Andrej Karpathy

Highlights: Karpathy responds to criticism about overhyping a popular site.

Worth reading: Reflects his engagement with public discourse on AI hype.

Highlights: Karpathy used an LLM to refine a blog post over 4 hours, showcasing iterative AI-assisted writing.

Worth reading: Demonstrates a practical workflow for leveraging LLMs to improve argument quality.

Highlights: Karpathy observes a widening gap in public understanding of AI capabilities and identifies a key issue.

Worth reading: Highlights a critical communication challenge in the AI field.

Highlights: Karpathy emphasizes the distinctiveness of the LLM stack across multiple dimensions.

Worth reading: Provides insight into the fundamental differences in LLM development compared to traditional software.

Highlights: Karpathy shares notes on using Claude for coding, focusing on workflow improvements from recent LLM advances.

Worth reading: Offers practical observations on AI-assisted coding with Claude.

Highlights: Karpathy announces the launch of Eureka Labs, an AI+Education company.

Worth reading: Marks a significant career move into AI education.

Highlights: Karpathy notes a shift from manual coding to agent-based coding due to improved LLM capabilities.

Worth reading: Demonstrates rapid adoption of AI coding agents in practice.

Highlights: Karpathy responds to criticism of overhyping a popular site.

Worth reading: Shows Karpathy's engagement with public perception of AI hype.

Highlights: Karpathy notes a growing gap in understanding AI capability.

Worth reading: Highlights public misunderstanding of AI progress.

Highlights: Karpathy observes LLMs are both smarter and dumber than expected.

Worth reading: Captures the paradoxical nature of LLM capabilities.

Highlights: Karpathy feels behind as a programmer due to AI-driven refactoring.

Worth reading: Reflects the impact of AI on software development.

Highlights: Karpathy used an LLM to refine a blog post argument over 4 hours.

Worth reading: Shows practical use of LLMs for iterative writing improvement.

Highlights: Karpathy announces joining Anthropic, emphasizing frontier LLM work and continued education passion.

Worth reading: Major talent move in AI industry, signaling Anthropic's growing influence.

Highlights: Karpathy announces his move to Anthropic, focusing on R&D while continuing his education passion.

Worth reading: Signals a major talent shift from OpenAI to Anthropic, impacting the AI landscape.

Highlights: Using LLMs to build personal knowledge bases for research topics, shifting token usage from code to knowledge manipulation.

Worth reading: Shows a practical application of LLMs for personal knowledge management.

Highlights: Shift from 80% manual coding to 80% agent coding due to improved LLM coding capabilities.

Worth reading: Illustrates the rapid adoption of AI agents in coding workflows.

Highlights: Highlights a gap in AI understanding due to outdated or limited use of AI tools.

Worth reading: Important perspective on public perception vs. actual AI capability.

Highlights: Karpathy responds to criticism of overhyping a popular site, likely related to AI or tech.

Worth reading: Provides insight into Karpathy's perspective on hype in the AI community.

Highlights: Karpathy explores bespoke software for personal health experiments, aiming to lower his resting heart rate.

Worth reading: Illustrates the potential of personalized software for health optimization, a trend in AI-driven lifestyle management.

Highlights: Karpathy uses LLMs to build personal knowledge bases for research topics, shifting token usage from code manipulation to knowledge manipulation.

Worth reading: Shows a practical application of LLMs for personal knowledge management, relevant to researchers and knowledge workers.

Highlights: Karpathy responds to criticism of overhyping a popular site.

Worth reading: Shows Karpathy's engagement with public perception of his commentary.

Highlights: Karpathy argues that LLMs benefit regular people more than experts, reversing typical tech diffusion.

Worth reading: Offers a fresh perspective on LLM societal impact.

Highlights: Karpathy coins 'vibe coding' as an AI-assisted coding style where developers rely on AI and ignore code details.

Worth reading: Introduces a popular concept shaping AI-assisted development.

Highlights: Karpathy notes a growing gap in AI understanding due to outdated or limited use of ChatGPT's free tier.

Worth reading: Highlights the disconnect between public perception and actual AI progress.

Highlights: Karpathy summarizes key paradigm shifts in LLMs during 2025, focusing on changes in the production stack.

Worth reading: Provides a high-level overview of LLM progress and infrastructure evolution.

Highlights: Karpathy describes his shift from manual coding to heavy reliance on AI agents for coding.

Worth reading: Illustrates the rapid adoption of AI coding agents in practice.

Highlights: Karpathy shares an amusing interaction with an early model version.

Worth reading: Provides insight into his experience with early LLM behavior.

Highlights: Karpathy mentions his habit of extensive reading.

Worth reading: Reflects his approach to staying informed in AI.

Highlights: Karpathy reviews paradigm changes in LLMs during 2025.

Worth reading: Summarizes key shifts in LLM landscape from a leading expert.

Highlights: Karpathy shares an amusing interaction with an AI model, possibly an earlier version.

Worth reading: Provides insight into Karpathy's experiences with AI model behavior.

Highlights: Karpathy mentions adopting a habit of extensive reading.

Worth reading: Reflects Karpathy's approach to continuous learning and information consumption.

Highlights: Karpathy expresses interest in the future of customized software.

Worth reading: Indicates Karpathy's forward-looking perspective on software development trends.

Highlights: Karpathy shares a review of LLM developments in 2025.

Worth reading: Summarizes key LLM advancements from Karpathy's perspective.

Highlights: Karpathy observes a growing gap in understanding AI capabilities, starting with a first issue.

Worth reading: Highlights a key concern about public perception versus actual AI progress.

Highlights: Karpathy shares his habit of extensive reading to stay informed.

Worth reading: Reflects his approach to continuous learning in AI.

Highlights: Karpathy recounts an amusing interaction with an early model version.

Worth reading: Provides insight into his hands-on experience with AI models.

Highlights: Karpathy discusses training LLMs using auto-generated data or objectives.

Worth reading: Reveals a technical approach to improving LLM training.

Highlights: Karpathy shares an amusing interaction with an AI model, likely referring to a chatbot or language model.

Worth reading: Provides insight into Karpathy's hands-on experience with AI models.

Highlights: Karpathy mentions developing a habit of extensive reading across various sources.

Worth reading: Reflects his approach to staying informed and learning.

Highlights: Karpathy observes a growing gap in understanding AI capabilities, pointing to a key issue.

Worth reading: Highlights a critical challenge in AI communication and education.

Highlights: Karpathy finds using LLMs to build personal knowledge bases for research topics very useful.

Worth reading: Shows a practical application of LLMs for personal knowledge management.

Highlights: Karpathy announces his new AI+Education company Eureka Labs.

Worth reading: Highlights Karpathy's latest venture in AI education.

Highlights: LLMs can be used to improve arguments, but they can also convincingly argue the opposite, revealing their persuasive power and potential pitfalls.

Worth reading: Illustrates the double-edged nature of LLMs in reasoning and argumentation.

Highlights: Natural language is becoming a dominant interface for programming, especially with LLMs.

Worth reading: Captures a key trend in AI-assisted coding and the shift towards natural language programming.

Highlights: LLMs can develop emergent reasoning-like behaviors through reinforcement learning with verifiable rewards.

Worth reading: Shows how training methods can lead to spontaneous reasoning capabilities in LLMs.

Highlights: Karpathy used an LLM to improve a blog post, but the LLM convinced him the opposite argument was true.

Worth reading: Shows how LLMs can challenge and refine reasoning, not just generate text.

Highlights: Karpathy suggests English is becoming the new programming language due to LLMs.

Worth reading: Captures a key insight about how LLMs are changing programming.

Highlights: Karpathy discusses training LLMs with auto-generated data.

Worth reading: Highlights a technique for improving LLM training efficiency.

Highlights: Karpathy recounts an amusing interaction with an early model version.

Worth reading: Shows his hands-on experience with AI model behavior.

Highlights: Identifies a key challenge in LLM personalization: memory distraction.

Worth reading: Highlights a fundamental problem in making LLMs personalized.

Highlights: Karpathy describes using an LLM to refine a blog post argument over 4 hours.

Worth reading: Demonstrates a practical use case of LLMs for writing improvement.

Highlights: Notes a growing gap in understanding AI capabilities, starting with a specific issue.

Worth reading: Reflects on public perception vs. reality of AI progress.

Highlights: Karpathy shares his habit of extensive reading across various formats.

Worth reading: Offers insight into his learning approach and information diet.

Highlights: Karpathy describes using LLMs to build personal knowledge bases, shifting his token usage from code manipulation to knowledge manipulation.

Worth reading: Illustrates a practical and novel use of LLMs for personal knowledge management.

Highlights: Karpathy responds to accusations of overhyping a popular site, likely related to AI.

Worth reading: Shows Karpathy's engagement with public discourse and his perspective on hype in AI.

Highlights: Karpathy posted a review of LLM developments in 2025, likely summarizing key trends and breakthroughs.

Worth reading: Provides a high-level perspective on the state of LLMs from a leading AI researcher.

Highlights: Karpathy advocates using LLMs to create personal knowledge bases for research.

Worth reading: Highlights a practical application of LLMs for knowledge management.

Highlights: Karpathy uses LLMs to construct personal knowledge bases, shifting his focus from code to knowledge management.

Worth reading: Illustrates a practical application of LLMs for personal productivity and research.

Highlights: Karpathy responds to criticism about overhyping a popular site, showing self-awareness about hype cycles.

Worth reading: Reveals Karpathy's perspective on the balance between excitement and realism in AI.

Highlights: Karpathy observes that LLMs are both smarter and dumber than expected, highlighting their paradoxical nature.

Worth reading: Provides a balanced perspective on LLM capabilities from a leading AI figure.

Highlights: Karpathy shares notes on using Claude for coding, reflecting on improvements in LLM-based coding workflows.

Worth reading: Offers practical insights into advanced AI-assisted coding from an expert practitioner.

Highlights: Karpathy recounts an amusing interaction with an early version of a model.

Worth reading: Provides insight into Karpathy's experience with early model capabilities.

Highlights: Karpathy notes that memory in LLMs can be distracting for personalization.

Worth reading: Highlights a key challenge in LLM personalization.

Highlights: Karpathy observes a growing gap in understanding AI capability.

Worth reading: Raises awareness about misconceptions in AI capability.

Highlights: Karpathy describes LLMs as both smarter and dumber than expected.

Worth reading: Captures the paradoxical nature of current LLM intelligence.

Highlights: Karpathy responds to accusations of overhyping a site.

Worth reading: Shows Karpathy's engagement with public perception of AI hype.

Highlights: Advocates for 'context engineering' as a more accurate term than 'prompt engineering' for industrial LLM applications.

Worth reading: Reframes how we think about optimizing LLM inputs in production.

Highlights: Using LLMs to create personal knowledge bases for research, shifting focus from code to knowledge manipulation.

Worth reading: Demonstrates a practical, personal use case for LLMs beyond coding.

Highlights: LLMs trained with verifiable rewards develop human-like reasoning strategies, breaking down problems into intermediate steps.

Worth reading: Highlights a key insight into how LLMs can learn reasoning without explicit instruction.

Highlights: Karpathy explores personalized software for health experiments, reflecting on bespoke software trends.

Worth reading: Shows how AI can enable highly personalized, data-driven self-improvement tools.

Highlights: LLMs trained with verifiable rewards develop reasoning-like behaviors, breaking problems into intermediate steps.

Worth reading: Key insight into how reinforcement learning can elicit reasoning in LLMs without explicit programming.

Highlights: Karpathy advocates using LLMs to create personal knowledge bases, shifting focus from code to knowledge management.

Worth reading: Highlights a practical application of LLMs for personal knowledge organization and research.

Highlights: Karpathy expresses feeling behind as a programmer due to rapidly advancing AI tools, but sees potential for 10X productivity.

Worth reading: Reflects the challenge and opportunity of integrating modern AI tools into programming workflows.

Highlights: Karpathy shares a humorous anecdote about an early model interaction.

Worth reading: Insight into early model behavior and user experience.

Highlights: Karpathy describes running an auto-research experiment tuning a small chat model.

Worth reading: Shows practical application of autonomous research in LLM tuning.

Highlights: Karpathy explains how training with verifiable rewards leads to emergent reasoning-like strategies.

Worth reading: Key insight into how reasoning emerges from reward-based training.

Highlights: Karpathy identifies a gap in current LLM learning paradigms.

Worth reading: Provocative thought on future directions for LLM training.

Highlights: Karpathy bought a Mac mini to experiment with 'claws' (likely a typo for 'Claude' or 'Claw' agent), noting that Apple Store staff said they are selling well and customers are confused. He is cautious about running OpenClaw due to security concerns with vibe-coded code.

Worth reading: Shows Karpathy's hands-on approach with new AI tools and his awareness of security risks in AI-generated code.

Highlights: Karpathy posted a summary of LLM developments in 2025, likely reflecting on key trends and milestones.

Worth reading: Provides Karpathy's perspective on the state of LLMs in 2025.

Highlights: Karpathy expresses interest in bespoke software and shares a personal experiment to lower his resting heart rate using a structured approach.

Worth reading: Illustrates Karpathy's methodical mindset applied to personal health, and his interest in personalized software.

Highlights: Karpathy bought a Mac mini to experiment with 'claws' (likely a typo for 'Claude' or an agent framework) and expresses concern about running open-source code with private data.

Worth reading: Shows Karpathy's hands-on approach to AI agent development and his security concerns with community code.

Highlights: Karpathy envisions a future of highly personalized software, using his own health experiment as an example.

Worth reading: Highlights Karpathy's interest in AI-driven personalization and self-experimentation.

Highlights: Karpathy observes a growing gap in AI capability understanding, noting that many base their views on outdated free-tier ChatGPT experiences.

Worth reading: Highlights a common misconception about AI progress and the importance of staying current.

Highlights: Karpathy explains how training LLMs with verifiable rewards leads to emergent reasoning behaviors.

Worth reading: Key insight into how reinforcement learning can produce reasoning capabilities in LLMs.

Highlights: Karpathy defends against accusations of overhyping a site, acknowledging the prevalence of spam and scams.

Worth reading: Shows Karpathy's nuanced view on hype versus reality in AI-related platforms.

Highlights: Karpathy reviews major paradigm changes in LLMs during 2025, noting shifts in the production stack.

Worth reading: Provides a high-level summary of key LLM trends from a leading AI researcher.

Highlights: The video provides a practical, example-driven walkthrough of how to effectively use Large Language Models in daily life, covering everything from basic interactions to understanding pricing tiers and model selection. It demystifies the growing LLM ecosystem by showing concrete applications and explaining when to use different models.

Worth watching: Andrej Karpathy's expertise and clear teaching style make complex AI concepts accessible, offering actionable insights for both beginners and experienced users looking to optimize their LLM usage.

Highlights: This video provides a comprehensive overview of how Large Language Models like ChatGPT are developed, covering the full training stack from data collection to deployment. It also offers practical mental models for understanding their 'psychology' and optimizing their use in real-world applications.

Worth watching: Worth watching because Andrej Karpathy, a leading AI researcher, delivers an accessible yet thorough explanation that bridges technical depth with practical application insights, making complex LLM concepts understandable for general audiences.

Highlights: Karpathy reflects on the key developments in LLMs over 2025, likely discussing training against auto-generated data.

Worth reading: Provides a high-level summary of LLM progress from one of the field's leading experts.

Highlights: Karpathy launches Eureka Labs, an AI+Education venture.

Worth reading: Shows his commitment to AI in education, a key area for democratizing learning.

Highlights: This video provides a comprehensive, hands-on walkthrough of reproducing the GPT-2 (124M) model from scratch, covering network architecture, training optimization, and hyperparameter tuning based on original papers. It demonstrates the full training pipeline with practical implementation details and concludes with generated text samples to evaluate model performance.

Worth watching: Worth watching for its educational value in understanding transformer-based language model implementation and training optimization, presented by a renowned AI educator with clear, practical demonstrations.

Highlights: The video explains that tokenizers are a separate, crucial component in LLMs, using Byte Pair Encoding to translate between text and tokens. It demonstrates building the GPT tokenizer from scratch, highlighting its distinct training process and core encode/decode functions.

Worth watching: Worth watching to understand a fundamental yet often overlooked part of how LLMs process text, presented clearly by an expert in the field.

Highlights: This talk demystifies Large Language Models (LLMs) by explaining them as a new computing paradigm analogous to operating systems, where models like ChatGPT serve as the core technical component. It covers their fundamental workings, future trajectory, and unique security challenges in an accessible way for general audiences.

Worth watching: Andrej Karpathy provides a clear, foundational understanding of LLMs from one of the field's leading educators, making complex concepts accessible while addressing practical implications and security considerations that remain highly relevant.

Highlights: Karpathy suggests that natural language is becoming the dominant way to program, thanks to AI.

Worth reading: Captures a paradigm shift in programming towards AI-driven code generation.

Highlights: This video provides a hands-on coding tutorial where Andrej Karpathy builds a GPT model from scratch, implementing the transformer architecture described in 'Attention is All You Need' and connecting it to real-world applications like GPT-2/3 and ChatGPT. It demonstrates the practical implementation of autoregressive language modeling while showing GitHub Copilot (itself a GPT model) assisting in writing the code, creating a meta-learning experience.

Worth watching: It's worth watching because it demystifies complex AI concepts through clear, practical coding examples and connects theoretical papers to real implementations, making advanced transformer architectures accessible to developers and enthusiasts.

Highlights: This video demonstrates how to evolve a simple 2-layer MLP into a deeper, tree-like neural network architecture that resembles DeepMind's WaveNet (2016). It shows the practical implementation process using PyTorch's torch.nn module while explaining the underlying mechanics of deep learning development.

Worth watching: It's worth watching because it provides a clear, hands-on walkthrough of building a complex neural network from simpler components, offering valuable insights into both PyTorch fundamentals and the architectural thinking behind influential models like WaveNet.

Highlights: This video demonstrates manual backpropagation through a complete 2-layer MLP with BatchNorm, covering gradients from cross entropy loss through embedding tables. It builds intuitive understanding of gradient flow at the tensor level, beyond scalar implementations like micrograd, while reinforcing core deep learning concepts.

Worth watching: Essential viewing for developers wanting to move beyond autograd black boxes and truly understand gradient computation in neural networks, presented by one of the field's most effective educators.

Highlights: This video examines the statistical challenges in training deep neural networks, focusing on how improperly scaled activations and gradients can cause instability. It introduces Batch Normalization as a key technique to stabilize training by normalizing layer inputs.

Worth watching: Worth watching for its practical insights into diagnosing and fixing common deep learning training issues, presented by an expert with clear visualizations of internal network behavior.

Highlights: This video demonstrates building a multilayer perceptron (MLP) for character-level language modeling, covering essential ML fundamentals like training, hyperparameter tuning, and evaluation. It provides practical insights into handling train/dev/test splits and diagnosing under/overfitting in neural networks.

Worth watching: It's worth watching for hands-on implementation of MLPs with clear explanations of core machine learning concepts, making it accessible for both beginners and practitioners looking to solidify their understanding.