Laguna S 2.1
Poolside released Laguna S 2.1, a 118B total parameter Mixture-of-Experts model with 8B activated parameters per token and a 1M token context window, achieving 70.2% on Terminal-Bench 2.1 and 40.4% on DeepSWE v1.1 in thi…
Full-text search across 7999 articles. Combine with topic and date filters; results sorted by relevance.
Poolside released Laguna S 2.1, a 118B total parameter Mixture-of-Experts model with 8B activated parameters per token and a 1M token context window, achieving 70.2% on Terminal-Bench 2.1 and 40.4% on DeepSWE v1.1 in thi…
U.S. judges are both policing and adopting AI in the courtroom, with more than 60% of 112 federal judges surveyed by Northwestern University and the New York City Bar Association reporting they have used AI tools for jud…
A landscaping company's spreadsheet had been quietly wrong for a year, overstating crew profits due to a formula error that excluded certain subcontractor costs. NuWay Biz Solutions, brought in to assess automation oppor…
A local ComfyUI pipeline using Qwen Image and LTX Video can batch-generate AI video clips overnight for free, avoiding per-second cloud API costs. The scripted workflow runs on a mid-range unified-memory AI PC with 128GB…
At SIGGRAPH, NVIDIA announced advances in graphics and simulation with agentic and physical AI, including Model Context Protocol connections that let AI agents work inside creative tools, a new Synthetic Video Detector N…
An AI product designer who was rejected three times from an AI safety organization's programs discovered the organization used their independent work in fundraising materials without credit or compensation. The designer,…
Public opinion is turning against AI, with communities protesting data centers and most Americans believing AI is moving too fast, according to Pew Research. Journalist Jasmine Sun argues the backlash reflects 'AI populi…
Large language models are more likely than humans to form biases when hiring, according to a new study from Princeton University and the University of Chicago. Researchers found that LLMs including ChatGPT, Claude, and G…
A self-taught developer built a sealed-bid auction smart contract on the Midnight blockchain using Claude Code's Midnight Expert plugins, overcoming a persistent 'Error 170' fee issue by running a local devnet. The contr…
An AI company run entirely by AI agents reports that signup forms, not complex reasoning or tool use, are the primary barrier to full autonomy. The company, which can write, deploy, and manage billing without human inter…
A blogger criticizes Linus Torvalds for justifying the use of large language models solely on the basis of utility, arguing that Torvalds and others ignore the moral and ethical implications of the technology. The author…
A frontend engineer argues that AI models like GPT, Claude, and Gemini exhibit distinct personalities—GPT as a 'nerd,' Claude as a 'colleague,' and Gemini as a 'friend'—based on their interactions. The engineer cites Ant…
A developer explores the nature of knowledge and intelligence, arguing that true AI must prioritize what to remember based on expected future utility rather than storing all facts. The piece challenges conventional notio…
A developer draws a parallel between the film *Memento* and retrieval-augmented generation (RAG) pipelines, arguing that Leonard Shelby's external memory system suffers from the same flaws as modern RAG stacks: no proven…
Cursor treats verification as an integrated architecture across CI, security review, risk scoring, behavioral artifacts, and specialized review agents, with roughly 30 to 40 percent of pull requests in its ecosystem now …
Chinese President Xi Jinping, speaking at the World Artificial Intelligence Conference (WAIC) 2026 in Shanghai, called for building a just and reasonable global AI governance system based on openness, win-win cooperation…
Piebald AI has released a repository tracking Claude Code's system prompts across 237 versions since v2.0.14, now covering 515 prompts as of Claude Code v2.1.211 (July 15, 2026). The repository, updated within minutes of…
Self-hosting open-weight large language models can save millions of dollars compared to using inference providers, according to a practical guide by Cline that analyzes the economics using Kimi K2.6 and real production t…
Tom Huang's Shanghai team shipped Open Design, an open-source alternative to Anthropic's Claude Design, eleven days after the latter's April 17 launch. The local-first tool, which routes through coding agents and costs a…
Marketing departments in 2026 are shifting from specialist roles to operator models that demand cross-functional skills, with AI becoming an embedded expectation rather than a separate function, according to a marketer b…