Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
This article details the approach used to fine-tune a 350 million parameter model to generate better structured outputs efficiently.
Latest AI News & Updates — Page 35
TechCrunch Disrupt 2026’s new Real World AI Stage features Nvidia, robots, and extinct animals
The Real World AI stage at TechCrunch Disrupt 2026 will showcase the convergence of digital and physical technology, including Nvidia and robots.
Accessing OpenAI models on Amazon Bedrock from Australia with global cross-Region inference
Amazon Bedrock now offers Australian teams access to OpenAI GPT-5.6 Sol, Terra, and Luna models with global cross-Region inference from the Sydney and Melbourne Regions.
OpenAI’s new reasoning technique alarms AI safety experts
OpenAI's Astra model introduces "recurrent depth," a reasoning technique that departs from sequential thinking, prompting concern from AI safety experts.
Google says its new Gemini 3.8 Flash model ‘works harder’ but might cost more
Google released Gemini 3.8 Flash weeks after 3.7 Flash, claiming it 'works harder' with more reasoning and iterative tool calls, though higher token use could raise costs.
Modernizing and scaling support operations with generative AI on AWS
AWS describes building a generative AI support operations platform that converts training videos into structured SOPs, applies RAG to guide ticket resolution, and predicts SLA risk to prioritize work.
From code to diagrams: Agentic architecture documentation with Amazon Bedrock AgentCore
AWS describes how a global interdealer broker used Amazon Bedrock AgentCore to automate architecture documentation, analyzing .NET code bases and generating searchable diagrams.
Trinity: Agentic AI-powered transition planning for students with disabilities
University Startups and AWS partner g/d/n/a built Trinity, a serverless multi-agent AI on Amazon Bedrock that creates IDEA-aligned transition plans for students with disabilities.
Proactive cyber defense for governments and enterprises
Google has launched the Fairwind Program, providing cyber defense tools to governments and trusted partners seeking proactive security measures.
ATV Big Air Tour turned 3 days of work into 3 hours with ChatGPT
ATV Big Air Tour reports using ChatGPT Work to speed up marketing and merchandising, including converting merchandise photos into an inventory website in 15 minutes.
Anthropic launches Claude Fable 5.1 and says it’s up to 45 percent cheaper for agentic work
Anthropic released Fable 5.1 and Mythos 5.1, saying the new models cost less—up to 45 percent cheaper for complex agentic tasks—while improving performance and addressing customer concerns.
BenchMIRT: What are LLM benchmarks actually measuring?
The article discusses the metrics used in LLM benchmarks and what they truly assess in language models.