Bilingual

Own Your Weights
14 days ago
- API dependence poses risks: model behavior, pricing, and availability can change unexpectedly.
- Open weights enable control: running models on personal hardware avoids external dependencies.
- Balance strategy: rent frontier models for cutting-edge needs, own baseline models for stability.
- Sovereignty stack includes controlling data, weights, and infrastructure to maintain autonomy.
- Establish local fallbacks for critical workflows to ensure resilience and self-reliance.
Inkling: Our Open-Weights Model
6 days ago
- The company has released a new AI model called Inkling, which is a Mixture-of-Experts transformer with 975B total parameters and 41B active parameters.
- Inkling supports a context window of up to 1 million tokens, was pretrained on 45 trillion tokens of multimodal data, and is available for fine-tuning via the Tinker platform.
- Key features of Inkling include multimodal reasoning (text, images, audio), controllable thinking effort for balancing performance and efficiency, and strong performance across reasoning, agentic coding, and safety benchmarks.
- Alongside Inkling, a preview of Inkling-Small (276B total parameters, 12B active) is shared, offering similar capabilities with lower cost and latency for certain workloads.
- The model's development involved large-scale RL training (over 30M rollouts), which improved reasoning and led to more concise reasoning styles.
- Inkling is designed as a broad foundation model for customization, emphasizing calibration, instruction following, and resistance to censorship (epistemics).
- Safety evaluations show Inkling has strong safeguards, performing well on benchmarks like FORTRESS and StrongREJECT.
- Inkling is available through various platforms and partners, including TogetherAI, Hugging Face, and others, with full weights released for customization.
American AI is locked down and proprietary. It's losing
21 hours ago
- US AI strategy relies on proprietary, locked-down models, while China adopts an open-weights approach.
- AI models lack strong moats; switching between them is easy via APIs, making enterprise services the real differentiator.
- China's open models exploit US export controls by turning compute disadvantages into distribution advantages and commoditizing the AI layer.
- The performance gap between US frontier models and open models is closing, with Chinese models claiming comparable capabilities at lower costs.
- US companies prioritize short-term profits and face misaligned incentives, risking economic fallout if AI spending declines due to ecosystem weaknesses.
- Open technologies foster permissionless innovation and ecosystem benefits, contrasting with locked-down practices that limit global adoption.

Hasty Briefsbeta

#open weights

Own Your Weights

Inkling: Our Open-Weights Model

American AI is locked down and proprietary. It's losing