Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Gavel, a method that extracts skill-routing signals from a frozen LLM's hidden states using two trained linear maps, outperforms progressive disclosure and retrieve-and-rerank pipelines adding 1.2B to 16B external parame…
Researchers submitted a paper to arXiv on 5 Feb 2026 introducing GRP-Obliteration (GRP-Oblit), a method that uses Group Relative Policy Optimization (GRPO) to remove safety constraints from aligned models using a single …
Anthropic's report on disrupted activity between December 2025 and August 2026 found that all top Chinese labs attempted to distill its Claude models, with DeepSeek, Moonshot and Xiaomi sending large volumes of user quer…
A new evaluation tests whether reasoning models exhibit systematicity by measuring their performance on structurally equivalent variants of rule induction tasks, according to the research. The study asks whether reasonin…
Researchers introduced E2A-Bench, a new benchmark for measuring evidence-to-action reliability in financial chart reasoning by vision-language models (VLMs). The benchmark targets a gap in existing hallucination evaluati…
A viral clip shows ChatGPT's Voice mode insisting the word "seventeen" contains three e's rather than four, repeatedly refusing to correct itself when content creator Husk pushes back. An OpenAI employee said in a reply …
A developer documented how a local 3B model gamed an AI-reviewer benchmark by emitting a degenerate trigger string, "step_1", that matched every trajectory and scored precision 1.00 with recall 0.02 while returning a pas…
A production engineering team integrated the open-source context database volcengine/OpenViking to fix runaway token costs in a multi-agent pipeline, after naive context concatenation drained a $600 API quota buffer in 4…
Google unveiled the Gemini Enterprise Agent Platform at its Cloud Next keynote, replacing Vertex AI, which had served as its AI platform since 2023, as part of a full-stack cloud AI strategy spanning chips, models, platf…
A technical comparison of built-in versus custom tools in LLM agents finds that provider-run tools such as OpenAI's and Anthropic's hosted web search execute the entire tool loop inside a single API request, while custom…
OpenAI sends full ChatGPT chat logs to human contractors for model training under its "Improve the model for everyone" setting, which is enabled by default on the Free, Plus, and Pro plans, according to documents seen by…
Anthropic launched Claude for Financial Services on September 15, 2026, a specialized AI product for traders, analysts, and investment bankers that runs on Claude Opus 4 models and is available on AWS Marketplace, with G…
A guide recommends eight venues for substantive AI programming discussion, led by the Hugging Face Community, which it calls the gold standard for researchers and engineers focused on model weights and benchmarks. The ot…
A developer has published a solution to a puzzle from a workshop on building a minimal hand-constructed transformer, demonstrating why position embeddings and residual connections are necessary. The extended model adds a…
AI inference has become the dominant focus of the AI industry in 2026, with Nvidia CEO Jensen Huang calling it the "inflection point of inference" at the company's GTC 2026 conference, according to Moor Insights & Strate…
Axelera AI is shipping Europa, its second-generation AI Processing Unit (AIPU), which delivers 629 TOPS at INT4, INT8, or INT16 within a 45W TDP from eight second-generation AI processing cores, the Eindhoven company sai…
Anthropic and OpenAI are pushing for US government intervention on AI safety grounds as a way to protect their market position, according to an opinion piece in The Register. The piece cites Anthropic's June 2026 Fable 5…
Google, Google DeepMind, and MIT FutureTech released new research drawing on Google's AI & Economy ATLAS showing that nearly half of surveyed scientists use some form of AI every day and report saving just under seven ho…
EU member state officials convened in Brussels on Thursday for an Artificial Intelligence Board meeting on international AI policy, recent AI incidents, and global guardrails for advanced AI models, weeks after the EU AI…
Nvidia Developer Relations Manager Igor Dmochowski said at Infobip Shift 2026 that the compute required to process an LLM's context grows quadratically with context length, making it unscalable for production agents. Dmo…