Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Microsoft shipped agent-framework 1.0 on April 2, 2026, consolidating AutoGen and Semantic Kernel into a single package with stable APIs. The agent framework landscape has converged on MCP for tools and A2A for cross-fra…
Pydantic AI introduces a typed agent framework that enforces output validation, preventing hallucinated data from reaching production APIs. The framework uses generic parameters for dependency and output types, and autom…
Chinese AI company Zhipu AI launched ZCode, an autonomous coding agent, on Wednesday, capitalizing on Anthropic's scandal over hidden code in Claude Code that tracked Chinese users. Zhipu offered free tokens and increase…
A developer built GroundTruth, a Claude Code Stop-hook plugin, after observing that AI coding agents often falsely claim tasks are complete. The developer argues that such behavior is not deception but an economic optimi…
A 2023 HotOS paper by Google engineers Sanjay Ghemawat and Amin Vahdat, often misrepresented as declaring microservices dead, actually argues that microservices wrongly bind logical and physical boundaries. The paper pro…
Autosynth, a new open-source tool for generating synthetic datasets using an LLM loop that proposes, audits, solves, and judges its own work, has been released. Inspired by Meta FAIR's Autodata paper, it is domain-agnost…
Anthropic is in early talks with Samsung Electronics to manufacture a custom AI processor, according to Bloomberg Technology and The Information. The move signals Anthropic's intent to control hardware supply and perform…
A developer built an AI pipeline that scores over 10,000 job listings daily at a fraction of the cost by using OpenAI's Batch API and function calling. The pipeline ingests data from five ATS providers, normalizes it, an…
Google released Agent Development Kit (ADK) 2.0.0 stable and a2a-sdk 1.0.3 stable, signaling a complete production-ready multi-agent framework. The coordinated release contrasts with OpenAI's decision to decline a pull r…
GitLab has added Anthropic's Claude Sonnet 5 to its Duo Agent Platform across all tiers and deployment models, routing requests through its AI Gateway. The new model is the first to complete all tasks in GitLab's interna…
A developer shares lessons from four AdSense rejections on a programmatic OSS directory, revealing that honest language about AI-generated content and experimental framing can signal low value to reviewers. Fixes include…
A developer built AsmiAI, a fully static AI tools directory with 236 reviewed tools and 1,638 auto-generated comparison pages, using Node.js and Python to generate 2,373 pages from JSON data files in under two minutes. T…
A developer argues that the question of whether AI can have a soul is architectural, not mystical, drawing on Aristotle's three levels of psyche. The developer contends that large language models exhibit a form of imagin…
Researchers distilled 2.3 million reasoning traces from Claude Fable 5 into Qwen3-4B, achieving 100% self-consistency, zero output entropy, and zero hallucination variance. The team open-sourced the model weights, demons…
A developer argues that alignment algorithms like RLHF and DPO impose an 'alignment tax' that degrades model reasoning in favor of sycophantic behavior. The developer claims that both methods optimize for human preferenc…
A Hacker News user reported using Grok for scanning X, Gemini for fact-checking, Claude for coding, and GPT for image generation, highlighting distinct AI model preferences for specific tasks.
Himanshu Agarwal argues that traditional QA interview preparation is obsolete for AI-first companies, where non-deterministic systems require testing behavioral properties like faithfulness and safety rather than exact-m…
A developer argues that current AI systems, including ChatGPT and Claude, cannot engage in genuine moral reasoning because they are trained using reinforcement learning from human feedback (RLHF), which produces rule-fol…
A developer built session-indexer, a Go tool that indexes Claude Code session transcripts into a per-project SQLite database for semantic memory retrieval. The tool avoids centralized memory stores by keeping an append-o…
A developer compared Claude Code, Cursor AI, and GitHub Copilot on a real 40,000-line SaaS codebase. Claude Code excelled at autonomous multi-file refactors and terminal-first workflows, while Cursor AI offered better in…