Skip to content
AI News HubLIVE

This edition’s highlights

Agents

How Google Duplex evolved to also call businesses for you in Gemini

  • Duplex moved from custom recurrent networks limited to narrow domains to a unified Gemini stack spanning multimodal understanding, planning, and dialogue.
  • Live calls still pose pragmatic grounding and sub-second latency challenges in noisy acoustic settings that web pre-training alone cannot solve.
Google Research BlogIn-site articleHow Google Duplex evolved to also call businesses for you in Gemini

New in AWS Quick: More capable flows, an updated web experience, and agents that work together

  • Users can describe a business process in chat and Quick builds a flow that orchestrates agents, skills, tools, and data, with deterministic steps for consistency.
  • The updated web app surfaces what needs attention, adds a daily briefing, and keeps agents working after the browser tab is closed.
AWS Machine Learning BlogIn-site articleNew in AWS Quick: More capable flows, an updated web experience, and agents that work together

Cut self-hosted LLM cost and latency with the distributed KV caching agent skill

  • Distributed KV caching adds a shared tier beside GPU fleets to reuse KV state, reducing cost up to 40% and TTFT up to 80%.
  • The Agent Skill guides fit checks, KVTrace savings estimates, cache cluster setup, and troubleshooting.
AWS Machine Learning BlogIn-site articleCut self-hosted LLM cost and latency with the distributed KV caching agent skill

Security event: Prompt leakage for requests in June and July of 2026

  • The leak occurred when an AI coding agent committed a credential to a public GitHub repository, providing access to a private Hugging Face repository.
  • Affected prompts came from requests to the Minimax M3, GLM 5.2, and Kimi 2.6 models, primarily on June 24, July 4, July 8, and July 14, 2026.
Together AI BlogIn-site articleSecurity event: Prompt leakage for requests in June and July of 2026

Improved controls for the new era of cyberattacks

  • Threat actors are shifting from AI-assisted to fully autonomous AI-led attacks, with agent swarms exploiting zero-days and exfiltrating data in minutes.
  • Baseten uses AI-assisted defense including defense-in-depth, second-level detection and response, and agentic red teaming to match attack speed.
Baseten BlogIn-site articleImproved controls for the new era of cyberattacks
Models

Qwen4 Max is live on Together AI

  • Qwen4 Max handles 1M-token context, processing all 344 files of the Tornado repository (652,519 tokens) in one request
  • The model's patch passed the maintainers' tests, costing just 69 cents
Together AI BlogIn-site articleQwen4 Max is live on Together AI

Meet bison-1 and groundhog-1, frontier American decision models

  • bison-1 ranks #1 of 19 decision models on Artificial Analysis Decision Index v2.1 with a score of 76, ahead of d1.2-ultra (75) and mammoth-1 (71).
  • Both models are built on Google Gemma 4 and offer 128k context windows (8x previous models) with text and image input.
Baseten BlogIn-site articleMeet bison-1 and groundhog-1, frontier American decision models

How Standard Bank Group automates AI risk assessment for generative AI

  • The assessment runs entirely in Standard Bank Group’s AWS account using Amazon S3, AWS Well-Architected Tool, AWS Step Functions, and Amazon Bedrock.
  • A pre-assessment agent collects evidence automatically, and the pipeline tests the application directly, including jailbreak attempts and consistency checks.
AWS Machine Learning BlogIn-site articleHow Standard Bank Group automates AI risk assessment for generative AI
Research

Ravi Vakil on Alessandro Masullo's Proposed GRH Proof: 'It Is All Human, with No AI'

  • Alessandro Masullo has a proposed proof of the Generalized Riemann Hypothesis, which Brian Conrad has been reviewing closely for months and giving extensive feedback on.
  • Vakil stressed that the work is entirely human, with no AI involved.
Simon Willison's WeblogIn-site articleRavi Vakil on Alessandro Masullo's Proposed GRH Proof: 'It Is All Human, with No AI'

Levent Alpöge Claims Riemann Hypothesis Proof Discovered with Claude

  • Alpöge says he and Claude proved the Riemann Hypothesis and the Generalized Riemann Hypothesis
  • He claims the arguments were formalized without assumptions
Simon Willison's WeblogIn-site articleLevent Alpöge Claims Riemann Hypothesis Proof Discovered with Claude

You have reached the end of this edition.

Your reading list
Other updates (2)
Chips

Introducing Amazon SageMaker HyperPod Global Inference Gateway: One endpoint across Regions and clusters

AWS launches the SageMaker HyperPod Global Inference Gateway, a managed endpoint that routes inference requests across multiple clusters and Regions. It scores clusters based on KV cache utilization, queue depth, Region proximity, and error rate, handles failover automatically, and authenticates via SigV4 and IAM over AWS PrivateLink.

AWS Machine Learning BlogIn-site articleIntroducing Amazon SageMaker HyperPod Global Inference Gateway: One endpoint across Regions and clusters
Agents

Announcing Amazon Textract Insights: Verifiable generative AI for documents

AWS announced the general availability of Amazon Textract Insights, a generative AI feature that lets users describe fields in natural language and submit a single document or mixed packet to get structured results with confidence scores and page-level grounding, with no model training required.

AWS Machine Learning BlogIn-site articleAnnouncing Amazon Textract Insights: Verifiable generative AI for documents