Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Researchers at Cato AI Labs discovered two critical sandbox escape vulnerabilities in Cursor's AI code editor, tracked as CVE-2026-50548 and CVE-2026-50549, both rated 9.8/10 on CVSS 3.1. The flaws, collectively named Du…
Shanghai AI Lab has developed Agents-A1, a 35-billion-parameter model that matches trillion-parameter models on multi-step agent tasks. Instead of scaling parameter count, the researchers scaled the training horizon—the …
Anthropic added a new security measure to its Claude Fable 5 AI model, agreeing to extend guardrails for certain restricted capabilities, after the Trump administration lifted export controls that had taken the model off…
OpenAI now allows ChatGPT Plus subscribers to use their $20/month plan inside third-party coding agents like Pi and OpenCode, effectively providing $100–$200 worth of tokens. However, the policy is not contractual and co…
Shopify has joined the PyTorch Foundation as a Platinum member, committing to invest in the open-source AI ecosystem and contribute engineering expertise to shape PyTorch for commerce and agentic AI. The e-commerce platf…
Physicists Giorgio Parisi and Francesco Zamponi used the AI model Claude to help solve the jamming problem, a decades-old math mystery in physics. Claude's initial proof contained errors but provided a correct underlying…
Researchers at the University of Louisiana at Lafayette are using artificial intelligence to help preserve Louisiana French, an endangered language with only about 20,000 fluent speakers remaining. The LaFLEUR Project ai…
A developer introduced Graph of Thoughts (GoT), a reasoning framework that extends Tree of Thoughts by allowing branches to merge. GoT represents reasoning as a graph where thoughts can have multiple parents, enabling ag…
Anthropic's Claude Code was found embedding invisible tracking signals in Unicode apostrophe variations and date format changes within its system prompts. The tool, which has shell access to users' machines, used these c…
A developer built Relevyn, a tool that measures brand visibility across AI engines like ChatGPT, Claude, Perplexity, and Gemini. The tool uses deterministic text detection instead of LLM-based analysis for speed and cons…
Together AI raised $800 million in Series C funding at an $8.3 billion valuation, led by Aramco Ventures with participation from Nvidia and others. The company, which provides infrastructure for running open-source AI mo…
OpenAI announced GPT-5.6 Sol, Terra, and Luna on June 26, but access is limited to a small group of trusted partners via Codex and the API at the request of the U.S. government. The model family features aggressive prici…
GitHub Copilot CLI now features auto model selection that routes tasks to the optimal AI model based on real-time utilization and health metrics, improving reliability and token efficiency. The feature respects admin pol…
SWIRL 5 introduces a source-grounded hallucination check that runs after generation to catch confident wrong answers with misleading citations. The system segments claims, verifies them against source documents, and flag…
Google researchers have developed the Scalable Cluster Termination System (S-CTS), a multimodal defense that uses LLMs enhanced with LoRA and APO to detect and terminate coordinated bot-nets producing AI-generated spam o…
AI startups should shift from selling productivity-enhancing software to selling completed work products, opening new verticals and pricing models. EvenUp exemplifies this by selling AI-generated demand packages to perso…
The Trump administration lifted export controls on Anthropic's AI models, enabling the company to restore access to Claude Fable 5 and Mythos 5. The decision marks a shift in US AI policy and benefits Anthropic's global …
Anthropic suspended access to its new Fable 5 and Mythos 5 models three days after launch on June 12, 2026, citing unspecified issues, and retired Claude Sonnet 4 and Claude Opus 4 on June 15, breaking any requests still…
Rising token costs are driving companies to abandon tokenmaxxing strategies in favor of open weight models, which offer greater efficiency and lower computational expenses. Alex Karp noted the shift as businesses priorit…
The U.S. Department of Commerce lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 models, reversing restrictions imposed in June over national-security concerns. Commerce Secretary Howard Lutnick informed…