Full-text search across 3496 articles. Combine with topic and date filters; results sorted by relevance.
Liquid AI released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, built on the LFM2 hybrid backbone with an 8,192-token context. The 350M model achieves a 17-task mean of 81.02, rank…
Antics has released an SDK that adds real-time multiplayer to AI-generated games without requiring server-side code, using a hosted relay for room management and host-authoritative state sync. The SDK supports up to 8 pl…
A developer built a pipeline to auto-generate AI model comparison pages for aiappdex.com, pairing models by HuggingFace pipeline tag and using Claude Haiku to produce structured side-by-side evaluations. The approach gro…
A developer proposes using typed memory categories with Mem0 to improve LLM agent memory management, addressing failures from treating all facts equally. By storing memories as 'user', 'feedback', 'project', or 'referenc…
Semiconductor stocks are sliding across US and Asian markets as investors shift from the AI hype cycle to demanding proof of returns on massive capital expenditures. Big Tech is spending billions on Nvidia H100s and B200…
A developer warns that the term 'agent' is being diluted, causing real engineering mistakes, and argues that most real agent deployments are narrow, purpose-built pipelines. The developer defines an agent as a system wit…
A developer created a userscript that merges Hacker News articles and their comment threads into a single view, eliminating the need to switch between tabs. The script adds a resizable side panel for side-by-side reading…
LearnVector is building an AI-powered personalized education platform that uses a multi-agent workflow to manage a student's zone of proximal development, providing minimal hints rather than answers. The system employs d…
A developer building a Jira Backlog Analyzer found that a single Retrieval-Augmented Generation (RAG) pipeline produced generic or outdated recommendations. By splitting project knowledge into three separate RAG sources—…
Manim WebGPU, a browser-based version of the math animation tool, now runs entirely in the browser with GPU-accelerated rendering via WebGPU, WebAssembly, and Pyodide. The tool features a built-in AI assistant that lets …
A developer built an HRV anomaly detector using the Isolation Forest algorithm from Scikit-learn to flag potential overtraining days from Oura Ring data. The system fetches daily readiness scores via the Oura Cloud API a…
A developer's hands-on comparison of AI agent frameworks finds LangGraph best for production due to its precise state machine, despite a steep learning curve, while CrewAI and AutoGen are better suited for demos and rese…
A security researcher found that Granola's OAuth flow lacked strict state parameter verification, enabling account takeover via a crafted callback URL that could hijack a victim's session and inherit webcam permissions. …
Mazu AI is scaling weather forecasting for the Global South by using AI-driven pattern recognition on historical atmospheric data, reducing warning latency from hours to minutes without requiring supercomputers. The syst…
OpenAI's GPT-5.4 nano Model failed to correctly sort unique words from a passage of Homer's The Odyssey, a task that is trivial with a scratch workspace but difficult for current LLMs lacking effective mutable memory, ac…
A new supply chain attack called 'slopsquatting' exploits AI hallucinations by tricking developers into installing malicious packages that LLMs invent, with research across 500,000 code samples and 16 LLMs showing a 19.7…
Proposed trade restrictions on Chinese humanoid robots could slow embodied AI development in the US by limiting access to affordable, high-performance hardware, according to an analysis of the hardware-software gap. Deve…
A hands-on comparison of MCP server support in Cursor, Claude Code, and Windsurf finds that Claude Code's CLI implementation has the lowest tool latency at ~0.8 seconds, while Windsurf handles multi-tool agentic chains w…
A team transitioning to agentic AI deployment found that treating LLM agents as advanced chatbots failed, so they shifted to a tool-use architecture with sandboxed Docker containers, human-in-the-loop permission layers, …
A staff letter from AI labs including OpenAI and Anthropic calls for government-backed safety standards to prevent a 'race to the bottom' in risk management, arguing that shipping fast often takes precedence over rigorou…