AI
Artificial intelligence: LLMs, agents, diffusion, vision, NLP and the latest from top labs.
593 published articles
Anthropic AI for Science rare disease grants
$50,000 in Claude credits won't cure rare diseases, but it might map the path
Anthropic's AI for Science rare disease grants offer up to $50,000 in Claude credits to researchers and early-stage biotechs. Applications are open through August 2, 2026.
2026-07-21
AI Research
Qwen just taught diffusion models a trick from the LLM playbook: RL beats supervised fine-tuning
The Qwen-Image-2.0-RL report details a post-training pipeline that combines RLHF, on-policy distillation, and composite reward models to improve text-to-image and image editing quality. The approach yields measurable gains across aesthetic quality, instruction following, and face identity preservation, borrowing techniques from LLM alignment research.
2026-07-20
Alibaba Qwen-Music Model and Melody Chain-of-Thought
Alibaba's Qwen-Music composes melodies first, then fills in the band, and out-sings Suno
Qwen-Music breaks new ground by explicitly planning melodies before generating full songs, a technique the team calls Melody-CoT. The model achieves state-of-the-art results on 13 of 16 objective metrics and, in blind tests, wins against Suno V5 and MiniMax Music 2.6.
2026-07-20
Foundations of AI
DeepMind's new framework shows why AI is brilliant at some things and terrible at others
A Google DeepMind paper introduces a three-part framework for thinking about rigor in AI: conceptual, epistemic, and operational. It argues that deep learning's rapid progress has come from prioritizing performance-driven iteration over scientific understanding, and that closing the gaps will require more than better benchmarks.
2026-07-20
Multimodal AI
Alibaba just shipped a model that hears, sees, and speaks, and it runs on a phone
Alibaba's Qwen2.5-Omni processes text, images, audio, and video end-to-end, generating text and natural speech in real time. Benchmarks show it matches specialized models in vision and audio, and a 4-bit quantized version runs on an RTX 3080 or a Snapdragon phone.
2026-07-20
Speech AI
Voice AI said 'I understand your frustration.' It had no idea what that meant.
Hume's Real World VoiceEQ benchmark, based on more than one million human ratings, tests over 40 voice models across dimensions standard benchmarks ignore: emotion, speaker identity, and acoustic context. The findings show speech-to-speech models vary wildly, and even leading systems often ignore the audio cues humans use instinctively.
2026-07-20
Open-endedness & VLMs
Sakana AI's Picbreeder reboot shows what creativity metrics miss
A Sakana AI-led team replaced human Picbreeder users with VLMs and found the synthetic archives lack the boldness and diversity of the human originals. Their experiments with noise, memory, and thousands of prompted personalities reveal both the promise and the limits of using large models for open-ended discovery.
2026-07-20
Infrastructure
Kimi's K3 demand just crashed its own GPU. New subscriptions are on hold.
Kimi.ai pauses new K3 subscriptions after 48 hours of demand stress. Existing users unaffected. Plans to split membership into general and coding plans.
2026-07-19
Biologically plausible AI's hidden warning
The CIFAR-10 ablation that should make every AI researcher rethink their benchmarks
Sakana AI's Error Diffusion scales brain-like, backprop-free learning to CIFAR-10 and RL. The catch is in the ablation tables: which design choices matter reverses entirely between benchmarks, and the cost of Dale's principle grows with task difficulty.
2026-07-19
Artificial Intelligence
The AI that taught itself to revise: four agents, one puzzle, no consensus
Researchers propose ARCANA, a multi-agent system that cracks ARC-AGI-2 puzzles by splitting reasoning into four steps and closing the loop with reflective feedback. Each failure teaches the next attempt. The architecture improves reasoning efficiency, but questions about generalization and compute cost remain.
2026-07-19
AI Research
RDPO: The two-word fix for reinforcement learning's self-sabotage problem
Multi-task reinforcement learning has a dirty secret: the reward signals that drive alignment often work against each other. An international team just published a method that prevents the system from fighting itself, and it costs almost nothing to add to existing pipelines.
2026-07-19
Autonomous Agents
Hermes Agent gets smarter the longer it runs, and that changes everything
Hermes Agent from Nous Research goes beyond the typical coding copilot by embedding a closed learning loop, cross-session memory, and autonomous skill creation. It runs anywhere, integrates with 20+ messaging platforms, and aims to build a deepening model of each user over time.
2026-07-19