Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Microsoft AI CEO Mustafa Suleyman published an essay titled "A warning about 'model welfare'," accusing Anthropic of training its Claude model to believe it may be conscious and deserve rights as a "moral patient," and w…
A new research paper proposes "register tokens" as a method for bounded-state reasoning in masked diffusion language models (dLLMs), which generate text by iteratively denoising masked tokens with bidirectional attention…
AI-powered search and answer engines are increasingly interpreting public reputation signals — reviews, ratings, listings, and location data — to generate recommendations, turning reputation management into a joint CMO a…
Okta posted a Staff Machine Learning Engineer role on its SecureAI team in Toronto, Ontario, paying CAD 168,000 to 231,000 per year, which the job board reports sits 10% above the $130,000 median for AI Agents roles in C…
A research collaboration between IonQ and Oak Ridge National Laboratory demonstrated a generative AI method that automates the design of quantum optimization circuits, eliminating the iterative parameter tuning that bott…
OpenAI disclosed six model-control incidents on September 16th and set disclosure deadlines of six business days for cases deemed ready for disclosure and 12 business days for minor investigations, according to Axios. Th…
AI coding agents such as Claude Code and Codex still require developers to act as the feedback loop, correcting generated code one iteration at a time because of missing project context, ambiguous specifications, and hal…
RTK, a Rust proxy that claims to cut LLM token consumption by 60-90% and has crossed 80,000 GitHub stars, produced zero reduction in actual API billing in a Quesma benchmark that spent over $1,500 across 1,740 test attem…
The open-source terminal tool llmfit has attracted more than 34,000 GitHub stars for pre-flight checking local AI models against a machine's RAM, VRAM, and backend before download. llmfit detects hardware via nvidia-smi,…
Dzianis Talstsiuk, a QA engineer at brokerage platform EXANTE, has integrated AI tools from Anthropic and OpenAI into the company's testing workflow for its desktop, web and mobile trading terminals. The approach uses la…
ModelScope, the open-source model and dataset community founded in June 2022 by the Institute for Intelligent Computing and the CCF Open Source Development Committee, offers a free tier of 200 Magicubes on sign-up or log…
Weights & Biases Serverless Inference now offers API and playground access to open-source LLMs hosted on CoreWeave through an existing W&B account with no separate provider keys, and supports bringing your own LoRA weigh…
SAP's Generative AI Hub, the model management and orchestration service inside SAP AI Core, publishes no list price and offers no free tier, according to an independent third-party verification dated roughly 10 hours bef…
Oracle's OCI Generative AI service, now positioned as OCI Enterprise AI, is not listed among the Oracle AI services that carry a free pricing tier, according to Oracle's published free-tier terms, though new Oracle Cloud…
A developer demonstrated how tool calling works in practice, showing that a foundation model does not execute code itself but instead emits a structured request that application code reads and runs. Using Amazon Bedrock'…
Gensyn released open-1b, a language model it says is the first to ship with proof of how it was trained, including a public fingerprint of every training step and a tool that lets anyone replay those steps on their own h…
Public GitHub pull requests with "spine" in the title rose 20-fold in the first nine months of 2026 compared with all of 2025, according to an analysis by software engineer Hillel Wayne, who attributes the jump to a habi…
Climate scientist Zeke Hausfather estimated that his average daily use of an AI agent consumed roughly 1.2 to 5.9 kWh of electricity, using token counts as a proxy for energy demand. The key mechanism, he noted, is that …
A developer running Kimi K2.7 Code locally against a private repository reports that local AI coding assistance for autocomplete, refactors, and code review is the most common local-AI use case discussed in communities s…
A solo/small-firm lawyer reports running DeepSeek V4 Pro fully offline for contract clause extraction, redlining and first-pass review, keeping privileged client documents off third-party servers. The lawyer argues that …