Former OpenAI CTO Mira Murati has taken a bold step forward in the world of AI with the launch of Thinking Machines Lab, an artificial intelligence research and product startup. The lab aims to create a future where AI is fully accessible, allowing everyone to access powerful tools to make AI work for their unique needs and goals.
“Our goal is simple: advance AI by making it broadly useful and understandable through solid foundations, open science, and practical applications,” Murati, who is serving as the chief operating officer of the startup, said on X.
The lab’s core leadership team also includes chief scientist John Schulman and chief technology officer Barret Zoph – both of whom are former OpenAI employees.
Inside Murati’s Lab
Headquartered in San Francisco, the startup is focused on building multimodal models for text and images, with an actual focus on human collaboration. As per reports, the startup was in talks to raise $100 million in funding last year.
The company believes that “scientific progress is a collective effort”, and with this philosophy in mind, it has committed to sharing its research and code with the larger research community. Thinking Machines Labs plans to frequently publish technical blog posts, papers, and code
Murati revealed that the company is trying to develop strong foundations to build more capable AI systems and foster a culture of open science that will help the whole field understand and improve these systems.
“While current systems excel at programming and mathematics, we’re building AI that can adapt to the full spectrum of human expertise and enable a broader spectrum of applications,” the company further stated.
Although the blog does not directly mention AGI (artificial general intelligence) or ASI (artificial superintelligence), it emphasises the importance of developing strong foundations for building more capable AI systems.
Notably, the push towards open source is interesting. Since DeepSeek’s launch, more frontier labs and startups, including OpenAI, are adopting open source.
Murati’s startup is hiring for roles in engineering, research, and machine learning. However, unlike traditional hiring, the startup values strong personal and independent projects and welcomes both self-taught and formally trained applicants.
“Our team combines rigorous engineering with creative exploration, and we’re looking for collaborators to help shape this vision,” the startup stated in a post.
This reflects a growing trend where tech startups increasingly value personal insights and side projects in hiring, acknowledging them as strong indicators of skills and creativity.
In a way, albeit smaller, Thinking Machines Lab mirrors OpenAI’s early days – a tight-knit, elite team pushing AI forward.
The OpenAI Mafia Continues to Grow
Nearly half of the company’s 30 listed staff members previously worked at OpenAI alongside others from Mistral, Meta, and DeepMind.
“Very strong team, a large fraction of whom were directly involved with and built the ChatGPT miracle,” Andrej Karpathy, one of the founding members of OpenAI, wrote on X while congratulating Murati on her new startup.
Murati left OpenAI last October after six years at the company. “I’m stepping away because I want to create the time and space to do my own exploration,” she said at the time of leaving. Even Schulman, who joined Anthropic last year, left the company in just five months to join Murati’s team.
OpenAI alumni have already seeded multiple AI companies, and Thinking Machines Lab is part of this trend.
Moreover, Ilya Sutskever, former chief scientist at OpenAI, also announced his resignation and launched a new AI startup called Safe SuperIntelligence Inc (SSI), which raised $1 billion in funding, led by Nat Friedman and SSI co-founder Daniel Gross, within a span of three months.
Jan Leike, a former OpenAI executive, joined Anthropic following his resignation.
Perplexity AI, founded by former OpenAI researcher Aravind Srinivas, emerged after his 2018 stint at the company. Elon Musk, who left OpenAI the same year, also launched xAI in 2023. Notably, Anthropic, the company behind the Claude family of models, is led by former OpenAI employees Dario and Daniela Amodei.
Other spinouts include Aidence (generative modelling), Cresta (AI for customer service), Cleanlab (ML dataset fixes), and Symbiote AI (3D avatars).
OpenAI is becoming a launchpad for future AI founders, driving an explosion of AI entrepreneurship, much like how PayPal’s old team shaped Silicon Valley giants.
