Episode Details

The open-source AI revolution is shattering the boundaries of enterprise customization and mobile hardware limits. Mira Murati’s highly anticipated startup, Thinking Machines Lab, has officially entered the arena with "Inkling", a staggering 975-billion parameter Mixture-of-Experts (MoE) model built to compete directly against tech giants through bespoke business fine-tuning.

We’ll talk about:

• Mira Murati's startup launches a 975B parameter MoE architecture that only awakens 41B active parameters per task, featuring an adjustable "thinking effort" dial and native uncertainty flagging.
• How Thinking Machines plans to monetize by using Inkling as a baseline for enterprises to build hyper-customized, private-data workflows using their training hub.
• Bonsai 27B utilizing aggressive 1-bit quantization to shrink a 27B model down to an unprecedented 3.9GB, running local coding and vision workflows entirely on-device without cloud costs.
• OpenAI launching "Codex Micro," a dedicated physical keyboard featuring built-in workflow dials, while Apple reportedly taps Alibaba's Qwen to finally power Apple Intelligence in China.

Keywords: Thinking Machines Lab, Inkling, PrismML Bonsai 27B, phone local LLM, Codex Micro, Tinker.

Links:

1. Newsletter: Sign up for our FREE daily newsletter.
2. Our Community: Get 3-level AI tutorials across industries.
3. Join AI Fire Academy: 700+ advanced AI workflows ($14,500+ Value)

Our Socials:

1. Facebook Group: Join 295K+ AI builders
2. X (Twitter): Follow us for daily AI drops
3. YouTube: Watch AI walkthroughs & tutorials