Alibaba Releases Qwen 3.8 Models with Open Weights
Alibaba's Qwen team released Qwen 3.8 models with open weights under Apache 2.0 license on August 14, 2026.
Archive
Browse all published articles.
ARCHIVE
-
-
Connecticut Judge Blocks AI Prompt Injection in Court Filings
A Connecticut judge ruled that hidden AI prompts in court filings are a 'dangerous' tactic, citing a case where a plaintiff attempted to influence a court's decision using AI-generated text.
-
Google Lets Users Remove Visible Watermark From AI Generations
Google announced on Friday it will let users remove visible watermarks from AI-generated content, including images, videos, and songs.
-
Grok 4.6 Now Available in GitHub Copilot
GitHub Copilot has integrated Grok 4.6, allowing developers to access the latest coding model starting August 14, 2026.
-
Indonesia Launches First University-Based AI Tech Center with NVIDIA
Indonesia's Ministry of Communication and Digital Affairs, Indosat, NVIDIA, and UGM launched the UGM Indosat NVIDIA AI Technology Center, the country's first university-based AI center, to address national challenges through AI innovation.
-
Tim O'Reilly Advocates Open-Source AI to Empower Users
Tech visionary Tim O'Reilly warns that major AI labs are misreading the future, prioritizing large models over user needs.
-
OpenAI and Anthropic slash prices in AI model price war
OpenAI and Anthropic are cutting prices for their AI models as companies seek cheaper alternatives, with costs for mid-tier products dropping by almost a quarter since mid-July.
-
Hyperscalers May Regret Embracing Natural Gas for AI Data Centers
Natural gas prices could triple in some U.S. regions by 2027, according to Noreva, as hyperscalers expand their reliance on the fossil fuel for AI data centers.
-
Kog Targets LLM Inference Speed with GPU Optimization
French startup Kog claims its software can deliver 3,000 tokens per second on standard GPUs, aiming to unlock faster AI inference for enterprises.
-
Meta Releases Open-Weight AI Model Glimmer
Meta launched Glimmer, an open-weight AI model, this week, allowing users to run it on their own hardware.
-
AWS Combines SageMaker AI and Bedrock AgentCore for Multi-Agent Workflows
AWS demonstrates how to integrate SageMaker AI with Bedrock AgentCore to build multi-agent systems, using Qwen 3.5 9B and Claude models.
-
Amazon Nova Forge Enables Custom Reward Functions for Multi-Turn Reinforcement Learning
Amazon Nova Forge now supports multi-turn reinforcement learning with a serverless option, allowing teams to design reward functions for complex tasks.