Thinking Machines Engineer Retrospective on Open-Source Model Development
jonlachman · x · 2026-07-16
A Thinking Machine team member reflected on the entire process of launching the open-source model Inkling. They emphasized that building the entire pipeline from scratch—spanning data, pre-training, post-training, and final release—is an incredibly complex engineering feat ("taking a village"), especially the release of open weights. While there were many potential pitfalls and areas for future improvement, they expressed satisfaction with the current results and readiness for the next challenge.
More from Models
- DeepSeek cut KV cache per token from 389KB to 890B in two years as open-closed gap debate rages — teortaxesTex · 2026-10-27
- GPT-6 Astra controls dexterous hands zero-shot at just 5.2% success in 500 trials — stingning · 2026-10-27
- Unverified: fable 5.5 speculated at 4-5T parameters to outscale 1-3T Opus — haider1 · 2026-10-27
- Kimi K3 Is 'Open' Until You Succeed: License Caps Free Use at $20M Revenue — alex_verem · 2026-10-27
- Screenshot shows 'GPT-6 Astra Ultra' in ChatGPT model picker — unverified leak or joke — justalexoki · 2026-10-27
- DeepSeek Safety Eval Finds "Over-Attribution": Model Mistakes Ordinary Tasks for Tests — Much_Boysenberry9124 · 2026-10-27