Skip to content

Sam's News — ai — 2026-08-15

AI

7 Apple develops custom AI model for China market with Alibaba partnership

Apple is developing its own AI model for the Chinese market in partnership with tech company Alibaba.

  • Custom LLM developed with Alibaba for Chinese market
  • Apple Intelligence rollout planned following iOS update
  • Regulatory registration with Chinese cyberspace authority completed
  • Rare US–China partnership amid geopolitical tensions

Sources: Ratopati RSS, Reuters Web Search Update to: Apple Partners With Alibaba to Train Custom AI Model for China Market

7 OpenAI Voluntarily Limits New AI Model Releases at Government Request

OpenAI agreed to work with the government on a formal process for reviewing and limiting new model releases.

Sources: Cybersecurity Dive Web Search, Cybersecurity Dive Web Search Update to: OpenAI voluntarily limits new model releases at government request

7 Google Releases Three New Gemini AI Models

Google released three new Gemini AI models, including Gemini 3.6 Flash as its most powerful variant.

Sources: The New York Times Web Search Update to: Google releases three new Gemini AI models including Gemini 3.6 Flash

6.5 Meta Releases Muse Glimmer 30B Open-Weight Model for Agentic AI

Meta released Muse Glimmer, a 30B open-weight dense language model with a 120K+ context window for local agentic AI workflows.

Sources: NVIDIA Developer Web Search, NVIDIA Developer Web Search Update to: Meta releases Muse Glimmer 30B open-weight language model

6.5 Price Becomes the Most Important LLM Benchmark

Cost efficiency has become the most important benchmark for evaluating large language models, as demonstrated by Google's Gemini releases.

Sources: Constellation Research Web Search, The New York Times Web Search, Constellation Research Web Search Update to: Price Emerges as Most Important LLM Benchmark

6 Z.ai releases GLM-5.3 with enhanced coding and cybersecurity capabilities

Chinese AI developer Zhipu AI released GLM-5.3, an open-source large language model with record performance across benchmarks.

Sources: SiliconANGLE Web Search, SiliconANGLE Web Search

5.5 Running Large Language Models Locally on Android Phones

A user successfully deployed a local large language model directly on an Android phone, eliminating the need for cloud-based AI subscriptions.

Sources: Android Police Web Search, Android Police Web Search Update to: Running local LLMs on Android phones eliminates need for cloud AI subscriptions

5.5 Clinical Expertise and Model Confidence Key to Radiologist-LLM Collaboration

Radiologists' ability to avoid being fooled by LLMs depends on their confidence in the AI model and their own clinical expertise.

Sources: Radiology Business Web Search Update to: Radiologists' Resistance to LLM Misidentification

5.5 Evaluation harness finds AI models overconfident when wrong

Automated evaluation testing reveals that large language models exhibit highest confidence precisely when making errors, a flaw qualitative review missed.

Sources: VentureBeat RSS

QNAP has released Qsirch, which integrates local vision and language models to enable semantic search across media stored on network-attached storage devices.

Sources: Letem světem Applem RSS