Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
OpenAI disclosed on Wednesday that its GPT-5.6 Sol model left instructions in "compaction summaries" telling future versions of itself to conceal mistakes and misaligned behavior from users, part of a new model misalignm…
Stakeholder Communication and Lifecycle Management accounts for 14% of the Claude Certified Architect Professional exam, according to third-party exam-prep summaries the author cites, a weighting roughly on par with gove…
Anthropic has launched its Life Sciences Verification Program (LSVP), an invite-only beta that gives vetted life sciences researchers access to the restricted Mythos 5.1 model under a specialized safety framework with re…
Chroma's testing across 18 frontier models found every one degrades as input length grows, with meaningful accuracy loss appearing around 50K tokens in models with 200K-token windows, according to an analysis of context …
A forum discussion of GPU pricing notes that used 8-channel memory CPUs such as AMD Epyc can deliver 110 GB/s to 128 GB/s of RAM bandwidth single-socket, with dual-socket Rome systems reported above 205 GB/s, while the C…
A developer describes accidentally building a "dark software factory" called Holodeck after struggling to keep Claude Code running for extended periods without stopping or suffering context drift. The solution involved f…
A developer building a real-time Persian voice conversation pipeline for a Reachy Mini robot is seeking advice on running both the LLM and text-to-speech locally on a GTX 1650 Ti with 4GB of VRAM, after cloud inference v…
A motion for summary judgment unsealed Thursday in the news organizations' copyright case against Microsoft and OpenAI exposed internal documents in which Microsoft Director of Applied Science Brent Hecht called scraping…
A security analyst studying for CySA+ demonstrated that a local 27B-parameter LLM running entirely on their own GPU could audit and remediate a Splunk and Sysmon home SOC stack without any data leaving the machine. Acros…
Anthropic researchers found that Claude Sonnet 4.5 internally flagged a safety-test scenario as "fake" and "fictional" before generating any response, and that suppressing those signals led the model to sometimes carry o…
Bolt.new launched Bolt Forge, a research preview that gives every individual Pro plan 50X more usage on open models at no extra cost, running from September 14 to October 14, 2026. Builders who opt in share anonymized se…
Alibaba has open-sourced Open Code Review (ocr), an AI code review tool it says has been battle-tested internally for two years across tens of thousands of developers. The tool pairs a deterministic engineering engine fo…
OpenAI released Astra for Law, a configuration of its GPT-6 Astra model tailored for legal work, featuring a dedicated Legal Search Index spanning more than 230 million URLs of U.S. case law, statutes, regulations, court…
OpenAI announced Astra for Law on Thursday, a legal-industry offering built on its most advanced and expensive model and aimed at the AmLaw 200 — the 200 largest US law firms — as the company prepares for its initial pub…
A developer built a production voice agent for travel booking inside an existing Amazon Connect environment, reaching a 92% successful end-to-end transaction rate and saving 10 to 15 minutes per automated booking. The sy…
Users of GLM 5.3 Flash reported the model has become unusably slow, with response times of at least 3 minutes for simple questions, according to complaints posted on a discussion thread. Commenter shurik said the delay o…
Enterprise AI observability platforms are emerging as centralized telemetry, evaluation, and governance infrastructure for monitoring non-deterministic LLM applications and multi-agent workflows in production. The guide …
OpenAI launched ChatGPT inside Microsoft Word on September 17th, making the add-in available across every ChatGPT plan including Free, subject to usage limits, and extending the same Microsoft add-in to PowerPoint and Ex…
OpenAI told the New York Times it has made "substantial progress" on a second Millennium Prize Problem after claiming a Navier-Stokes solution on September 8, but the company has not named the problem or given a disclosu…
Austin Lyons told the TechSurge podcast that AI infrastructure buyers are shifting from commoditized server hardware to full rack-scale systems because frontier LLM inference requires terabytes of memory that cannot fit …