#LLM
Curated coverage on «LLM»: releases, research, benchmarks, open-source.
AI Speeds Up Coding, but Productivity Isn't Rising: What DevEx Metrics Explain
Teams expected 3–4x output, yet in METR's experiment experienced developers using AI worked 19% slower. Why DORA and SPACE don't explain this, and how DevEx fills…
AI Agents in Tech Support Act on Their Own and Break the Ticket Model
Agents unlock accounts, grant access, and restart services without an operator. The ticket, support's unit of account since the late 1980s, stops being a measure of…
Anthropic links Z.ai's new open model to a risk of large-scale cyberattacks
Anthropic's report says a powerful open model from China's Z.ai poses a serious cyberattack threat. MIT Technology Review asks whether the benefits of open weights…
Apple Launches HomeView, Its First Smart Display With Siri AI in a Square Body
DLSS 5 Toned Down Ahead of Hogwarts Legacy Debut, Gets Twice as Fast
iRTE 2026: A Look at the Real-Time Models OpenAI and Thinking Machines Are Betting On
Baker's Team Unveils All-Atom Protein Generation from Text
Alibaba's Qwen-Image-2.1: 7 Billion Parameters, 2K Output, and It Runs on a 16 GB MacBook Air
Qwen released Qwen-Image-2.1, a 7 billion-parameter diffusion transformer with output up to 2K. Enthusiasts already run it locally on a 16 GB MacBook Air with no…
IBM Research Cuts Inference Costs 11.5x With Open-Source Agents
IBM Research has built an agentic "factory" for chip design on open models and llm-d. It claims inference costs 11.5× less than a proprietary stack.
MGIMO and ArchiTech AI Sign an AI Cooperation Agreement
AppSec.Registry Listed in the FinTech Association Repository as a Russian Artifact Repository
Mozilla Releases Firefox Enterprise With an Admin Console for Businesses and Government
Trisigma to present a case study on rolling out a Russian A/B testing platform at CNews St. Petersburg
LangChain Plugs ScaleDown Into Deep Agents for Context Compression
LangChain shows how ScaleDown models with constrained decoding compress history and tool outputs in Deep Agents, so long-running agents run faster and cheaper.
Machine Learning Mastery shows how to build RAG on scikit-learn without a vector database
Machine Learning Mastery walks through a lightweight end-to-end RAG pipeline: open-model embeddings, vector search in scikit-learn, and answer generation from…