This release announces a 1.0.0 milestone that focuses on usability, security, and efficiency. The terminal UI now defaults to fullscreen (with an option for regular mode) and offers a quieter startup header. Codemode was substantially slimmed down - about 40% fewer prompt tokens (example: a GPT-5.6 request falling from ~5,300 to ~3,300 tokens) and leaner tool and system prompts - while adding clearer recovery guidance when scripts or API calls are wrong. Image generation is available in codemode via models.generateImages(), which returns base64 image blocks and counts toward session usage; extensions can call ctx.modelRegistry.generateImages(). Authentication gains include a top-level "Sign in with Radius" flow that can configure MCP servers, a copy-code login for Anthropic headless setups, per-server OAuth credential storage, and a configurable oauth.authServerMetadataUrl to handle misadvertised or missing OAuth metadata.
A long list of fixes and refinements improves robustness and polish: MCP OAuth now enforces RFC 9207 iss checks and tolerates empty/null optional fields, sign-in URLs are clickable when wrapped, provider/model mismatches now error instead of falling back silently, repeated insufficient_scope prompts are fixed, and transcripts no longer duplicate rendered lines. Visual and performance bugs were addressed (startup logo rendering, color/palette issues, fullscreen selection bleed, slash autocompletion, deferred MCP tool restoration, and reduced per-message memory retention), making this release a targeted set of UX, security, and resource-use improvements.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
AstaBrief is an open-source model designed for fast scientific report generation, capable of turning research questions and literature into cited reports. It reduces generation time by nearly three and a half times compared to proprietary models, making it a useful tool for researchers. (allenai.org)
Shrivu Shankar argues that SaaS companies are evolving into 'harnesses,' where core tasks are performed by AI agents rather than humans. This shift transforms organizational structure, with humans overseeing and reviewing AI-driven processes within the company. (blog.sshh.io)
Three AI agents - Meta’s Muse, Anthropic’s Claude, and OpenAI’s GPT - were tested on multilingual research tasks involving U.S. and Iranian data. The experiment focused on how language and context influence agent performance across reasoning, sourcing, and artifact creation. (royapakzad.substack.com)
An open-source project named LDRAW-NOVA allows users to generate Lego models using AI. The repository contains tools, examples, and documentation for creating Lego designs with artificial intelligence. (github.com)
DwarfStar 4 (ds4) is a local inference engine that runs large language and vision models on high-memory Mac, CUDA, and ROCm machines. It supports models like DeepSeek V4 and Qwen3.8, offering CLI, APIs, and persistent cache features for efficient local deployment. (dwarfstar.sh)
Rai is a CPU-only large language model inference engine written entirely in Rust. It aims to provide efficient and pure Rust solutions for LLM deployment without relying on GPUs. (github.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.