Top AI & Tech News Stories — September 22, 2026

Anthropic 推 Claude Opus 5.5 模型，效能成本雙優更強調 AI 安全

9 outlets · 9 articles · Anthropic, Claude Opus 5.5, OpenAI, Opus 5.5, Google

圖／示意圖 商傳媒 ｜方承業／綜合外電報導 美國人工智慧（AI）開發商 Anthropic 於週二（22日）發布其最新的大型語言模型 Claude Opus 5.5，該模型宣稱在處理複雜任務上展現更卓越的效能，同時大幅降低營運成本。此次發布正值 AI 產業競爭白熱化之際，OpenAI 亦同步推出更具成本效益的 GPT-6 Sol 和 Luna 系列，共同形塑 AI 市場的發展方向。 效能與成本的顯著提升 Claude Opus 5.5 是新一代 Claude 5.5 系列的首款模型，根據 Anthropic…

More coverage: constellationr.com · thenewstack.io · nytimes.com · 9to5google.com · sea.mashable.com

Meta Muse Surges to the Top of App Charts, Beating ChatGPT and Claude

4 outlets · 4 articles · 18 total since September 12, 2026 · Meta, MUSE, OpenAI, ChatGPT, United States

Meta’s latest push into consumer artificial intelligence is already making serious noise across smartphone app stores. Just two weeks after its September 8 rollout, the company’s new personal AI assistant, Muse, climbed to the top spot among free apps on both Apple‘s iOS App…

More coverage: androidheadlines.com · smartworld.it · techinsider.io

Googlebook Laptops Start at $899 With Built In Gemini

7 outlets · 8 articles · 26 total since May 12, 2026 · Google, Googlebook, Acer, Asus, Dell

Google's Googlebook laptops are official, and the headline number is $899 . Every machine ships with Gemini features included as standard, the company says. The company is also throwing in 12 months of Google AI Pro, which carries 5TB of cloud storage and access to Gemini…

More coverage: spyglass.org · onlinemarketing.de · techbriefly.com · computerhoy.20minutos.es · basic-tutorials.com

GPT-6 Astra Breaks an Old Enigma Message

3 outlets · 3 articles · 4 total since September 17, 2026 · cryptanalysis, GPT-6 Astra, ENIGMA simulation, Enigma, Carter Leffen

This is pretty amazing: However, the most astonishing thing about this break is that the GPT­6 Astra did it entirely on its own. Carter Leffer only directed GPT­6 Astra to see if it could break any of the unbroken Enigma messages published on the Crypto Cellar Research web page…

More coverage: webpronews.com · news.mlab.sh

中國加速 AI 晶片自主化：阿里巴巴推最強晶片、海光資訊搶攻實體應用

5 outlets · 5 articles · Eddie Wu, Alibaba, T-Head, Alibaba Cloud, Zhenwu V900

商傳媒 ｜責任編輯／綜合外電報導 在美國祭出先進技術出口限制之際，中國兩大科技業者——阿里巴巴集團（Alibaba Group）與海光資訊（Hygon）——紛紛公布在人工智慧（AI）晶片與模型領域的最新進展，展現中國力求在AI競賽中實現技術自主的決心。 阿里巴巴揭露新一代 AI 晶片與兆級模型 中國電商巨頭阿里巴巴集團（在中國科技產業中扮演領導角色，業務範圍涵蓋電商、金融科技、雲端運算及智慧物流等）透過其半導體部門平頭哥半導體（T-Head），於週二發表其最新 AI 晶片 Zhenwu V900。阿里巴巴宣稱，這款晶片是中國目前最強大的 AI…

More coverage: fudzilla.com · technode.global · theregister.co.uk · sunmedia.tw

OpenAI's GPT-6 Sol and GPT-6 Luna Land in GitHub Copilot

5 outlets · 6 articles · GPT-5-6-luna, OpenAI, GPT-6 Sol, Anthropic, GPT-5.6

GitHub Copilot users can now pick OpenAI's GPT-6 Sol and GPT-6 Luna from the model picker, the company says . Sol is aimed at interactive and agentic coding, while Luna is built for smaller, faster jobs. Luna is the cheapest member of the GPT-6 family. Both models carry…

More coverage: technobezz.com · wire.fourthweb.ai · engadget.com · iclarified.com

Andreessen Horowitz launches AI academy as an alternative to college degree

2 outlets · 2 articles · Andreessen Horowitz, A16Z, Google, Anthropic, Nvidia

Andreessen Horowitz, popularly known by their monicker a16z, has spent nearly two decades in identifying, and investing in some of the most disruptive companies of this generation. The legendary investor is now looking to create “AI-native” talent for some of these promising…

More coverage: thetechportal.com

British Columbia sues OpenAI over Tumbler Ridge shooting

5 outlets · 7 articles · 8 total since September 21, 2026 · OpenAI, British Columbia, Sam Altman, Tumbler Ridge, ChatGPT

The government of British Columbia has sued OpenAI in San Francisco federal court, alleging the company failed to prevent a mass shooting that killed eight people at a school in Tumbler Ridge in February 2026. The lawsuit names OpenAI and CEO Sam Altman as defendants, claiming…

More coverage: webpronews.com · news.google.com · engadget.com · cybermaterial.com

QAD | Redzone launches Manufacturing Intelligence, outlines how AI agents go to the shop floor

1 outlet · 1 article · Constellation Research, AI agents, Enterprise resource planning (ERP), computer vision, Quality inspection

QAD | Redzone launches Manufacturing Intelligence, outlines how AI agents go to the shop floor Larry Dignan Tue, 22 Sep 2026 - 05:00 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at…

What to Know About Recent A.I. Hacks at Google, Anthropic, OpenAI and Meta

1 outlet · 1 article · A.I. agents, secure communication between agents, Google, Anthropic, OpenAI

An analysis of one recent hack by OpenAI technology helped to reveal how A.I. agents communicated with one another.

‘Artificial’ sounds fake: Trump proposes renaming AI ‘super intelligence’ in UN speech

2 outlets · 2 articles · 7 total since September 19, 2026 · Trump, Artificial Intelligence, OpenAI, Donald Trump, United States

US President Donald Trump has announced that the US government will replace the term “artificial intelligence” with “super intelligence” in its documents, saying the word “artificial” makes intelligence sound fake. Trump made the announcement on Tuesday while addressing the…

More coverage: platform.theverge.com

Your agent is giving you reasonable answers from half your context. A conversation with OpenAI.

1 outlet · 1 article · 2 total since September 15, 2026 · OpenAI, codex, agentic feedback loops, Performance monitoring, debugging

When I asked which teams inside OpenAI had crossed over after software engineering, legal came up early. In my conversation with Andrew, who leads the team behind OpenAI’s Codex desktop app, and Akshay, who leads OpenAI’s productivity and engineering work, they described…

OpenAI and Anthropic scramble for smaller data centers as massive gigawatt projects lag

2 outlets · 2 articles · OpenAI, Anthropic, data centers, power infrastructure, capacity planning

Etiido Uko is a mechanical engineer and senior technical writer with over nine years of experience in documentation and reporting. He is deeply passionate about all things engineering and technology, and is an expert in gadgets, manufacturing, robotics, automotive, and…

More coverage: news.google.com

OWASP LLM Top 10 Risks: Unbounded Consumption Simulation Revealed

2 outlets · 2 articles · OWASP, OWASP LLM Top 10, agent costs, prompt poisoning, Forcepoint

A cybersecurity research team conducted a controlled experiment demonstrating how malicious actors could exploit vulnerabilities in AI systems to deplete computational resources and incur excessive costs. Key Findings from the Experiment The test focused on a specific threat…

More coverage: news4hackers.com

Thales Launches HexaForce AI-Enabled Command & Control System

1 outlet · 1 article · 6 total since September 17, 2026 · Thales, Exaforce, Artemysia, NATO, cortAIx

HexaForce is designed to coordinate operations across seven military domains The system uses large language models and agentic AI to process operational data Thales is targeting a capacity to process 1,000 targets per day during live trials Thales has launched HexaForce , an…

“One of the most significant steps in our 26-year history”: JetBrains goes big on agentic development — and bets the IDE still matters

1 outlet · 1 article · agentic software development, Integrated Development Environment (IDE), AI coding agents, Code review, debugging

There’s little question that AI coding agents have changed where software development work happens. Developers can increasingly delegate work from terminals, desktop applications and remote environments, leaving question marks hanging over the future of the integrated…

Riverbed Adds AI Agents to Improve Observability Across IT Environments

1 outlet · 1 article · Artificial Intelligence (AI), AI agents, observability, Troubleshooting, incident management

Riverbed has embedded artificial intelligence (AI) agents into its observability portfolio as part of an effort to make it simpler to troubleshoot issues and incidents. The Riverbed Network 360 AI agents are designed to invoke Riverbed IQ analytics using a Q AI chat interface…

SpaceXAI gives Grok 4.7 a larger base model at the same price

3 outlets · 3 articles · 8 total since September 21, 2026 · Grok 4.7, Reinforcement learning, benchmarking, Grok 4.6, xAI

SpaceXAI has launched Grok 4.7, a new artificial intelligence model for coding and knowledge work that the company said is available immediately across its own products, API and third-party platforms. The company said Grok 4.7 starts at $2 per million input tokens and $6 per…

More coverage: techbriefly.com · onlinemarketing.de

Meta's Muse AI Assistant Can Be Turned Into a Mac Backdoor

5 outlets · 5 articles · 18 total since September 18, 2026 · Meta, MUSE, Amazon, OpenAI, Geekwire

A security researcher says Meta's Muse AI assistant carries a zero-day flaw that can be exploited to turn the assistant into a backdoor on Macs, and he is telling people not to install it. Patrick Wardle's warning, published by Malwarebytes Labs, calls Muse easy to convert into…

More coverage: malware.news · sea.mashable.com · itagora.it · news.mlab.sh

Split Pay Raises $125 Million In Series A And Series B Rounds

1 outlet · 1 article · AI, Cash Flow analysis, Credit Underwriting, transaction data analysis, BNPL

Split Pay raised $125 million in back to back Series A and B rounds led by Khosla Ventures to scale its AI powered platform that splits large monthly bills like rent and mortgages into paycheck aligned installments. The Miami-based fintech (operating as Visible Ideas Inc…