Former OpenAI Chief Technology Officer Establishes New Team
On February 18, former OpenAI Chief Technology Officer Mira Murati announced the establishment of “Thinking Machines Lab,” a team composed of 30 researchers and engineers from OpenAI, Meta, and Mistral, among others.
Unlike other companies founded by former OpenAI executives, Thinking Machines Lab does not solely pursue the elusive Holy Grail of AI—the AGI (Artificial General Intelligence)—but aims to popularize AI applications, placing greater emphasis on AI safety and transparency.
“We not only want to create fully autonomous AI systems but also aim to develop AI systems that can closely collaborate with humans,” stated Thinking Machines Lab. “Existing AI systems are impressive in programming and mathematical calculations, but we aspire to create AI with a broader range of applications.”
Mira Murati on Founding Thinking Machines Lab: Aiming for More Human-Centric AI Systems?
“I co-founded this company with a group of outstanding scientists, engineers, and developers.” Regarding the goals of Thinking Machines Lab, Murati pointed out that in addition to allowing users to customize AI systems to meet individual needs, they also hope to develop more powerful AI systems and promote a culture of open science.
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
– Fostering a…— Mira Murati (@miramurati) February 18, 2025
Murati emphasized that a significant issue in the current AI landscape is that while AI systems are powerful, they are often difficult for the average person to understand and use. “The scientific community’s understanding of cutting-edge AI systems has not kept pace with the rapid development of AI functionalities. The training methods of these systems are largely confined to top research laboratories, which limits public discourse on AI and the general ability to utilize AI,” stated Thinking Machines Lab.
Although Thinking Machines Lab has yet to release any products or models, the team’s goal is to develop “multimodal” AI systems that can collaborate with humans. “We not only want to develop fully autonomous AI systems but are committed to building multimodal systems that can work alongside humans.” The team emphasized that such systems focus more on human-machine collaboration and can adapt to various professional fields. They mentioned that they are developing AI models in areas like science and programming to facilitate new scientific discoveries and engineering innovations.
AI safety and transparency are also core objectives of Thinking Machines Lab. Compared to other competitors, the team stated that they will enhance safety through multiple approaches, including sharing best practices and safety development guidelines with the industry. The research and product teams will adopt a “co-design” model and commit to maintaining research transparency through shared code, datasets, and model specifications to accelerate external research.
In contrast to other competitors like OpenAI, the team stated that the research and product teams will adopt a “co-design” model and commit to maintaining research transparency through shared code, datasets, and model specifications to accelerate external research.
Image / DALL·E
Furthermore, Thinking Machines Lab emphasizes that the value of AI should be reflected in its contributions to the real world. They believe that the most important breakthroughs often come from rethinking objectives rather than merely optimizing existing metrics.
Mira Murati: Background of a Former Tesla and OpenAI Figure
Before joining OpenAI, Murati worked at Tesla for three years as a senior product manager for the Model X, during which Tesla released early versions of its Autopilot driving assistance system. In 2018, Murati joined OpenAI and was promoted to Chief Technology Officer in 2022, leading the development of ChatGPT, DALL-E, and Codex. During Sam Altman’s brief ousting from OpenAI’s board in November 2024, she served as the interim CEO of OpenAI and supported Altman’s return during the turmoil, leaving OpenAI in October of the same year.
Today, we are excited to announce Thinking Machines Lab (https://t.co/gD5QlPMfWw), an artificial intelligence research and product company. We are scientists, engineers, and builders behind some of the most widely used AI products and libraries, including ChatGPT,…— Thinking Machines (@thinkymachines) February 18, 2025
Currently, the team at Thinking Machines Lab consists of about 30 people, including researchers from AI giants like Google DeepMind, Character.ai, and Mistral, as well as many former high-level executives from OpenAI. In addition to having former OpenAI Chief Researcher Barret Zoph as the CTO, OpenAI co-founder John Schulman serves as the Chief Scientist.
References: Tech Crunch, The New York Times, Reuters, The Verge, AXIOS
This article is collaboratively republished from: Startup Gathering.