Thinking Machines Lab

7 stories

Stories

A Safe Path to Open Weights
Strategic openness can strengthen AI safety and support broader access as defenses and safety science mature.
Tinker BlogJul 31, 202610 min read
The Future Worth Building Is Human
AI built for autonomy crowds people out, making us passive observers of what’s coming. We’re building toward a different future.
Tinker BlogJul 10, 202610 min read
Interaction Models: A Scalable Approach to Human-AI Collaboration
Interaction models move beyond turn-based AI interfaces by handling multimodal, real-time collaboration natively across audio, video, and text.
Tinker BlogMay 11, 202616 min read
On-Policy Distillation
On-policy, dense supervision is a useful tool for distillation
Tinker BlogOct 27, 202525 min read
LoRA Without Regret
How LoRA matches full training performance more broadly than expected.
Tinker BlogSep 29, 202526 min read
Modular Manifolds
A geometric framework for co-designing neural net optimizers with manifold constraints.
Tinker BlogSep 26, 202518 min read
Defeating Nondeterminism in LLM Inference
Reproducibility is a bedrock of scientific progress. However, it’s remarkably difficult to get reproducible results out of large language models.
Tinker BlogSep 10, 202521 min read