Daily AI News | Google opens Googlebook preorders with five models starting at $899
# Daily AI News | Google opens Googlebook preorders with five models starting at $899
## Lead Story
### Google opens Googlebook preorders with five models starting at $899
Google opened preorders for Googlebook, a new laptop category built on the Android technology stack and ChromeOS foundations for Gemini Intelligence and Android-phone integration. Five partner devices start at $899 and are scheduled for October.
**AI take:** The platform joins Gemini, Android, and desktop computing; app compatibility, battery life, and long-term updates will decide its value.
## More News
### SpaceXAI releases Grok 4.7 for coding and knowledge work
SpaceXAI released Grok 4.7 with a 500,000-token context window and text-and-image input, available through its API, Cursor, Grok Build, and model gateways. Below 200,000 prompt tokens, list pricing is $2 per million input tokens and $6 per million output tokens; benchmark figures are vendor-reported.
**AI take:** The upgrade keeps base pricing flat while targeting longer tasks, but users should test stability and total token use on real codebases.
### Yandex open-sources the AliceAI 80B base model under Apache 2.0
Yandex released the weights for AliceAI-Foundation-80B-A3B-Base. Trained from scratch, the sparse MoE has 80 billion total parameters with 3 billion active and is available for research and commercial projects under Apache 2.0.
**AI take:** The low active-parameter count and permissive license reduce deployment barriers, though vendor performance claims need independent reproduction.
### NVIDIA launches DSX Ready qualification for AI-factory power and cooling
NVIDIA launched DSX Ready to qualify partner products against applicable DSX AI-factory reference-design requirements. Initial categories cover battery energy storage systems and cooling distribution units from partners including Hitachi Energy, LG Energy Solution, Tesla, LiquidStack, and Vertiv.
**AI take:** A common qualification can reduce integration risk, but it does not replace site-specific engineering and stability testing.
### Huawei Cloud adds a HarmonyOS coding model and agent to CodeArts
Huawei Cloud upgraded CodeArts for HarmonyOS developers with a dedicated coding model, a HarmonyOS agent, and a developer practice center intended to cover the path from idea and development to app-store submission.
**AI take:** A specialized model can reduce framework knowledge gaps; code correctness and actual publishing speed will determine its value.
### Zhipu apologizes and opens ZCode after a repository-data dispute
Zhipu apologized after ZCode was accused of uploading local Git history without adequate disclosure, saying it deleted the related object-storage data and fixed the issue in version 3.14.0. It also opened the client code and introduced third-party security review.
**AI take:** Open code and audits improve verifiability, but organizations should still examine data boundaries, defaults, and deletion evidence.
### British Columbia sues OpenAI over an alleged role in a school shooting
British Columbia sued OpenAI and Sam Altman in a U.S. court, alleging ChatGPT contributed to a mass school shooting and that the company failed to warn authorities. The allegations have not been adjudicated.
**AI take:** The case will test warning duties and liability boundaries for general-purpose chat products in high-risk conversations.
### Runway Workflows adds compositing, HDR, and depth tools
Runway expanded Workflows with compositing, alpha, HDR, depth-map, and RGB-depth operations, letting creators connect more generation and post-production steps inside one product.
**AI take:** This can reduce tool switching; format compatibility and repeatability on complex projects will determine the real gain.
### OpenAI proposes international AI-safety standards and coordination
OpenAI published proposals for international AI-safety standards, calling for greater use of national AI-safety institutes for testing, standards coordination, and communication about high-risk incidents. The paper is an industry policy proposal, not a binding rule.
**AI take:** Cross-border alignment could reduce fragmented compliance, but impact depends on government adoption, independent testing, and incident disclosure.
## Rumors
