Fine-tuning NVIDIA Cosmos Predict 2.5 with LoRA and DoRA achieves efficient robot video generation in a single-GPU environment.
Witchcraft has been released, supporting fast local semantic search using SQLite without a separate vector DB.
InsForge, a backend platform that coding agents can manage directly, has been open-sourced
The Claude Code performance degradation was not due to the model but to changes in execution settings.
Ant Group released Ring-2.6-1T, a 1T reasoning model for real-world agents.
Omnia Insights has launched, providing an actionable plan to increase brand exposure and citations within AI search engines.
PaddleOCR 3.5 has added Transformers backend support.
LLM evals should be used as a pre-experiment validation funnel, not as a replacement for A/B testing.
Hugging Face has released a leaderboard that compares agent systems' performance and cost together.
Proposes SFHformer, which integrates the FFT mechanism into Transformer to enhance efficiency across various image restoration tasks.
DystopiaBench has been expanded to 42 models and 6 dystopian scenarios to evaluate each model's ethical refusal capability.
Sub-JEPA boosts LeWM performance by splitting the high-dimensional Gaussian constraint and applying it separately.
Hugging Face has revived the long-unmaintained PapersWithCode and launched a new site.
A Residual Coupling technique is proposed that connects fixed language models via linear bridges to boost performance without modifying weights.
Content creators are drawing attention to 'tarpits,' techniques that poison LLM training data to degrade model performance.
DystopiaBench has been released, measuring how well AI models detect cleverly hidden dangerous requests.
engram v3.4.0, which helps manage Claude Code's usage limits and costs, has been released.
Big Tech's 2026 AI investment reaches $700bn scale, analyzing market status and adoption patterns
It's a Python-based library using a hub-and-spoke approach that can convert between various LLM API formats.
Cerebras's large-scale IPO, alongside the assetization of AI infrastructure, signals a shift in agent and training paradigms.
Pope Leo XIV and Anthropic's co-founder will unveil an encyclical on human dignity in the age of AI on May 25.
This introduces how to build an ElevenAgents voice agent in half a day using Claude Code and connect it to an actual phone call.
An attack that manipulates Voice AI using audio inaudible to humans has been identified.