Hugging Face Launches Open Alignment Initiative, Backs Anthropic's Embedded Evaluators
Hugging Face announces the Open Alignment Initiative and asks to join Anthropic's embedded evaluators program to make AI safer through transparency.
Global AI Digest
-
-
OpenCode Go Reaches Break-Even for Two Consecutive Weeks
Developer dax reports that OpenCode Go has broken even for two straight weeks, calling it a major milestone and arguing that rivals burning cash or quietly sell…
-
Anthropic Pledges Permanent Third-Party Access to Its AI Systems as CEO Calls for Industry to Slow Down
Anthropic CEO Dario Amodei published a new essay arguing the AI industry should slow down and outlined a three-part plan. Anthropic is immediately committing to…
-
Command Code CEO Calls DeepSeek V4.1 Flash the 'GOAT Model' After Record-Breaking Launch
Command Code CEO Ahmad Awais praised DeepSeek V4.1 Flash as a 'GOAT model' after the model's record-breaking launch, with the platform reporting 3.1 trillion pa…
-
DeepSeek V4.1 Debuts as Command Code's Top Model, Serving 3.1 Trillion Paid Tokens in 24 Hours
Command Code announces DeepSeek V4.1 as its new top model, claiming record-breaking usage in the first 24 hours with 3.1 trillion paid tokens and sustained thro…
-
Command Code: DeepSeek V4.1 Flash Sets Record With 3.1 Trillion Paid Tokens in 24 Hours
Command Code reports that DeepSeek V4.1 Flash is now its most-used model, claiming 3.1 trillion paid tokens served in the first 24 hours and sustained throughpu…
-
OpenCode Extends 4x Usage Boost for Another Week
OpenCode, the open source coding agent, is extending its boosted usage plan for another week, continuing day-0 support after raising limits to 4x for 72 hours.
-
Qoder Advances Agentic Coding with Custom Models, Kimi K2.8 Preview, and Canvas Mobile Use
Qoder, an agentic coding platform, has rolled out a series of updates: custom model integration for all CLI users, Kimi K2.8 Preview support, and a new Canvas m…
-
OpenAI Astra Resets After Fixing Quality Issues
OpenAI resets Astra for users after fixing skill over-triggering and disabling an opt-in context experiment that affected an estimated 4-5k users.
-
Qwen3.8-27B Now Runs on Cerebras with Rapid Inference
Alibaba Qwen's Qwen3.8-27B dense open-weight model is now available on Cerebras for fast inference, scoring 34 on the Artificial Analysis Intelligence Index and…
