DeepMind Alumni Attract Investment with Startups Beyond Large Language Models
Researchers originating from DeepMind are raising billions of dollars in funding for their new startups developing reasoning, reinforcement learning, and world models.
Experts leaving DeepMind are successfully drawing massive investor interest with artificial intelligence startups that move beyond large language models to focus on reasoning, reinforcement learning, and world models.
Metis Reasoning and New Startups
Thore Graepel, one of the names behind AlphaGo, left Alphabet's research lab in the summer to seek tens of millions of dollars from a small group of investors for his new startup, Metis Reasoning.
Metis Reasoning develops systems capable of responding to unknown problems for use in robotics, science, and engineering to enable machines to plan and act under uncertainty.
Reinforcement Learning and World Models
David Silver, who led DeepMind's reinforcement learning team, raised $1.1 billion at a $5.1 billion valuation for Ineffable Intelligence in April with the backing of Sequoia, Lightspeed, Nvidia, and Google.
AMI Labs, founded by Yann LeCun, secured $1.03 billion in funding in March at a $3.5 billion valuation to build world models.
Emulate and Other Developments in the Sector
Emulate, founded in August by three DeepMind veterans, is conducting investment talks reaching up to 700 million at a $3.7 billion valuation.
These startups in the artificial intelligence ecosystem clearly demonstrate that funding pools are shifting toward areas distinct from classical language models.
OpenAI's GPT-6 Cyber Model
Within the scope of other developments in the sector, OpenAI is preparing to launch its latest cybersecurity-focused model, GPT-6 Cyber.
The company is also preparing a new gateway product that will enable users to securely deploy this model, automate workflows, and address security vulnerabilities.