Ex-OpenAI CTO Mira Murati Launches New Startup, Thinking Machines Lab

Mira Murati's New Venture: Thinking Machines Lab

Mira Murati, former CTO of OpenAI, has unveiled her latest project: Thinking Machines Lab. This new startup, which has just emerged from stealth mode, aims to revolutionize AI by developing tools that cater to individual needs and goals. The company's mission is to create AI systems that are not only more understandable and customizable but also more capable than what's currently on the market.

Leading the charge as CEO, Murati is joined by OpenAI co-founder John Schulman as chief scientist and Barret Zoph, formerly of OpenAI's model post-training team, as CTO. In a blog post shared with TechCrunch, Thinking Machines Lab highlighted the significant strides in AI technology, yet pointed out the "key gaps" that still exist.

"The scientific community's understanding of frontier AI systems lags behind rapidly advancing capabilities," the blog post explains. It goes on to note that the knowledge of how these systems are trained is largely confined to a few elite research labs, which restricts both public discourse and the general public's ability to effectively use AI. Additionally, despite their vast potential, these systems are often hard to tailor to individual needs and values.

Focus on Multimodal and Collaborative Systems

Thinking Machines Lab is set to tackle these issues by focusing on building "multimodal" systems designed to collaborate with people. These systems will be adaptable to a wide range of human expertise and open up new applications. "We are building models at the frontier of capabilities in domains like science and programming," the blog post states, suggesting that the most advanced models will lead to groundbreaking applications and benefits, including new scientific discoveries and engineering breakthroughs.

Commitment to AI Safety

AI safety is another cornerstone of Thinking Machines Lab's mission. The company plans to contribute to the field by preventing misuse of their models, sharing industry best practices for building safe AI systems, and supporting external research on AI alignment by providing code, datasets, and model specifications.

"We'll focus on understanding how our systems create genuine value in the real world," the blog post emphasizes. "The most important breakthroughs often come from rethinking our objectives, not just optimizing existing metrics."

Mira Murati announced the launch of Thinking Machines Lab on Twitter, stating, "I started Thinking Machines Lab alongside a remarkable team of scientists, engineers, and builders. We're building three things: Helping people adapt AI systems to work for their specific needs, developing strong foundations to build more capable AI systems, and fostering a..."

Murati's Journey and the Team

Murati left OpenAI in October after a six-year tenure, where she initially joined as VP of applied AI and partnerships in 2018. She was promoted to CTO in 2022 and played a pivotal role in developing ChatGPT, DALL-E, and Codex, the latter of which powered early versions of GitHub's Copilot. She also served as OpenAI's interim CEO briefly following Sam Altman's sudden departure, and Altman has described her as a close ally.

Rumors of Murati assembling a team of high-profile AI experts have circulated for months, and the Thinking Machines Lab blog lists 29 employees from top firms like OpenAI, Character AI, and Google DeepMind. The company is currently hiring machine learning scientists, engineers, and a research program manager.

There have been whispers of Murati seeking over $100 million in funding from unnamed VC firms, though the blog post neither confirms nor denies this. Before her time at OpenAI, Murati was a senior product manager at Tesla, working on the Model X and early versions of Autopilot. She also served as VP of product and engineering at Leap Motion.

Joining the Ranks of AI Innovators

Murati's launch of Thinking Machines Lab adds her to the growing list of former OpenAI executives venturing into new startups, including Ilya Sutskever's Safe Superintelligence and Anthropic. Her new endeavor promises to push the boundaries of AI, focusing on customization, collaboration, and safety.