Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A developer released Codeflowmap, an open-source tool that maps a codebase's dependency and call graph to surface read, write, and auth paths between files and functions. It can annotate files using local or remote AI mo…
Vitrus, a new knowledge management tool, provides sourced answers and explicitly lists what it doesn't know, using deterministic gap detection to avoid hallucination. It offers agent-native API integration, onboarding pa…
A Hacker News user asks the community for parameter count estimates of the hypothetical models Opus 4.8 and GPT-5.5, noting that frontier labs keep flagship sizes secret.
Cisco Foundation AI open-sourced FAPO, a Claude Code-driven system that autonomously optimizes multi-step LLM pipelines by attributing failures at the step level and proposing variants. In evaluations, FAPO outperformed …
Google DeepMind released Gemma 4 12B, an encoder-free multimodal model that processes visual and audio inputs directly through a single decoder-only transformer, enabling local agentic workflows on standard 16GB laptops.…
Callimachus, a new open-source desktop app, indexes and searches conversations from 11 AI coding agents into a local SQLite store, enabling hybrid keyword and semantic search across tools like Claude Code and Cursor. The…
Two former OpenAI founders, Thomas Dimson and Joey Flynn, launched In the Weights, a free tool that measures how strongly AI models like GPT-5.5 and Claude Opus 4.8 can identify and describe a person from their internal …
Developer Dom Sadarangani released Domvault, an open-source, self-hostable second brain tool that integrates with Obsidian and Claude AI to create a self-improving knowledge graph. The tool features a consent-based publi…
A new benchmark called FunnyBench tests AI models' ability to tell funny jokes by asking each model to generate ten jokes and letting users vote on their humor. The live leaderboard uses a Bayesian score to rank models b…
A developer released TLA+ Process Studio, an open-source tool that uses LLMs to help stakeholders model and discuss business processes as state machines, aiming to improve alignment before coding. The tool generates TLA+…
A new study reveals that 45% of consumers now use AI assistants to find local services, yet ChatGPT recommends only 1.2% of all local business locations. This shift is causing a dramatic decline in click-through rates fo…
Maccha, a lightweight file architecture and set of markdown templates, enables cross-agent persistent memory for AI coding assistants like Antigravity, Claude Code, and OpenCode, allowing them to share a unified digital …
NVIDIA has reportedly acquihired the team from Essential AI, including co-founder and Transformer paper co-author Ashish Vaswani, to join its Nemotron research group. The move strengthens NVIDIA's in-house model research…
Save, a new API that converts any URL into clean Markdown optimized for large language models, launched with a free tier of 1,000 pages per month. The service fetches web pages server-side, optionally renders JavaScript,…
A developer warns that language models cannot distinguish between data and commands, making AI agents vulnerable to prompt injection attacks via web pages, tool outputs, or supply chain components. The developer argues t…
A developer successfully ran a 35-billion-parameter Mixture-of-Experts model on a 2017 AMD RX 580 8GB GPU using Vulkan, bypassing CUDA and ROCm. The project, called Polaris Revival, achieved 17-18 tokens per second for L…
Bluffbench, an evaluation for LLM plot interpretation, is nearing saturation as models like Fable 5 achieve high scores on the hardest 'mocked' cases, though human-level performance remains elusive. The benchmark will so…
A new scheduling technique called TurboPrefill reduces waiting time for Vision Language Models by nearly half, from 9.0 to 4.6 seconds, without changing model weights or architecture. The optimization, validated on Qwen2…
A new class of post-trained AI security models, such as ArgusRed and PentestGPT, is enabling automated penetration testing by bypassing standard safety refusals and executing exploits in sandboxed environments. These sys…
DeusData released codebase-memory-mcp, an open-source MCP server that indexes codebases into a persistent knowledge graph, enabling AI coding agents to answer structural queries in under a millisecond instead of grepping…