AI researcher and educator
Andrej Karpathy
Andrej Karpathy is an AI researcher and educator, a founding member of OpenAI, and a former director of AI at Tesla working on computer vision.
About Andrej Karpathy
Andrej Karpathy is an AI researcher and educator whose public teaching covers neural networks and large language models. His educational videos include both technical lessons on building models and introductions for a general audience.
He was a founding member and research scientist at OpenAI, led Tesla's AI and computer vision work from 2017 to 2022, and returned to OpenAI in 2023 to work on midtraining and synthetic data. During his PhD at Stanford, he designed and taught the CS231n course on deep learning for visual recognition.
Mentioned in our coverage
79 articles name Andrej Karpathy, excluding author credits.
Page 7 of 7
- oh-my-pi (omp): the batteries-included terminal coding agent that gets edits right the first time →
Mario Zechner's Pi, extended with production-grade harness engineering: hashline edits that eliminate whitespace battles, LSP-driven renames, lldb/dlv/debugpy integration, 14-provider web_search, and task isolation across APFS/btrfs/overlayfs—5.5k stars and climbing.
- How Gary Tan Shipped 400x More Code Using Claude Code & AI Agents →
From 13-year coding hiatus to building open-source projects with 100K+ stars — while running Y Combinator...
- What is CLAUDE.md? Persistent Memory That Transforms Claude Code Sessions →
CLAUDE.md is Claude Code's onboarding script—a markdown file that gives the model persistent memory about your project, eliminating the need to re-explain your stack, conventions, and preferences in every session.
- Agent harness engineering: when the model stays fixed and the scaffolding wins →
The viral 2026 narrative is grounded in public numbers: benchmark gains from prompts, tools, and middleware—not a model swap. Here is what an agent harness is, who proved it, and how teams decide depth.
- Agentic fatigue meets vibe coding: the AI developer productivity paradox (2026) →
Social feeds show ambitious builders 'fully cooked' by mid-afternoon despite AI leverage. Token spend surges 13×, context switching exhausts cognition, and vibe-coded apps collapse under their own weight. Here is the paradox, the economics, and the escape hatch.
- What is a context window? LLM 'working memory' and a 2026 snapshot of top models →
Context length is the cap on 'how much the model can read at once,' not the same as how many parameters it has. This guide defines the window, input vs max output, long-context tradeoffs, and what Anthropic, OpenAI, Google, and Meta publish today.