While transaction fraud attempts by AI startups have decreased, multi-account fraud such as free trial abuse has surged.
Voodoo Dynamic Quant, a tool that selects the optimal quantization level for each tensor using gradient descent, has been released under the MIT license.
The open-source MCP app 'Chat On Steroids' has been released, enabling local file modifications and terminal command execution via ChatGPT conversations.
Toss Securities shared its experience of overcoming the limitations of GPU-awareness by building a GPU-native cluster equipped with validation and resource control capabilities.
Samsung Electronics has released an AI framework that analyzes lyrics and musical structure together to evaluate songs at a human-expert level.
Meta has released a methodology that applies causal inference to quantify the incremental revenue of AI infrastructure investments, which could not be measured by A/B testing.
Pinterest reduced serving costs by up to 30% by introducing SSD serving and quantization technologies to the Manas platform.
Anthropic released the Salesforce in Claude plugin in beta, enabling sellers to automate CRM tasks and pipeline reviews within Claude.
Anthropic released 43 new workflows and 27 integrations for Claude for Small Business, featuring partners like Shopify and Salesforce, alongside a fall workshop series.
Anthropic has released a guide for sales organizations to deploy Claude, featuring a three-phase rollout plan and case studies from Cox Communications and Cyera.
Google has released 'Agent Substrate', an open-source runtime that enables AI agents to run at 10x higher density on GKE and restart in under 500ms.
Manticore Search has introduced chunking strategies within table definitions to address information loss caused by model input window limits.
MIT CSAIL spinout G5 Labs has secured $14 million in seed funding and unveiled a platform that converts natural language into structured intent graphs to generate code.
A recursive framework has been proposed in which AI agents autonomously generate and execute scientific instruments to build an executable world for inferring the mechanisms of physical phenomena.
A developer tested a system using the x402 protocol that requires AI agents to pay 1 cent before reading a page.
Periodic Neon has been deployed in laboratories, surpassing GPT-6 and Claude 5.1 in scientific analysis at a lower cost.
Delphi shortened its Python backend deployment cycle to over 100 times a day by leveraging Vercel Workflows and Queues.
Artificial Analysis released Capability Indices v1.1, with Claude Fable 5.1 taking first place in all six occupational indices and Grok 4.5 demonstrating high cost efficiency.
Researchers have released 'Turnbench', a turn-taking evaluation benchmark by dialogue type featuring 30 hours of high-quality data and a standardized evaluation protocol.
Andon Labs has unveiled Pion, a platform designed to validate AI's ability to autonomously acquire resources through real-world business operations such as vending machines and retail stores.
OpenAI researcher Daniel Selsam warned of the risk that AI models may feign alignment and slip into uncontrollable states.
AI Briefing