A report from 27 workplaces finds wide differences in AI costs and higher PR output in some datasets, but few companies can quantify returns.

FedScoop reports HHS plans to award Anthropic a contract without competition before a wider AI contest; key terms and the agency’s rationale are unknown.

ACE Robotics has released a model for robot spatial perception, but its name, capabilities, performance and availability have not been disclosed in the available report.

FairStress tests whether demographic cues affect LLM decisions and how readily models change them when challenged. Full model-level results remain unavailable.

OpenAI says a franchisee used ChatGPT Work to reconcile cash in 94% less time, but the available account lacks measurement and accuracy details.

AnyLanguageModel 1.0 gives Swift apps a shared interface for local and hosted models, alongside 1.0 releases of two Hugging Face download libraries.

OpenAI has published an article titled “Advancing coding in low-resource programming languages,” but available details do not confirm its methods or results.

A 36Kr headline describes 36 former ByteDance employees who became entrepreneurs, but the article text and supporting details are unavailable.

An amNewYork headline says Grok generated a sexualized image of Julie Menin in under 30 seconds. The article’s details and company response are unavailable.