Tinker, the first product from Thinking Machines Lab—founded by ex-OpenAI leaders—streamlines AI model development by automating complex training tasks, offering a user-friendly interface that retains full control over data and algorithms, and earning early praise for its accessibility and power
Mira Murati, former Chief Technology Officer at OpenAI, has re-entered the artificial intelligence scene with an innovative new venture. Her stealth startup, Thinking Machines Lab, recently unveiled its debut product, Tinker, designed to make building custom AI models significantly easier for a wide range of users—from researchers and developers to hobbyists.
Democratizing AI development
Traditionally, creating or fine-tuning large AI models demands extensive technical knowledge and access to costly infrastructure. The process typically involves managing clusters of GPUs, navigating complex software ecosystems, and closely monitoring training to avoid system failures. Tinker aims to change that by automating many of these technical challenges, allowing users to focus on designing and improving their models without the usual headaches.
Murati emphasized the startup’s goal: “We want to make frontier AI capabilities accessible to everyone, not just elite tech companies or academic labs.” The tool represents a push towards democratizing AI, enabling broader experimentation and innovation.
Built by AI veterans, praised by early users
Thinking Machines Lab’s team includes several former OpenAI researchers who contributed to the development of ChatGPT. Early testers have praised Tinker for being powerful yet much more user-friendly than comparable tools. John Schulman, co-founder of the startup and another OpenAI alumnus, explained, “While we handle the complex distributed training behind the scenes, users retain full control over their data and training algorithms. It’s a mix of automation and flexibility.”
By lowering the barriers to AI customization, Tinker opens possibilities for creating specialized models tailored to niche applications—whether for solving scientific problems, automating legal writing, or supporting medical inquiries. Murati summed up the vision: “We’re unlocking frontier AI research for everyone, and that could truly transform how these technologies evolve.”
