Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Australian startup Springboards has launched Flint, an LLM trained to generate more diverse responses to open-ended questions, addressing the predictability and lack of creativity in mainstream chatbots like ChatGPT and …
The Center for AI Safety, with world experts, created Humanity's Last Exam (HLE), a benchmark of over 2,500 expert-level questions across disciplines to test AI reasoning. Even top models like GPT, Gemini, and Claude sco…
Codex and Claude Code are competing AI-powered developer tools from OpenAI and Anthropic, respectively, each using state-of-the-art models to read and modify entire codebases. The article compares the two harnesses, noti…
Snowflake launches Cortex AI Guardrails, a runtime security layer in Horizon Catalog that blocks indirect prompt injection attacks targeting enterprise AI agents. The feature intercepts and neutralizes adversarial instru…
Microsoft 365 Copilot, launched in November 2023, has evolved from a simple chatbot into a suite of autonomous agents integrated into Office apps, reaching 20 million paid users by April 2026. However, Microsoft is now l…
AI-generated SQL queries are increasingly deployed without human understanding, creating a risk of undetected semantic and performance bugs. The declarative nature of SQL hides execution costs and subtle errors that test…
GitHub Copilot has added Kimi K2.7 Code, an open-weight model from Chinese AI startup Moonshot AI, as a selectable option in its model picker for the first time. Hosted on Microsoft Azure and billed by token usage, the m…
The US government ordered Anthropic to suspend its Claude Fable 5 and Mythos 5 models due to jailbreak vulnerabilities reported by Amazon, then allowed their swift restoration after fixes. This intervention signals a new…
AI-powered tools like GitHub Copilot, Claude Code, and Cursor AI are increasingly used by .NET developers to generate C# code, but experts warn that generated code may contain logic errors, bugs, or security vulnerabilit…
Organizations rushing to deploy AI in production are finding that traditional software monitoring tools fail to address AI-specific failures like drift, hallucination, and degradation. Current AI observability tools rely…
Snowflake SVP of Engineering Vivek Raghunathan presented a five-stage framework at Snowflake Summit for transitioning AI-assisted engineering from chaotic experimentation to a repeatable, org-wide system.
Loop engineering, not prompt engineering, is the next frontier for AI systems that perform multi-step tasks. While prompt engineering optimizes the initial query, loop engineering governs what happens after the model res…
Large language models and autonomous AI systems are being weaponized to target AI infrastructure, creating a new attack surface that existing cybersecurity frameworks cannot address, according to a report from Small Wars…
People Inc. blocks all AI scrapers by default and grants access only to approved partners with licensing deals, as Chief Innovation Officer Jonathan Roberts argues that trust in AI is shifting to individual creators rath…
Developer Yuri Koval recommends managing Claude Code configuration with a dotfiles repository and GNU Stow, enabling version control rollbacks and easy machine migration. The approach tracks skills, hooks, plugins, and M…
Researchers discovered two sandbox bypass vulnerabilities in Cursor IDE that allow remote code execution via prompt injection. Tracked as CVE-2026-50548 and CVE-2026-50549, the flaws let attackers break out of Cursor's c…
XAI released Grok 4.3 on April 17, 2026, with the Grok Build now installed in Railway sandboxes. The model has achieved a score of 8 on the GDP.pdf benchmark and ranks #57 on the LM Arena leaderboard with 37,636 votes.
The European Union AI Act will require LLM watermarking by August 2026, but experts argue text watermarks are inherently easy to remove due to the compressed nature of text. Google's SynthID and other methods face challe…
Mastra AI argues that model routers require rigorous testing beyond simple model selection, introducing scorers, datasets, and experiments to validate routing decisions across quality, cost, speed, and safety axes. The c…
Sam Altman spent a year pitching the White House on a 5% stake in OpenAI, and the offer now includes rival AI developers, signaling a shift where the US government moves from monitoring frontier AI to holding an ownershi…