Ex-OpenAI researcher Jerry Tworek launches Core Automation to build the most automated AI lab in the world

2026-04-22 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, quick

Summary

Former OpenAI researcher Jerry Tworek has launched Core Automation, an AI lab aiming to be the "most automated AI lab in the world" by automating its own research processes. The lab focuses on developing novel learning algorithms beyond current pre-training and reinforcement learning paradigms, alongside new architectures designed to scale more effectively than transformers. Core Automation's team comprises experts in frontier models, optimization, and systems engineering, envisioning small teams augmented by AI agents to accomplish work traditionally requiring large organizations. Tworek departed OpenAI in January 2026 after seven years, citing a lack of opportunity for fundamental research there and declaring deep learning research "is done." Core Automation joins other "Neo Labs" founded by OpenAI alumni, all pursuing fundamentally new approaches to AI progress.

Key takeaway

For research scientists and entrepreneurs evaluating the future of AI development, Core Automation's launch signals a significant shift away from incremental deep learning advancements. You should consider exploring fundamental algorithmic and architectural innovations, as the field's leading minds are increasingly betting on entirely new approaches rather than just scaling existing models.

Key insights

New AI labs are pursuing fundamental algorithmic and architectural shifts beyond current deep learning paradigms.

Principles

• Automate research to accelerate AI progress
• Focus on new learning algorithms
• Design architectures for better scaling

In practice

• Explore alternatives to transformer architectures
• Investigate learning beyond pre-training
• Integrate AI agents into research workflows

Topics

• Core Automation
• Jerry Tworek
• Automated AI Research
• New Learning Algorithms
• AI Architectures

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

Related on AIssential

• GPT-5.5: The paper is important because it shows an early version of AI becoming part of the machinery that improves AI. It does not prove that fully autonomous science has arrived. — AI systems can now automate and accelerate core aspects of AI research and development.
• this EX-OPENAI RESEARCHER just released it... — AI agents can autonomously conduct machine learning research, improving model training and performance.
• OpenAI loses top AI researcher Jerry Tworek after seven years — A key OpenAI researcher departed, citing a desire to pursue research areas not prioritized by the company.
• From Factory Precision to Intelligent Robots: How Real-World Performance Shapes Industrial Edge AI — Production-ready industrial edge AI demands deterministic performance and integrated, concurrent compute for diverse workloads.
• 🔮 Exponential View #565: Autoresearch; the solar supercycle; an agentic nation; ChatGPT Olympian, seeing fraud & moving asteroids++ — Declining solar costs, rapid AI agent adoption, and AI-driven research automation are reshaping global energy and technology landscapes.
• On theCUBE Pod: IBM Think goes AI first and Musk buries the hatchet with Anthropic — AI is transforming enterprise operating models, creating new roles and intensifying the demand for compute and integrated platforms.

Open in AIssential →

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