Prominent AI researcher Andrej Karpathy picks Anthropic over former home OpenAI to get back into frontier LLM research

2026-05-19 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

Andrej Karpathy, a prominent AI researcher, is joining Anthropic's pretraining team as of May 19, 2026. He will focus on building the strongest possible base models, which are then fine-tuned for specific tasks like reasoning, coding, or math. Karpathy plans to establish his own pretraining team at Anthropic, utilizing Claude to accelerate pretraining research, betting on the idea that AI models can improve themselves exponentially. This move marks his return to frontier large language model (LLM) research after recently working on AI in education with Eureka Labs. Karpathy, a key figure in AI, previously worked at OpenAI in its early days, then led Tesla's Autopilot development, and returned to OpenAI before leaving in 2024. His decision to join Anthropic over OpenAI is seen as a significant gain for his new employer.

Key takeaway

For AI Directors and researchers evaluating frontier LLM development, Andrej Karpathy's move to Anthropic underscores the strategic importance of foundational pretraining and AI self-improvement. You should consider investing in robust base model development and exploring how existing LLMs can accelerate your pretraining research. This shift highlights a significant talent acquisition in the competitive LLM landscape.

Key insights

Andrej Karpathy's move to Anthropic signals a strategic bet on AI self-improvement for foundational LLM pretraining.

Principles

• AI progress can compound exponentially.
• Strong base models are crucial for fine-tuning.
• Agentic AI for coding shows rapid progress.

Method

Focus on building robust base models, then fine-tune using reinforcement learning for specific tasks like reasoning or coding, leveraging existing LLMs to accelerate pretraining research.

In practice

• Explore Claude for pretraining acceleration.
• Invest in foundational model development.
• Monitor agentic AI for coding advancements.

Topics

• Andrej Karpathy
• Anthropic
• Large Language Models
• LLM Pretraining
• AI Agents
• Claude

Best for: Investor, Research Scientist, AI Scientist, Director of AI/ML, Tech Journalist

Related on AIssential

• Andrej Karpathy Joins Anthropic: What Happens Next — Karpathy's move to Anthropic signals a critical industry shift towards AI-driven recursive self-improvement and automated research.
• Andrej Karpathy Leaves His Own Startup to Join Anthropic — Andrej Karpathy joins Anthropic to lead AI-assisted pre-training research, aiming to optimize compute efficiency for frontier LLMs.
• Anthropic Steals OpenAI's Crown — The AI landscape is rapidly diversifying with advanced models, specialized tools, and integrated agent platforms.
• MLWhiz Weekly Recsys/ML/GenAI Newsletter # 7 - The week Karpathy Joined Anthropic — Top AI talent and significant capital are rapidly consolidating around Anthropic, signaling a strategic industry shift.
• OpenAI VS Anthropic — Frontier AI labs are rapidly advancing generative coding, focusing on agentic systems and self-improvement.
• BREAKING: Karpathy Joins Anthropic's Pretraining Team - A Devastating Blow to OpenAI — Top AI talent is increasingly moving to Anthropic, signaling a shift in frontier model development strategy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.