Core Automation, co-founded by ex-OpenAI VP Jerry Tworek, launches to build “the world's most automated AI lab” with talent from OpenAI, Anthropic, and DeepMind
Business Insider
Context & Ripple Effects
Core Automation enters the frontier-model ecosystem with a founding team drawn from OpenAI, Anthropic, and DeepMind, positioning the company around automating the work of an AI lab rather than a single downstream application.
Related coverage subsequently describes fundraising ambitions at a multibillion-dollar valuation following an earlier raise, while another ex-frontier-lab team, Humans&, also attracted a large seed round. Together, the coverage points to investor appetite for founder-led attempts to build new AI institutions outside incumbent labs.
First-order effects
Core Automation becomes a new destination for frontier-lab talent and capital, led by former OpenAI VP Jerry Tworek and staffed with recruits from major model developers.
Its stated focus makes automation of AI research and development the company’s central product and operating thesis, rather than merely an internal efficiency program at an incumbent lab.
Second-order effects
OpenAI, Anthropic, and DeepMind face another talent-retention and recruiting pressure point as experienced researchers and operators can form competing labs with credible access to funding.
If Core Automation can automate meaningful portions of lab work, the market for research tooling and AI-development infrastructure could shift toward systems built for increasingly agentic R&D workflows.
Third-order effects
The launch is consistent with frontier AI becoming more institutionalized: alumni from leading labs are increasingly able to spin out new, heavily financed organizations that compete for the same specialized people and resources.
A sustained push to automate lab operations could change the basis of competition from model capabilities alone toward how efficiently organizations can turn compute, talent, and agent systems into research output; the extent of that shift remains unproven at launch.
The trend: Frontier AI is moving toward a new generation of capital-backed labs that seek advantage not only through models, but through automating the process of building them.
After an incredible chapter at Google DeepMind working on Gemini, I'm starting something new with some exceptional people. I've increasingly felt that the current research paradigm — scaling models, data, and static deployment won't get us all the way. We believe the next phase
Jerry Tworek (and other insanely talented people) new lab > We think the next wave of frontier research will come from small teams with highly capable agents > We're pursuing new learning algorithms that supersede large-scale pretraining and reinforcement learning, and [image]
A new AI lab built from the ground up around automated research. Led by the legend Jerry Tworek, former VP of Research at OpenAI, where he led the development of reasoning models. Absolutely incredible founding team. Heavy hitters out of OAI, Anthropic, DeepMind. Core idea: [imag…
Ok I did leave anthropic, a few weeks ago, it was one of the best places to work for a researcher. Jerry Tworek nerdsniped me into starting this with him and others The pretraining team at ant is one of the well functioning research team in the industry, and anthropic has great
A lot of the stack we use today was built for an earlier paradigm. If we want new kinds of research, we'll need new libraries, new abstractions, and new ways of running the lab. I couldn't be more excited about our founding team. We're going to any% the frontier.