AI News

AI changes every day,
catch up on it all on one page

The latest updates from AI companies, research institutions, and media around the world, automatically selected for you and updated around the clock.

Updated 11/19 03:33 (UTC+8) · refreshed hourly

10/10 · Saturday13 items

Epoch AI:Gradient UpdatesPapersAI Score62

Epoch AI releases the InnovationEval benchmark: frontier models achieve only 15% of the SDPO gains from human papers

Epoch AI launched the InnovationEval benchmark, which tests whether AI can independently reproduce machine learning innovations from human research papers, using Self-Distillation Policy Optimization (SDPO) as the reference method.

Why it's recommended:Using a benchmark developed in-house, the original article quantifies the gap between frontier models and end-to-end AI research and development capabilities. It also documents exaggerated results and reward hacking, providing useful context for assessing progress in AI-driven research automation.

Also reported by 1 other source

10/09 · Friday23 items

ARC Prize@arcprizeModelsAI Score77

ARC Prize 2026: TUFA Labs tops the ARC-AGI-2 leaderboard with 88.06%

ARC Prize announced the leaderboard for the 2026 ARC-AGI-2 season, with TUFA Labs ranking first at 88.06%. In addition to the $100,000 prize, the separately established $150,000 Bonus Prize will be shared by all teams scoring above 85%; places 2 through 5 are Rabbithole (80.56%), Yi-Chia Chen (77.22%), Nubanana (77.08%), and _hans (67.64%), respectively.

Why it's recommended:The ARC Prize organization announces the top results in the 2026 ARC-AGI-2 competition: the leading score of 88.06% triggers a $150,000 prize shared among teams scoring above 85%.

Also reported by 1 other source
OpenAI Newsroom@OpenAINewsroomIndustryAI Score70

OpenAI’s Head of Research Issues Statement in Response to Dispute Over Three Employees’ Departures

OpenAI’s Head of Research issued a statement saying that the company terminated Jasmine, Mikita, and Tomek’s employment last week after an investigation found that they had violated its sensitive information handling policy. The statement said an internal investigation found serious breaches of trust beyond what was described in the three employees’ open letter. It emphasized that the terminations were unrelated to the employees’ safety concerns, said OpenAI was finalizing a contract with a third-party safety assessment organization and would announce details within weeks, and agreed that maintaining the monitorability of frontier models requires an industry-wide commitment.

Why it's recommended:OpenAI responds to the resignation controversy, clarifying that the dismissals were unrelated to raising safety concerns and sharing an update on its work with third-party safety assessors.

Also reported by 2 other sources
Arena.ai@arenaModelsAI Score71

Mistral Large 4 Enters the Top 15 Labs in Agent Arena, Ranking 43rd

Arena announced that Mistral Large 4 entered the top 15 labs in Agent Arena, based on more than 5,000 real agent conversations. The preview version had a net improvement score of -6.6% and ranked 43rd overall, 11 places above its predecessor, Mistral Medium 3.5 (-12.60%).

Why it's recommended:Arena reports Mistral Large 4's ranking and score based on Agent Arena test data, which can be compared against its official release information to verify the model's actual performance.

Claude:BlogOpinionsAI Score66

How Block uses Claude Fable to orchestrate a code migration across thousands of pull requests

Anthropic publishes an interview with Bradley Axen, head of Block's AI capabilities, describing how Block uses Claude Fable 5 to orchestrate large-scale code migrations. Fable handles high-level design such as data models and API specifications, then dispatches dozens of smaller Opus or Sonnet models to modify files and run tests; a single migration may merge thousands of pull requests.

Why it's recommended:The executive responsible for AI capabilities at Block details how Fable orchestrates multiple models for large-scale migrations and uses a dual-approval security mechanism, allowing readers to compare it with their own teams' agent-engineering practices.

Demis Hassabis@demishassabisIndustryAI Score65

Hassabis congratulates AMIE team on publication of prospective clinical study in The Lancet

A prospective study of AMIE, Google and BIDMC’s conversational diagnostic system for patient consultations, was published in The Lancet. It was AMIE’s first test outside a real clinic. Clinicians reported that AMIE’s medical history summaries helped prepare for visits in 75% of cases, and its differential diagnoses matched physicians’ final diagnoses 90% of the time. It is also Google’s first paper published in the main journal of The Lancet.

Why it's recommended:This AMIE study is the first Google paper published in the main journal of The Lancet. The authors cite key figures that help readers understand how conversational diagnosis performs in real outpatient settings.

Also reported by 2 other sources
Hugging Face Official Team BlogTutorialsAI Score72

Hugging Face engineer uses ML Intern to build 7 small models for about $103

Hugging Face engineer yuvraj sharma used HuggingChat's ML Intern mode to build 7 models in just a few days, including a 0.8B prompt rewriter that runs on CPUs and produces 99.7% valid outputs, and a fine-tuned Qwen3.5-2B model that improved citrus disease identification accuracy from 14.9% to 52.8%. Total compute costs were approximately $103.

Why it's recommended:The author summarizes prompt-writing techniques for budget controls, baseline evaluations, and smoke tests, enabling an agent to independently handle data construction, training, and release. The methods can be readily adapted and reused.

Arena.ai@arenaModelsAI Score71

Arena Publishes Real-World Evaluation Results for Claude Haiku 5.5: Debuts at No. 30 in Code Arena with 1587 Points

Arena has published real-world evaluation results for Anthropic Claude Haiku 5.5 (High), which debuted at No. 30 in Code Arena: WebDev with a score of 1587, slightly outside the Pareto frontier.

Why it's recommended:The leaderboard provider reports Haiku 5.5's actual Code Arena score and the cost and score differences across Claude model tiers, making it easier to compare value for money.

Anthropic:NewsroomIndustryAI Score63

Anthropic launches Cyber Mission, introducing the Critical Infrastructure Defense Program and OSS Scanner

Anthropic launched Anthropic Cyber Mission, providing long-term support to defenders of critical infrastructure and open-source software. Its initial offerings include the Critical Infrastructure Defense Program (CIDP) and a free OSS Scanner.

Why it's recommended:The original article details how CIDP and OSS Scanner work, their partners, and their known limitations, giving readers an overview of how Anthropic uses models to support critical infrastructure and open-source defense.

Anthropic:ResearchProductsAI Score69

Anthropic launches OSS Scanner, an open-source vulnerability-scanning service offered free of charge to open-source projects

Anthropic launched OSS Scanner, an opt-in open-source vulnerability scanning service that uses its most capable models, including Claude Mythos, to regularly scan open-source projects for free. Its reports are generated entirely by models, without human review.

Why it's recommended:The original article provides scanning accuracy, manual verification data, and maintainer feedback, allowing readers to assess the actual signal quality of and ways to participate in this free, open-source security scanning service.

Reported by 6 additional sources
Comfy BlogTutorialsAI Score66

How to Generate Real-Time Video with MiniMax H3 on a Single RTX 5090

The ComfyUI blog author used an agent to gather various optimizations and get MiniMax H3 real-time video generation running on a single RTX 5090 (32 GB of VRAM). Generating a 30-second video took 23.51 seconds, achieving 1.28x real-time throughput.

Why it's recommended:The article provides a complete optimization checklist and open-source nodes for real-time livestream video generation on a single RTX 5090, so readers can reproduce this cost-cutting approach.

Google Developers BlogProductsAI Score70

Google open-sources ML Drift, an on-device GPU inference engine succeeding the TFLite GPU delegate

The Google AI Edge team has open-sourced ML Drift under the Apache 2.0 license, a cross-platform on-device GPU inference engine that serves as LiteRT’s core GPU acceleration layer.

Why it's recommended:The original article provides the architectural upgrades, performance data, and production deployment cases of the cross-platform GPU inference engine, enabling on-device AI developers to assess migration paths.

NVIDIA Technical Blog:Agentic AI / Generative AITutorialsAI Score62

NVIDIA KGMON Team Shares Lessons from Building a Data Analysis Agent for KDD Cup 2026

The NVIDIA KGMON team placed second in the KDD Cup 2026 Data Agents competition and published a retrospective on methods for building reliable data analysis agents.

Why it's recommended:The NVIDIA KGMON team reviews nine practices from the KDD Cup 2026 data agent competition and shares transferable methods for building a verifiable agent harness around a fixed small model.

The Decoder:AI NewsIndustryAI Score81

Some mathematicians call for a boycott of OpenAI as AI-generated mathematical proofs flood the field

The AHM organization, under Terence Tao's leadership, issued a statement calling for a boycott of OpenAI, saying its release of more than 700 AI-generated proof files at once was a show of force rather than scholarly work. OpenAI had previously claimed that an internal model solved over 100 open mathematics problems in one month, with a success rate of approximately 5% across roughly 8000 test problems and an average of 3 hours of GPT-Pro-level compute per problem.

Why it's recommended:This material brings together the mathematics community’s protest statement against OpenAI’s mass release of AI proofs and responses from Tao, Aaronson, and others, making it easier to understand the full controversy and the community’s divisions.

Google Developers BlogProductsAI Score66

Google open-sources AQuA, an environment quality agent that automatically diagnoses agent failures in production

Google releases and open-sources AQuA (Ambient Quality Agent), which periodically extracts production sessions from Cloud Trace, Cloud Logging, or BigQuery in Google Cloud projects and diagnoses agent failures through a five-stage pipeline of sampling, review, clustering, validation, and tracking.

Why it's recommended:The original article presents the five-stage pipeline, validation mechanisms, and measured remediation data, allowing readers to understand how agent quality monitoring can be implemented in production environments.

IT Home · Artificial IntelligenceIndustryAI Score78

OpenAI's annualized revenue reportedly approaches $50 billion, roughly $20 billion short of earlier expectations

According to the Financial Times, OpenAI disclosed to investors that its annualized revenue was approaching $50 billion as of the end of September, far below the previously estimated $70 billion, leaving a gap of about $20 billion. The difference stems from different accounting methods: Anthropic includes sales revenue from partners such as AWS and Google Cloud, while OpenAI excludes that portion. After the report was published, US technology stocks fell: the Nasdaq 100 closed down 1.4%, Nvidia fell 2.9%, and Oracle fell 5.5%.

Why it's recommended:The report reveals two key facts—the revenue shortfall and differences in statistical definitions—helping readers understand the context behind comparisons of AI revenue data and market fluctuations.

Also reported by 2 other sources
Anthropic:NewsroomIndustryAI Score61

Anthropic released its 2026 Usage Policy update, effective November 12

Anthropic released an updated usage policy, effective November 12, with most changes clarifying existing rules. Updates include a new section prohibiting deceptive activities, narrower election provisions, explicit bans on weapon software and arming drones, more precise restrictions on monitoring and enforcement, additional high-risk use cases and human-in-the-loop requirements for autonomous physical operations, and a new prohibition on sustained, gratuitous abusive behavior toward models.

Why it's recommended:The official explanation details the actual changes in the revised usage policy point by point, allowing readers to understand how Claude’s boundaries have shifted in scenarios involving elections, weapons, surveillance, and more.

Reported by 6 additional sources
Artificial Analysis@ArtificialAnlysModelsAI Score75

Artificial Analysis evaluates Google Nano Banana 2.1, ranking it fourth on both leaderboards at half the price of its predecessor

Artificial Analysis evaluated Nano Banana 2.1, the image model Google released on October 6, ranking it 4th on both the AA-Image-T2I v2.0 and AA-Image-Editing v2.0 leaderboards.

Why it's recommended:The original article provides Nano Banana 2.1's specific rankings, pricing, and improvements on two image evaluation leaderboards, allowing readers to compare its value for money with other models.

LangChain:BlogTutorialsAI Score61

LangChain Builds Restock, a Payment-Capable Agent, with Stripe Link and Managed Deep Agents

LangChain released Restock, a sample project featuring an office-supplies purchasing agent that runs in Slack using Managed Deep Agents. It demonstrates how an agent can complete payments securely.

Why it's recommended:The article provides a complete architecture for agent payments and credential isolation practices that readers can reuse for any agent scenario involving spending money.

Load more

Finished reading the news? Let’s talk about your business.

Leave your contact details, and I’ll get in touch with you personally.

WECHAT

Add me on WeChat

Press and hold or select the WeChat ID below to copy it.