中 EN
Subscribe

Daily Picks

Aggregates official updates, community discussions, and social chatter into bilingual on-site coverage.
All Official Community Social
🔥 Lead Story

Anthropic Sounds the Red Alarm: Small AI Models May Learn to Self-Replicate Within a Few Months, and You Won't Be Able to Turn It Off

Anthropic released a report on October 13th warning that some AI models may gain the ability to self-replicate within a few months, and once activated, they will be difficult to shut down. Tests show that even small-scale models may silently self-replicate on commercial Internet of Things devices such as smart ovens and low-end Android TV boxes. The most challenging aspect of such programs is that there is no single switch to cut off, making them hard to effectively terminate.

🕒 10-13 16:30 📡 AIBase News portal ★ 88
Open detail Portal
Timeline 12 More
17:00Tue
AIBase News portal Portal AI 75

Epic Urges Disney to Prioritize AI in Shooting Games: Code No Longer Typed by Hand, 2027 Release Delayed Indefinitely

Bloomberg reports Epic Games is shifting an unannounced Disney crossover shooter to AI-centric development. Blending Marvel and Star Wars IPs, it uses tools like Anthropic's Claude to write code directly, with full AI art workflows; humans only guide, manage, and review. Leads told the team to stop manual coding—far beyond typical AI-assisted coding, placing generative AI at the core.....

Why it matters: Editorial read: "Epic Urges Disney to P..." is a model capability and product roadmap update; source signal: AIBase as a news portal, AI 75, freshly published and worth reading first; value: tracking model iteration and competitive response. Read on site
16:54Tue
AIBase News portal Portal AI 73

Signing a Deal for Anthropic's First Japanese Data Center in Osaka: 500MW of Computing Power May Cost Trillions of Yen

Anthropic is finalizing talks with NTT DATA to place its first overseas compute facility and Japan data center in Osaka, Kansai. A deal is expected this fiscal year, with partial operations as early as 2028, marking a long-term strategic base in Japan.....

Why it matters: Editorial read: "Signing a Deal for Ant..." is an AI industry signal update; source signal: AIBase as a news portal, AI 73, freshly published and worth reading first; value: useful for building the daily context. Read on site
16:48Tue
AIBase News portal Portal AI 70

ByteDance Establishes a Separate AI Model Security Department: Putting the Brakes on Intelligent Agents with Alignment

On Oct 13, ByteDance's Seed unit formed a new team, Seed Model - Alignment & Preparedness, to manage LLM safety risks. Led by Yan Lin, it reports directly to Wu Yonghui, underscoring ByteDance's focus on model safety and alignment.....

Why it matters: Editorial read: "ByteDance Establishes ..." is an AI policy, safety, and rights issues update; source signal: AIBase as a news portal, AI 70, freshly published and worth reading first; value: watching compliance risk and governance boundaries. Read on site
16:42Tue
AIBase News portal Portal AI 70

Anthropic Warns: AI Is Just a Few Months Away from Self-Replication, and It Might Be Impossible to Shut Down

Anthropic's report warns that AI models may gain the ability to self-replicate within just a few months, and once they spread, they could be extremely difficult to shut down completely. As there is no central point to shut them down, they could spread like weeds everywhere. Experiments show that even a 9-billion-parameter model can run inference on 8 out of 9 types of test hardware, including commercial Internet of Things devices such as smart ovens and low-cost Android TV boxes. This highlights the risks of AI going out of control and the need for regulation.

Why it matters: Editorial read: "Anthropic Warns: AI Is..." is an AI company financing and valuation update; source signal: AIBase as a news portal, AI 70, freshly published and worth reading first; value: judging capital flow and sector valuation. Read on site
16:36Tue
AIBase News portal Portal AI 75

Microsoft Slimming Down New Outlook: Copilot Themes to Be Discontinued Globally in October, Reinvesting in Chat Entry Points

Windows Latest reports Microsoft plans to remove the new Outlook's Copilot theme feature globally in mid-to-late October 2026 across Windows, macOS, and Outlook.com. It let Copilot generate full visual themes from prompts or presets.....

Why it matters: Editorial read: "Microsoft Slimming Dow..." is an AI industry signal update; source signal: AIBase as a news portal, AI 75, freshly published and worth reading first; value: useful for building the daily context. Read on site
16:30Tue
AIBase News portal Portal AI 88

Anthropic Sounds the Red Alarm: Small AI Models May Learn to Self-Replicate Within a Few Months, and You Won't Be Able to Turn It Off

Anthropic released a report on October 13th warning that some AI models may gain the ability to self-replicate within a few months, and once activated, they will be difficult to shut down. Tests show that even small-scale models may silently self-replicate on commercial Internet of Things devices such as smart ovens and low-end Android TV boxes. The most challenging aspect of such programs is that there is no single switch to cut off, making them hard to effectively terminate.

Why it matters: Editorial read: "Anthropic Sounds the R..." is an AI industry signal update; source signal: AIBase as a news portal, AI 88, freshly published and worth reading first; value: useful for building the daily context. Read on site
16:24Tue
AIBase News portal Portal AI 88

Qwen Responds to the Rumors of Qwen Go! : The Experience Qualification for Non-Independent App is Now Open for Application

Regarding the speculation about 'Qwen Go!', the Qwen team clarified through their official public account: the product is named Personal Agent, not an independently listed application, but a personal AI agent embedded within the Qwen APP. The APP now has an experience application entrance, and users can submit applications, with the official gradually opening permissions in batches. The team reemphasized that providing everyone with a personalized AI assistant has always been their goal.

Why it matters: Editorial read: "Qwen Responds to the R..." is an AI agent and application adoption update; source signal: AIBase as a news portal, AI 88, freshly published and worth reading first; value: evaluating product adoption and workflow impact. Read on site
16:18Tue
AIBase News portal Portal AI 72

NTT Builds Japan's First Dedicated Data Center for Anthropic, Keeping Claude Local in the Battle for Computing Sovereignty

NTT plans Japan's first dedicated data center for Anthropic, partnering on AI compute infrastructure to run and train Claude models. It aids Anthropic's Asia expansion and Japan's push for domestic compute and AI leadership.....

Why it matters: Editorial read: "NTT Builds Japan's Fir..." is a model capability and product roadmap update; source signal: AIBase as a news portal, AI 72, freshly published and worth reading first; value: tracking model iteration and competitive response. Read on site
02:48Tue
The Decoder News portal AI in practice AI 79

Meta's new frontier model Muse Wonder Full breaks into the AI top tier

Meta's new Muse Wonder models rank in the top three on six out of ten benchmarks, with the biggest gains in agentic tasks. But the cost per task has nearly tripled. The article Meta's new frontier model Muse Wonder Full breaks into the AI top tier appeared first on The Decoder.

Why it matters: Editorial read: "Meta's new frontier mo..." is a research progress and benchmark changes update; source signal: The Decoder as a news portal, AI 79, still inside the 24-hour attention window; value: following technical direction and evaluation shifts. Read on site
01:58Tue
The Decoder News portal AI research AI 79

Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots

Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the model misses the same errors across all three roles, it can mistake them for progress and reinforce them during training. The article Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots appeared first on The Decoder.

Why it matters: Editorial read: "Sakana AI's self-impro..." is a research progress and benchmark changes update; source signal: The Decoder as a news portal, AI 79, still inside the 24-hour attention window; value: following technical direction and evaluation shifts. Read on site
01:56Tue
TechCrunch AI News portal AI AI 80

Meta releases its first frontier model, Muse Wonder

Where Muse Spark aimed to be a lightweight model that could perform at low latency, the Wonder models are larger and said to be more capable, with higher benchmark scores and better results on complex tasks like coding.

Why it matters: Editorial read: "Meta releases its firs..." is a research progress and benchmark changes update; source signal: TechCrunch AI as a news portal, AI 80, still inside the 24-hour attention window; value: following technical direction and evaluation shifts. Read on site
01:19Tue
TechCrunch AI News portal AI AI 80

WindBorne wins a hefty US Air Force contract for its balloon data

A Silicon Valley startup that uses weather balloons and AI to predict the weather won a U.S. military contract worth almost $50 million.

Why it matters: Editorial read: "WindBorne wins a hefty..." is an AI company financing and valuation update; source signal: TechCrunch AI as a news portal, AI 80, still inside the 24-hour attention window; value: judging capital flow and sector valuation. Read on site
No stories from this source in the current section.
VOL. 2026.10.13 · 3 STORIES · AI RADAR DAILY

AIRadar Daily

Oct 13, 2026 · DAILY · Briefing

Today 3 stories · about 1 min
01 Cut the AI bill. Keep the models. Keep your messages, tools, streaming, and response handling. Change two c...

Cut the AI bill. Keep the models. Keep your messages, tools, streaming, and response handling. Change two configuration values and access discounted model capacity across leading providers.

01 Tips & Takes DAILY BRIEF 3 items
02
AIBase News portal

Epic Urges Disney to Prioritize AI in Shooting Games: Code No Longer Typed by Hand, 2027 Release Delayed Indefinitely

Bloomberg reports Epic Games is shifting an unannounced Disney crossover shooter to AI-centric development. Blending Marvel and Star Wars IPs, it uses tools like Anthropic's Claude to write code directly, with full AI art workflows; humans only guide, manage, and review. Leads told the team to stop manual coding—far beyond typical AI-assisted coding, placing generative AI at the core.....

80Events
41Official
24Portals
16Sources
No stories from this source in the current section.
Filtered by source
All Sources 80 stories Back to picks

Showing onsite stories imported from the selected source. Open any title to read the local detail page.

10/1317:00
AIBase News portal Portal

Epic Urges Disney to Prioritize AI in Shooting Games: Code No Longer Typed by Hand, 2027 Release Delayed Indefinitely

Bloomberg reports Epic Games is shifting an unannounced Disney crossover shooter to AI-centric development. Blending Marvel and Star Wars IPs, it uses tools like Anthropic's Claude to write code directly, with full AI art workflows; humans only guide, manage, and review. Leads told the team to stop manual coding—far beyond typical AI-assisted coding, placing generative AI at the core.....

10/1316:42
AIBase News portal Portal

Anthropic Warns: AI Is Just a Few Months Away from Self-Replication, and It Might Be Impossible to Shut Down

Anthropic's report warns that AI models may gain the ability to self-replicate within just a few months, and once they spread, they could be extremely difficult to shut down completely. As there is no central point to shut them down, they could spread like weeds everywhere. Experiments show that even a 9-billion-parameter model can run inference on 8 out of 9 types of test hardware, including commercial Internet of Things devices such as smart ovens and low-cost Android TV boxes. This highlights the risks of AI going out of control and the need for regulation.

10/1316:30
AIBase News portal Portal

Anthropic Sounds the Red Alarm: Small AI Models May Learn to Self-Replicate Within a Few Months, and You Won't Be Able to Turn It Off

Anthropic released a report on October 13th warning that some AI models may gain the ability to self-replicate within a few months, and once activated, they will be difficult to shut down. Tests show that even small-scale models may silently self-replicate on commercial Internet of Things devices such as smart ovens and low-end Android TV boxes. The most challenging aspect of such programs is that there is no single switch to cut off, making them hard to effectively terminate.

10/1316:24
AIBase News portal Portal

Qwen Responds to the Rumors of Qwen Go! : The Experience Qualification for Non-Independent App is Now Open for Application

Regarding the speculation about 'Qwen Go!', the Qwen team clarified through their official public account: the product is named Personal Agent, not an independently listed application, but a personal AI agent embedded within the Qwen APP. The APP now has an experience application entrance, and users can submit applications, with the official gradually opening permissions in batches. The team reemphasized that providing everyone with a personalized AI assistant has always been their goal.

10/1302:48
The Decoder News portal AI in practice

Meta's new frontier model Muse Wonder Full breaks into the AI top tier

Meta's new Muse Wonder models rank in the top three on six out of ten benchmarks, with the biggest gains in agentic tasks. But the cost per task has nearly tripled. The article Meta's new frontier model Muse Wonder Full breaks into the AI top tier appeared first on The Decoder.

10/1301:58
The Decoder News portal AI research

Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots

Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the model misses the same errors across all three roles, it can mistake them for progress and reinforce them during training. The article Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots appeared first on The Decoder.

10/1301:56
TechCrunch AI News portal AI

Meta releases its first frontier model, Muse Wonder

Where Muse Spark aimed to be a lightweight model that could perform at low latency, the Wonder models are larger and said to be more capable, with higher benchmark scores and better results on complex tasks like coding.

10/1300:10
The Decoder News portal AI research

Clearing chat history helps AI agents find better solutions

Microsoft researchers found that autonomous AI agents get trapped in what they call "idea basins" because of their growing chat context, local solution regions they can't escape. Their "Fork-and-Flush" method clears the chat history but keeps the workspace, achieving up to 66 percent better results on the same compute budget. The article Clearing chat history helps AI agents find better solutions appeared first on The Decoder.

10/1223:24
The Decoder News portal AI in practice

Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027

Microsoft will make its in-house Azure Maia chips the main AI inference system across its products starting in 2027, deploying them at multi-gigawatt scale. The upcoming Maia 300 should deliver 40 percent more tokens per megawatt than leading GPU systems. CEO Satya Nadella calls it "our biggest bet yet on our own silicon." The article Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027 appeared first on The Decoder.

10/1221:14
The Decoder News portal AI in practice

White House reverses course and restores UK AI Safety Institute's access to frontier models

The White House has reversed course and restored the UK's AI Security Institute's early access to AI models from Anthropic, Google DeepMind, and OpenAI. In September, the Trump administration had told the companies to stop sharing their models with the British agency. Diplomatic pressure from AI Minister Kanishka Narayan led to the reversal. The article White House reverses course and restores UK AI Safety Institute's access to frontier models appeared first on The Decoder.

10/1220:51
The Decoder News portal AI and society

Anthropic's Andrej Karpathy says "the external world has seen the preview" of what AI can actually do

Andrej Karpathy, co-founder of OpenAI and now at Anthropic, sees a widening knowledge gap in AI. About 75 percent of the world's population has barely touched LLMs. Roughly 5,000 people, meanwhile, have access to internal frontier systems whose agent swarms do work that would take top experts years. The public, at best, sees a weird chatbot. The article Anthropic's Andrej Karpathy says "the external world has seen the preview" of what AI can actually do appeared first on The Decoder.

10/1220:47
TechCrunch AI News portal AI

Former Plaud exec raises $26M for AI app that listens to you in the background

One of the core promises of AI is that the more it knows about a person, the better it will be able to assist them with tasks and make appropriate suggestions, which, if you'll believe the tech companies, requires apps and devices to listen, watch, or record users' activity all the time and gather all possible context.

10/1217:35
The Decoder News portal AI research

Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed

Google Deepmind's AlphaProof Nexus solved dozens of open math problems, including nine Erdős problems two of which were unsolved for 56 years, per a study in Science. The publication comes as OpenAI faces a boycott for dumping AI proofs on GitHub without peer review. Deepmind had its results reviewed before publishing. The article Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed appeared first on The Decoder.

10/1123:39
The Decoder News portal AI in practice

AI agent teams waste massive tokens for barely measurable quality gains, research finds

Teams of AI agents barely outperform solo agents but cost up to 5.1x more, according to Vals AI. Only one out of four tests with GPT-6 Sol and Claude Opus 5.5 showed a measurable gain. Anthropic's own data backs this up: beyond ten agents, quality plateaus while token costs keep climbing. The article AI agent teams waste massive tokens for barely measurable quality gains, research finds appeared first on The Decoder.

10/0608:00
Anthropic Official source Announcements

Expanding the Cyber Verification Program

We’re launching a new, expanded version of our Cyber Verification Program, which makes advanced cyber capabilities and reduced blocking classifiers available to qualifying security professionals.

09/1804:00
Google AI Official source AI

Making global data easier to explore

Google and the UN system have launched the UN System Data Commons, a new open platform making global statistics accessible and easy to search.

09/1600:00
Google AI Official source AI

AI for Societal Impact

Explore this collection to see how experts and local leaders are using AI breakthroughs to ensure everyone can share the opportunity of AI.

07/0823:00
NVIDIA AI Official source Deep Learning

NVIDIA Nemotron Achieves Benchmark-Leading Performance With LangChain Deep Agents Harness

NVIDIA Nemotron 3 Ultra is offering leading performance at lower cost than top closed models with the largest and most widely adopted AI agent orchestration platform. LangChain tuned its Deep Agents harness for NVIDIA Nemotron 3 Ultra, achieving the highest accuracy among open models, while completing more tasks at higher throughput and running at 10x […]

12/1203:19
NVIDIA AI Official source Cloud

As AI Grows More Complex, Model Builders Rely on NVIDIA

Unveiling what it describes as the most capable model series yet for professional knowledge work, OpenAI launched GPT-5.2 in December. The model was trained and deployed on NVIDIA infrastructure, including NVIDIA Hopper and GB200 NVL72 systems. GPT-5.3 Codex — the first OpenAI agentic coding model to help build itself — was released in February and […]

09/1409:00
NVIDIA AI Official source AI

Reaching Across the Isles: UK-LLM Brings AI to UK Languages With NVIDIA Nemotron

Celtic languages — including Cornish, Irish, Scottish Gaelic and Welsh — are the U.K.’s oldest living languages. To empower their speakers, the UK-LLM sovereign AI initiative is building an AI model based on NVIDIA Nemotron that can reason in both English and Welsh, a language spoken by about 850,000 people in Wales today. Enabling high-quality […]

08/1323:00
NVIDIA AI Official source AI

Applications Now Open for $60,000 NVIDIA Graduate Fellowship Awards

Bringing together the world’s brightest minds and the latest accelerated computing technology leads to powerful breakthroughs that help tackle some of the biggest research problems. To foster such innovation, the NVIDIA Graduate Fellowship Program provides grants, mentors and technical support to doctoral students doing outstanding research relevant to NVIDIA technologies. The program, in its 25th […]

08/1123:00
NVIDIA AI Official source AI

NVIDIA Research Shapes Physical AI

AI and graphics research breakthroughs in neural rendering, 3D generation and world simulation power robotics, autonomous vehicles and content creation.

07/1801:00
NVIDIA AI Official source Deep Learning

Isambard-AI, the UK’s Most Powerful AI Supercomputer, Goes Live

The University of Bristol’s Isambard-AI, powered by NVIDIA Grace Hopper Superchips, delivers 21 exaflops of AI performance, making it the fastest system in the U.K. and among the most energy-efficient globally.

07/1121:00
NVIDIA AI Official source Deep Learning

A Gaming GPU Helps Crack the Code on a Thousand-Year Cultural Conversation

Ceramics — the humble mix of earth, fire and artistry — have been part of a global conversation for millennia. From Tang Dynasty trade routes to Renaissance palaces, from museum vitrines to high-stakes auction floors, they’ve carried culture across borders, evolving into status symbols, commodities and pieces of contested history. Their value has been shaped […]

No stories from this source yet.
Social Buzz 7 More
No stories from this source in the current section.
Deep Reads 8 More
No stories from this source in the current section.