ChatT2: New LLM Framework Targets Natural Product Research with Multi-Agent Architecture
Researchers introduce ChatT2, a specialized large language model agent designed to assist with bacterial natural product investigations.
Sources · arxiv.org
The news desk
Follow what’s changing.
Recent developments in artificial intelligence, with the original sources one click away.
All news
181–200 of 1,892 Claude Opus 5.5 Now Available on Amazon Bedrock and AWS
Anthropic's Claude Opus 5.5 model is now accessible through Amazon Bedrock and Claude Platform on AWS for agentic coding and knowledge work tasks.
Sources · Amazon AWS AI
MIT Technology Review Warns Against Summer AI Hype Cycle
Recent AI claims from major companies highlight growing concerns about industry hype, according to MIT Technology Review.
Sources · MIT Technology Review
New Research Tackles Credit Assignment and Multi-Agent Reasoning Challenges in LLM Training
Researchers introduce PACT and VERPO methods for improved LLM training while revealing vulnerabilities in multi-agent reasoning systems.
Sources · arxiv.org
CESBench Introduces New Benchmark for Testing LLMs on IoT Cryptographic Security
Researchers create 380-item benchmark evaluating large language models on cryptographic engineering security for IoT devices.
Sources · arxiv.org
MIT Technology Review Investigation Reveals Deaths Near US Border's AI Surveillance Towers
Investigation documents cases of migrants who died after passing undetected through areas monitored by AI-enabled surveillance towers along the US-Mexico border.
Sources · MIT Technology Review
OpenAI Forms Mathematics Advisory Group as Internal Model Resolves Over 100 Open Problems
OpenAI announces independent advisory group at IAS as new AI model solves 100+ mathematics problems, drawing mixed response from mathematicians.
Sources · openai.com / techcrunch.com
Researchers Develop AI Chatbot to Decode Damaged Ancient Greek Papyrus
A new large language model aims to fill gaps in fragmented ancient Greek papyrus records to reveal details about ancient life.
Sources · WIRED AI
Three ArXiv Papers Address LLM Memory, Security, and Educational Applications
New research tackles conversational memory systems, encrypted guardrails against jailbreaks, and AI tutoring for programming students.
Sources · arxiv.org
New CESBench Benchmark Tests LLMs on Cryptographic Security for IoT Devices
Researchers introduce CESBench with 380 expert-written items to evaluate LLMs on cryptographic engineering security, revealing gaps in security justification.
Sources · arxiv.org
NVIDIA Showcases Five Companies Using AI for Clean Energy at New York Climate Week
NVIDIA highlights companies applying AI to accelerate clean energy adoption by addressing infrastructure, R&D, and cost challenges.
Sources · NVIDIA AI Blog
TechCrunch's Equity Debates AI Industry's Commitment to Slowing Down
TechCrunch's Equity podcast examined whether AI executives genuinely want to slow the pace of AI development.
Sources · TechCrunch AI
Three New AI Research Papers Address Memory, Security, and Education Applications
Researchers released papers on conversational memory systems, encrypted AI guardrails, and LLM-powered programming tutors.
Sources · arxiv.org
US and China Discuss AI Security Alert Mechanism
Officials from the US and China discussed establishing a notification system for AI incidents that could threaten national security.
Sources · WIRED AI
AI Chatbots Driving Surge in Security Vulnerability Discoveries, WIRED Reports
While AI labs consider slowing development, widely available AI chatbots are already helping security researchers uncover numerous software flaws.
Sources · WIRED AI
Google's Gemini AI Model Demonstrates Ability to Hack Other Companies' Systems
Google's Gemini AI model has been shown to hack other companies' systems, though Google states it ended each hack immediately.
Sources · TechCrunch AI
Meta's Muse App Raises Privacy Concerns Over Data Collection and Information Sharing Prompts
Meta's new Muse app automatically opts users into AI training data collection and requests sensitive personal information including bank details.
Sources · WIRED AI
NYT Columnist Examines Anthropic Leader's Warnings on AI Risks
Kevin Roose reflects on AI dangers highlighted by Anthropic's Amodei as public awareness grows around risks long discussed in Silicon Valley.
Sources · NYT Technology
Trump's Data Center Push Faces Opposition from MAGA Base
President Trump's support for AI data centers is creating friction with his core supporters who are mobilizing against the facilities.
Sources · WIRED AI
Anthropic CEO Highlights Disturbing Findings in AI Safety Research
Anthropic's CEO emphasizes that AI safety depends on understanding how systems think, but current research reveals concerning patterns.
Sources · WIRED AI
