an AI tutor: chapterpal.com/s/9b8yltb1/att…

I added a new feature to @ChapterPal : "Simplify!" If you are using ChapterPal to read complex literature, you know that when you select some text in the reader, a contextual menu with "Why?", "How?", "Explain", and "Example" buttons appears. Now, the menu also contains a "Simplify" button. When you press this button, the Q&A that appears below the selection contains the content of the selected paragraphs rewritten in simple English. The simplified version doesn't make the text more abstract. Instead, it provides the same level of detail as the original but without assuming that the reader is an expert in the field. If the selected text uses some jargon experts are supposed to understand, the simplified text would introduce these concepts in an intuitive way. Try this new feature and don't hesitate to send your feedback to feedback@chapterpal.com.

A new learning curriculum on @ChapterPal: Must-read papers in 3D generation and neural radiance fields This curriculum traces the evolution of neural rendering and 3D reconstruction from foundational concepts to state-of-the-art large-scale models, beginning with core techniques like neural radiance fields and progressing through viewpoint-conditioned diffusion models before culminating in massive transformer-based architectures for single-image 3D reconstruction. Learners start by understanding how neural networks can represent 3D scenes through continuous functions and view synthesis, then advance to how diffusion models can condition generation on camera viewpoints and leverage Internet-scale pre-training to dramatically improve reconstruction quality. The sequence culminates with Large Reconstruction Models that scale transformer architectures to hundreds of millions of parameters, enabling real-time 3D object prediction from single images trained on massive multi-view datasets of over one million objects. By following this progression, learners gain practical knowledge of neural rendering fundamentals, modern generative techniques for 3D content, and how scaling both model capacity and training data dramatically improves generalization across diverse real-world and synthetic inputs. Learn from the best with AI tutor: chapterpal.com/curriculum/17e…

I started a new series of curricula on @ChapterPal: Prep reading for AI/ML interviews for different roles. The first curriculum in the series: **Prep reading for the model training infrastructure interview** This curriculum provides a journey through the essential landscape of AI model training infrastructure. Learners will gain a deep understanding of fundamental and cutting-edge techniques for optimizing, scaling, and managing large-scale deep learning model training, equipping them with the knowledge to excel in their job interview. Learn with an AI tutor: chapterpal.com/curriculum/213…

A new curriculum on @ChapterPal: Must-read papers on diffusion language models Diffusion language models represent a fundamental reimagining of how machines generate text — moving away from the autoregressive, left-to-right token prediction that has dominated the field, toward an iterative denoising process borrowed from continuous diffusion in vision. Rather than committing to each word sequentially, these models learn to refine entire sequences from noise, enabling parallel generation, flexible conditioning, and novel forms of control over the output. Though the paradigm originated in image synthesis, adapting it to the discrete, structured nature of language has required deep innovations in noise schedules, representation, and training objectives — innovations that are still actively unfolding. By reading the papers of this curriculum, learners will master the foundational theories, innovative architectures, and critical techniques of diffusion language models, enabling them to master and advance the science of artificial language generation from first principles. Learn from the best with AI tutor: chapterpal.com/curriculum/aef…

A new curriculum on @ChapterPal: Must-read paper on neural video synthesis This curriculum provides the essential knowledge to learn modern neural video synthesis from its foundational principles. Learners will master techniques spanning generative adversarial networks, autoregressive models, and the transformative diffusion models, gaining the expertise to synthesize photorealistic, controllable, and long-form video content, crucial for building advanced visual AI. Read from the best with AI tutor: chapterpal.com/curriculum/fe6…

A new curriculum on @ChapterPal: Must-read papers on autoregressive generation This curriculum provides a foundational journey through the most critical innovations in autoregressive generation, tracing a precise intellectual lineage from early neural language models to the large-scale systems that define the current frontier. Learners will gain a deep understanding of the core mathematical principles, architectural breakthroughs, and practical techniques essential for building and advancing intelligent sequential systems from first principles — from distributed representations and attention mechanisms through scaling laws, alignment techniques, and inference optimizations that make modern generative AI both capable and deployable. Learn from the best with an AI tutor: chapterpal.com/curriculum/29b…

A new curriculum on @ChapterPal: Must read papers on variational autoencoders Variational Autoencoders (VAEs) are a class of generative models that learn compact, meaningful representations of data while enabling the creation of new, realistic samples. By combining principles from deep learning and probabilistic modeling, VAEs provide a structured framework for understanding complex data distributions, making them essential for tasks such as image generation, anomaly detection, representation learning, and latent space exploration. This curriculum provides a foundational and advanced understanding of Variational Autoencoders, equipping the learner with the critical knowledge to build and innovate upon this powerful class of generative models. Learners will trace the evolution of VAEs from their inception to their state-of-the-art applications, mastering the core principles and key advancements. Learn from the best on ChapterPal: chapterpal.com/curriculum/c57…

A new curriculum on @ChapterPal: Must-read papers on generative adversarial networks This curriculum distills the essential breakthroughs in Generative Adversarial Networks, providing the student with enough knowledge to build this powerful technology from scratch. Learners will gain a comprehensive understanding of GAN theory, architectures, training stabilization techniques, and advanced applications for image synthesis and translation. Learn from the best with an AI tutor: chapterpal.com/curriculum/3b5…

A new curriculum on @ChapterPal: Pruning techniques for compact and efficient neural networks This curriculum dives into state-of-the-art pruning techniques, equipping practitioners with the knowledge to significantly reduce model size and computational cost without sacrificing performance. Learners will explore foundational concepts, advanced structured methods, and groundbreaking hypotheses that challenge conventional wisdom, enabling them to build highly efficient and deployable AI models. Learn from the best on ChapterPal: chapterpal.com/curriculum/315…

A new paid Reader plan on @ChapterPal. It's "PDF only" for $5.99/month. The plan includes 20 monthly PDF uploads and it doesn't do PDF to markdown conversion default on other plans. The new PDF reader looks very much like the original conversational reader except that instead of gradually revealing markdown sentences and paragraphs, the user reveals parts of the PDF: paragraphs, figures, tables, and equations. Because there's no PDF to markdown conversion, getting to reading the uploaded document is much faster, and the backend cost for me is substantially less. $5.99/month is only $1 above what Adobe charges for its AI assistant in Acrobat, but 1) the ChapterPal UX is way better, 2) includes access to a large library of high-quality books and curricula, and 3) who really wants to give Adobe money?

The biggest drawback of @ChapterPal has been the need to wait for the uploaded document to be converted into markdown before the user can start reading it. Because of the OCR involved, the waiting time could be between minutes and half an hour. I just pushed a major update to ChapterPal that allows users to start reading the document right after it is uploaded while it's being converted into markdown. The experience looks like you are reading the original PDF, except that you can stop at any moment and ask a contextual question before continuing with the PDF. As soon as the conversion into markdown is complete, the user gets a link to switch to reading a more feature-rich markdown-based version. The new feature is super raw but looks awesome, and there's almost no waiting time anymore! Try it and please send me your feedback at feedback@chapterpal.com or by leaving a comment under this post.

A new curriculum on @chapterpal: Must-read papers on diffusion models This curriculum provides a meticulously curated journey through the most critical papers that define the modern science of diffusion models. Learners will reconstruct the foundational theory, practical breakthroughs, architectural innovations, and diverse applications of diffusion models, ensuring the students' ability to learn this pivotal generative AI technology from first principles. Read the curriculum: chapterpal.com/curriculum/80c…

A new curriculum on @ChapterPal: "Must-read papers on mixture-of-experts (MoE) models" This curriculum provides a definitive journey through the evolution of Mixture-of-Experts (MoE) models, from foundational concepts to cutting-edge research. Learners will gain a comprehensive understanding of MoE architectures, training methodologies, scaling techniques, and practical applications, essential for learning this critical field of AI from scratch. Learn from the best on ChapterPal: chapterpal.com/curriculum/85c…

A new curriculum on @ChapterPal: "Must-read papers on graph neural networks" This curriculum provides a meticulously curated sequence of foundational papers, designed to enable the students to learn the entire field of Graph Neural Networks from first principles. Learners will gain a deep, chronological understanding of GNN development, from theoretical foundations and architectural innovations to advanced applications and practical considerations, ensuring the mastery of this critical AI domain. Learn everything about AI from the best: chapterpal.com/curriculum/bbf…

A new curriculum on @ChapterPal: "Must-read papers on convolutional neural networks" This curriculum provides a critical historical and technical journey through the most impactful papers that define convolutional neural networks. Learners will gain the essential knowledge to master the entire field of modern CNNs from fundamental concepts to advanced architectures, enabling sophisticated computer vision applications. Learn from the best with an AI tutor on ChapterPal: chapterpal.com/curriculum/44d…