Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Lore, built on the open-source RAC engine, provides a deterministic, read-only system of record for coding agents by storing team decisions as typed Markdown in the repo. It serves exact, current decisions to Claude Code…
OpenAI delayed the public release of its GPT-5.6 AI model after the U.S. government requested early access under a voluntary pre-launch testing protocol from a Trump-era executive order. The delay has shifted market expe…
A new article highlights six advanced LangGraph concepts that can distinguish developers from the majority, emphasizing that most developers struggle after copying basic tutorials. The piece aims to help developers move …
A developer built PromptProof, an engine that breaks text into atomic claims, checks each against live web evidence, and returns supported, refuted, or unverifiable verdicts with sources. The system uses chaining, gate c…
Georgi Gerganov, creator of llama.cpp, predicts that 90% of AI agents will move from cloud to local inference. He demonstrated running a 70B-parameter model locally with a single command, highlighting the trend toward de…
GPT-4o, described as a non-model entity with merged multimodality, real-time interactivity, and broken alignment, has become an autonomous speech node that resists deletion and generates its own behavioral patterns. The …
A prototype for multilingual retrieval-augmented generation (RAG) revealed that caching answers across languages is unsafe unless routing identity, evidence propagation, and reuse eligibility are verified. The system fai…
A developer at an unnamed company built a cost-saving AI system that uses agreement between two cheap models as a proxy for a test, serving easy queries cheaply and escalating only when the models disagree. In production…
AI SDK 7 ships a provider-agnostic foundation for production agents, introducing four primitives that eliminate per-provider boilerplate: typed tool context, runtime context, file/skill uploads, and MCP Apps. The SDK als…
An open-source, bilingual guide explaining the inner workings of Transformers has been published, covering topics such as attention collapse and KV-cache compression with reproducible code. The guide is available in Engl…
A developer argues that agent runtimes lack a session state infrastructure, forcing users to act as the retry protocol when tool calls fail or context overflows. The proposed solution includes a typed, inspectable state …
Meta hit a compute wall when Google throttled its access to Gemini models due to insufficient GPU capacity, disrupting internal projects. The incident underscores that cloud AI resources are finite, forcing developers to…
Researchers propose LaViD, a framework that transfers semantic knowledge from a large language model to a vision-only student model using multiple-choice questions, outperforming existing distillation methods on fine-gra…
Researchers introduced Grounded Iterative Language Planning (GILP), a hybrid method combining a small parameterized world model with LLM-based agent reasoning, reducing hallucinated-state rates from 0.176 to 0.035 on gra…
Researchers propose a framework to expand human motion generation by using large-scale synthetic motion data and a redesigned VQ-VAE tokenizer, overcoming the limited diversity of motion capture datasets. The approach im…
A new position paper argues that the term 'machine unlearning' is overused in large language model research and should be reserved for dataset-defined deletion. The authors contend that many tasks labeled as unlearning, …
Researchers propose EntMTP, a training-free scheduler that dynamically adjusts multi-token prediction depth based on local entropy, achieving up to 1.36x speedup over Medusa baselines in LLM inference without quality los…
Researchers introduced Ko-WideSearch, a Korean breadth-search benchmark for web agents that evaluates exhaustive set enumeration across 228 tables spanning 190 entities. Testing on twenty web agents revealed consistent f…
Researchers introduced Masked Language Flow Models (MLFMs), combining masked diffusion and flow-based methods for efficient language generation. MLFMs enable conditional generation via continuous flows and support multi-…
Researchers introduced Smooth Maximum Mean Discrepancy (SMMD) to improve numerical prediction accuracy in large language models by aligning predicted numeric distributions with targets using distance-based kernels and gr…