The Daily Token

Frontier

Top Story

voicesLatent Space

[AINews] Much ado about Open Weights

Everyone is writing a lot, but only Kimi K3 shipped today

1h ago/Latent.Space

Techmeme

Cursor introduces Cursor Start, a ~$7 monthly subscription exclusive to India, its third-largest market, priced below its standard $20 Pro monthly subscription

3h ago

TechCrunch AI

Cursor makes its biggest India push yet ahead of SpaceX acquisition with localized pricing

3h ago

Top Headlines

arXiv cs.AI·4h ago

Concept-based Visual Counterfactual Explanations with Diffusion Models

arXiv cs.CL·4h ago

Explaining GAND: A Resource on Gender-Ambiguous Natural Data & Contrastive Attribution

arXiv cs.LG·4h ago

Semalith v1.4: A Calibrated 184M Safety Classifier Achieving State-of-the-Art Prompt-Injection Detection at 44x Fewer Parameters than Llama-Guard-3-8B

Bloomberg Tech·6h ago

Anthropic’s Amodei Rejects Open Model Ban, Calls for Testing

From the Labs

Frontier model builders

Hugging Face

NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics

22h ago

Hugging Face

Bringing Nunchaku 4-bit Diffusion Inference to Diffusers

5d ago

Business & World

Market · companies · moves

The Verge AI

Why China is giving away its best AI models

People visit the booth of Kimi, an LLM developed by the Chinese startup Moonshot, during the World AI Conference in Shanghai, China, July 20th. | Image: LONG WEI/ Feature China/Fut

15h ago

Ars Technica AI

Canadian legislator reads out apparent LLM response in floor speech

Voices

Live from X

Import AI@Import AI·18h

Import AI 466: The bitter lesson for robotics, AIs complete week-long programming tasks; and OpenAI’s accidental AI hacker

Latent Space@Latent Space·3d

[AINews] Claude Opus 5: Fable-level performance at Opus price (half Fable)

The Wire

Latest · auto-refreshed

arXiv cs.AI

SeT-Diff: Towards Semantic Foundation Models for HPC Telemetry and Time-Series

4h ago arXiv cs.AI

QFoldAgent: An Autonomous Quantum Optimization Multi-Agent System for Protein Structure Prediction

4h ago arXiv cs.AI

Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy

4h ago arXiv cs.AI

DeepLens Diagnosis Agent: Agentic Workflow Design Lets a Small Reasoning Model Compete with Frontier LLMs

4h ago

[AINews] Much ado about Open Weights

Cursor introduces Cursor Start, a ~$7 monthly subscription exclusive to India, its third-largest market, priced below its standard $20 Pro monthly subscription

Cursor makes its biggest India push yet ahead of SpaceX acquisition with localized pricing

Concept-based Visual Counterfactual Explanations with Diffusion Models

Explaining GAND: A Resource on Gender-Ambiguous Natural Data & Contrastive Attribution

Semalith v1.4: A Calibrated 184M Safety Classifier Achieving State-of-the-Art Prompt-Injection Detection at 44x Fewer Parameters than Llama-Guard-3-8B

Anthropic’s Amodei Rejects Open Model Ban, Calls for Testing

From the Labs

NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics

Bringing Nunchaku 4-bit Diffusion Inference to Diffusers

Business & World

Why China is giving away its best AI models

Canadian legislator reads out apparent LLM response in floor speech

Voices

The Wire

SeT-Diff: Towards Semantic Foundation Models for HPC Telemetry and Time-Series

QFoldAgent: An Autonomous Quantum Optimization Multi-Agent System for Protein Structure Prediction

Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy

DeepLens Diagnosis Agent: Agentic Workflow Design Lets a Small Reasoning Model Compete with Frontier LLMs

[AINews] Much ado about Open Weights

Cursor introduces Cursor Start, a ~$7 monthly subscription exclusive to India, its third-largest market, priced below its standard $20 Pro monthly subscription

Cursor makes its biggest India push yet ahead of SpaceX acquisition with localized pricing

Concept-based Visual Counterfactual Explanations with Diffusion Models

Explaining GAND: A Resource on Gender-Ambiguous Natural Data & Contrastive Attribution

Semalith v1.4: A Calibrated 184M Safety Classifier Achieving State-of-the-Art Prompt-Injection Detection at 44x Fewer Parameters than Llama-Guard-3-8B

Anthropic’s Amodei Rejects Open Model Ban, Calls for Testing

From the Labs

NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics

Bringing Nunchaku 4-bit Diffusion Inference to Diffusers

Business & World

Why China is giving away its best AI models

Canadian legislator reads out apparent LLM response in floor speech

Voices

The Wire

SeT-Diff: Towards Semantic Foundation Models for HPC Telemetry and Time-Series

QFoldAgent: An Autonomous Quantum Optimization Multi-Agent System for Protein Structure Prediction

Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy

DeepLens Diagnosis Agent: Agentic Workflow Design Lets a Small Reasoning Model Compete with Frontier LLMs

Anthropic’s Dario Amodei responds: doesn’t oppose open-weight models, but fears Chinese AI

Claude chats showed in Google and Bing search results, despite Anthropic's robots.txt saying not to crawl them, likely because the pages lacked a "noindex" tag

Accelerating the frontiers of scientific discovery: Google’s $40M commitment to the Genesis Mission

3 Google updates from Galaxy Unpacked 2026

The SpaceX IPO and Data Centers in Space

MIITA: Memory-Induced Inference-Time Adaptation for Continual Learning with Small Language Models

SF-AMS: Strategic Forgetting for Structured Memory in LLM Agent

MioFFAn: an Annotation Software for Formula Formalization with LLM Automation Capabilities

Between Suppression and Collapse: Evaluating Narrative Unlearning with LENS

PatiGonit22K: A Comprehensive Dataset for Solving Complex Bengali MWPs

CORVUS: Context Optimization and Reduction Via Underlying Synchronization for LLM Coding Agents

CausalGate: Causal Importance Distillation for Transformer Module Pruning

Progress-conditioned Group Policy Optimization for Long-Horizon Agentic Tasks

QFedPolyp: A Communication- and Inference-Efficient Federated Learning Framework for Polyp Segmentation

Learning to Access Computation: Accessibility Plasticity as a Principle of Adaptive Intelligence

Hierarchical Grading in Large Language Models

Satya Nadella says companies that trust one AI for everything may not survive

PSA: Your Claude shared chats and Artifacts may have ended up on Google

Microsoft launches its first cybersecurity model, plus a new agentic cybersecurity system

OpenAI’s Hugging Face breach has reignited the debate over alignment and control

Threads users can now chat with Meta AI in their DMs

Nvidia, Microsoft launch open AI security alliance — without OpenAI, Google, or Anthropic

China Draws Lines on US Tariffs, AI Ahead of Xi-Trump Summit

How AI is expanding what people do at work

Anthropic's Opus 5 is about token efficiency, not a capability leap

Midjourney bought the astrology app Co-Star