Mira Murati's AI Model Inkling is Open-Source ... but Huge

Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has released Inkling—a massive open-weights AI model with 975 billion parameters.

Although Inkling is free to download under the Apache 2.0 license, running it is another story. Even the compressed version reportedly requires around 600GB of GPU memory, while the full-quality model may need more than 2TB of VRAM and a cluster of high-end NVIDIA data-center GPUs.

In this video, we examine the difference between open-source and open-weights AI, Inkling’s sparse architecture, the hardware required to operate it, and the business strategy behind Thinking Machines Lab’s commercial Tinker platform.

Is Inkling really an attempt to democratize artificial intelligence—or is it a carefully designed strategy to attract enterprise customers into a highly profitable infrastructure ecosystem?

Watch the full video to understand Mira Murati’s multi-billion-dollar AI gamble and what it could mean for the future of enterprise AI.

#MiraMurati #InklingAI #ThinkingMachinesLab #ArtificialIntelligence #OpenSourceAI #OpenWeights #AIModels #EnterpriseAI #GenerativeAI #MachineLearning #NVIDIA #OpenAI #AITechnology #TechNews #FutureOfAI