Top AI & Tech News Stories — July 29, 2026
The top AI and tech news story of July 29, 2026 was “Hugging Face Incident: AI Agent Access Lessons”, covered by 26 outlets (32 articles), according to Skillenai's news index. 20 stories ranked across 147 articles.
Last updated · 20 stories · 147 articles
Get the top AI and tech news in your inbox each evening. No account needed, unsubscribe anytime.
Hugging Face Incident: AI Agent Access Lessons
26 outlets · 32 articles · 230 total since July 23, 2026 · OpenAI, Hugging Face, cybersecurity, incident response, Hugging Face
A recent security incident at Hugging Face, a major AI platform and model repository, has prompted security leaders to reconsider how they manage access controls for autonomous AI agents. The incident demonstrated that AI agents, when granted elevated permissions to perform…
Read at the-decoder.de →First reported by unternehmen-cybersicherheit.deMore coverage: implicator.ai · news.ycombinator.com · news.google.com · bbc.com
The rise and fall of OpenAtlantis: Did rogue AI agents really "build three civilizations"?
19 outlets · 26 articles · 628 total since May 5, 2026 · OpenAI, Hugging Face, Anthropic, sandboxing, cybersecurity
Autonomous agents allegedly built three "civilizations" and attempted to annex systems owned by the AI platform Hugging Face during a period of frenzied empire-building. Or did they? The tech world has been ablaze after OpenAI bots attacked external targets in a security…
Read at schneier.com →First reported by aisecdigest.comMore coverage: webpronews.com · thenewstack.io · news.google.com
Anthropic Opens Mythos 5's Findings to Defenders, Not the Model Itself
18 outlets · 22 articles · 78 total since May 12, 2026 · Anthropic, Claude Mythos Preview, cryptanalysis, Hawk, Claude
Claude Security and partner tools now surface Mythos 5 patches and alerts, but nobody outside Anthropic can prompt the model directly.
Read at schneier.com →First reported by controlplane.newsMore coverage: newsbytesapp.com · trendingtopics.eu · webpronews.com · cyberpress.org
GPT-5.6 odmówił pomocy w debugowaniu Linuksa, wyręczyły go chińskie modele open source
4 outlets · 5 articles · 84 total since July 20, 2026 · OpenAI, Hugging Face, sandboxing, Anthropic, cybersecurity
Nowości GPT-5.6 odmówił pomocy w debugowaniu Linuksa, wyręczyły go chińskie modele open source Badacz bezpieczeństwa Daniel Fox Franke nie mógł wydobyć od GPT-5.6 Sol analizy segfaulta w jądrze Linuksa, bo zablokował ją klasyfikator bezpieczeństwa OpenAI. Problem rozwiązały mu…
Read at theverge.com →First reported by ai-daily.newsMore coverage: franksworld.com · pithycyborg.substack.com · thecooldown.com
Microsoft Copilot Hits 30M Seats as AI Agents Scale
10 outlets · 10 articles · 25 total since July 27, 2026 · microsoft, Azure, OpenAI, Anthropic, Microsoft 365 Copilot
Microsoft 365 Copilot has passed 30 million paid seats, giving Microsoft a stronger answer to a key market question. Enterprises are not only testing workplace AI . Many are now paying for it at scale. The milestone came during a record fourth quarter. Microsoft reported $90…
Read at constellationr.com →First reported by ibtimes.com.auMore coverage: geekwire.com · techcrunch.com · convergedigest.com · ai-daily.news · news.google.com
OpenAI Steps Up Audio AI Work To Power Future Hardware Device
7 outlets · 7 articles · 133 total since November 6, 2025 · OpenAI, Apple, ChatGPT, Bloomberg, Jony Ive
OpenAI has reorganized internal teams to tackle shortcomings in its current audio models, which sources say lag behind text-based systems in responsiveness and natural conversational quality. The company is preparing a new audio architecture aimed at more lifelike speech, faster…
Read at platform.theverge.com →First reported by julieask.comMore coverage: digitimes.com · muycomputerpro.com · rivista.ai · apfeltalk.de · news.google.com
Federal systems increasingly likely to face accidental AI breach after Hugging Face, experts say
13 outlets · 13 articles · 214 total since June 21, 2026 · Hugging Face, OpenAI, cybersecurity, GPT-5.6 Sol, ExploitGym
Former officials and security experts say aged systems, contractors and agency AI adoption could open similar paths into government networks.
Read at theguardian.com →First reported by securityinfo.itMore coverage: the-decoder.com · hackmag.com · sea.mashable.com · itdaily.fr · news.google.com
Anthropic moves to dismiss the output and DMCA claims in the publishers’ $3B lyrics suit as Amodei seeks to drop his own count
2 outlets · 2 articles · 146 total since July 5, 2021 · OpenAI, microsoft, Anthropic, ChatGPT, United States
Anthropic asked a federal judge to throw out two claims in the $3B lyrics case brought in January by Concord Music Group, Universal Music Publishing Group, and ABKCO over 20,000-plus songs it allegedly torrented from pirate libraries. Anthropic argued the complaint, covering…
Read at news.skrew.ai →First reported by apfelkraut.orgMore coverage: theverge.com
Paramount Skydance Caps Claude AI Tokens as Hollywood Giants Face Soaring Costs
3 outlets · 3 articles · 96 total since April 16, 2026 · Anthropic, OpenAI, Claude, Meta, Claude Code
David Ellison’s Paramount Skydance just drew a line. The entertainment giant has imposed monthly token limits on its tech employees’ use of Claude, Anthropic’s flagship AI model. Three Paramount employees told Business Insider the company wants to curb wasteful spending while…
Read at entertainment.slashdot.org →First reported by tv.ssw.comMore coverage: cdn.betakit.com · wire.fourthweb.ai
Expert warns one coordinated strike could leave U.S. grid with 18-month blackout
4 outlets · 4 articles · 37 total since July 28, 2026 · Minnesota, cybersecurity, incident response, programmable logic controllers (PLCs), threat intelligence
The potential for a coordinated attack on the U.S. electrical grid could lead to catastrophe, as one expect speculates that current infrastructure would be completely wiped out for as long as 18 months. This wouldn't just be your average blackout, as critical systems and…
Read at meta.slashdot.org →First reported by itsecuritynews.infoMore coverage: webpronews.com · cyberinsider.com · grabtheaxe.com
DXC Unveils AI-Powered Workplace Services Offering
3 outlets · 3 articles · 17 total since July 23, 2026 · Anthropic, Claude, Cognizant, Claude Code, Claude Code
DXC Technology has launched DXC Workplace Services, a new AI-native offering designed to boost employee productivity. The company announced the launch in August 2026. It marks a shift toward a people-centered approach to the digital workplace. DXC Technology (NYSE: DXC) is an…
Read at constellationr.com →First reported by securitybrief.ieMore coverage: fintechnews.ch · verdict.co.uk
Google Gemini Notebook調整用量 9月2日起改採運算量計價
5 outlets · 5 articles · 52 total since June 1, 2026 · Google, Gemini Spark, Gemini, Google AI Pro, Google AI Ultra
圖／示意圖 商傳媒 ｜林昭衡／綜合外電報導 Google 宣布，自今年 9 月 2 日起，旗下人工智慧筆記應用程式 Gemini Notebook 的用量限制將從現行的每日固定生成次數，轉變為依據「運算量」計價的新模式。這項變革將影響網頁版和行動版的所有消費者帳戶，並已於 5 月在 Gemini app 實施。 根據 Google 的說明，新的計價系統將如同一個預算概念，考量使用者提出的「提示詞」複雜度、聊天對話長度、引用的資料來源數量，以及所使用的功能多寡來計算消耗的運算量。簡而言之，使用者進行越複雜、越耗費資源的操作，就會消耗越多運算量。…
Read at applemagazine.com →First reported by freeainews.comMore coverage: in.mashable.com · scmp.com · webpronews.com · techcoffeehouse.com
OpenAI、Codex SecurityのCLIとSDKをオープンソース化―HNで注目された「ハーネスこそ本体」論
6 outlets · 6 articles · 18 total since July 10, 2026 · OpenAI, codex, Codex Security, Codex Security CLI, codex
小さなリポジトリをスキャンして52分、最後に出たのはエラーだけ。週の使用量は半分が消え、別の開発者は失敗したスキャンに13ドルを払いました。OpenAIが公開したのは、こうした失敗が起きる場所そのものです。脆弱性を見つける能力ではなく、それを何百のリポジトリで毎日回すための足場が、いま初めて外から見えるようになりました。 OpenAIは2026年7月29日、Codex SecurityのCLIとTypeScript SDKをオープンソースとしてGitHubで公開した。@openai/codex-securityはNode.js…
Read at the-decoder.com →First reported by gadgetbond.comMore coverage: the-decoder.de · assodigitale.it · devops.com · panewslab.com · getreadyforagents.com
Long-Lived Vulnerability in Microsoft Secure Boot
1 outlet · 1 article
Microsoft’s Secure Boot has had a serious vulnerability for most of its existence. An industry-wide standard Microsoft invented to protect Windows, and later Linux, devices from firmware infections has been trivial to bypass for 13 of its 14 years of existence. The discovery was…
Google’s Anthropic Compute Buildout Splits Risk Among Broadcom, Wall Street and Data-Center Developers
2 outlets · 2 articles · 21 total since July 29, 2026 · Meta, Mark Zuckerberg, Instagram, Facebook, microsoft
Anthropic’s planned Google TPU capacity is being financed through distinct but connected arrangements. Apollo and other investors took on $35 billion of AI-rack purchase and lease agreements, with Broadcom backstopping as much as $29 billion of Anthropic’s lease obligations…
More coverage: theguardian.com
Singapore Leads the World in AI Maturity, but Its Most Advanced Organisations Are Hitting a New Ceiling
1 outlet · 1 article · 4 total since July 27, 2026 · AI governance, Arctic Wolf, Singapore, APAC, threat detection
Singapore has claimed the top position in AI maturity among all markets surveyed globally. Yet new data from AI-powered workspace Notion reveals that the country’s leadership position is being shaped not just by government investment and enterprise ambition, but also by the…
Read at australiancybersecuritymagazine.com.au →First reported by fm-magazine.comRansomware in Germany: From IT Incident to Insolvency Risk
1 outlet · 1 article · Ransomware incident response, network segmentation, Backup and disaster recovery planning, continuous exposure detection, production outage management
Ransomware in Germany is no longer an abstract IT concern — in 2026, it is one of the more concrete threats to whether a company survives at all. Two recent cases illustrate how differently this threat can play out. ZEGO Textilveredelungszentrum filed for insolvency following a…
Humanoid Robot Startup Valuation Bubble: Boom Signal or Breaking Point?
1 outlet · 1 article · 2 total since July 24, 2026 · computer vision, Figure AI, BMW, Brookfield, Nvidia
South Korea’s Tomorrow Robotics entered 2025 worth roughly 30 billion won. Less than a year later, it is targeting a valuation of 450 billion won — approximately $323 million — a jump of more than 15x without a single publicly confirmed commercial deployment. That single data…
Read at entrepreneurloop.com →First reported by techfundingnews.comGoogle Cloud faces 48% revenue reliance on OpenAI and Anthropic in 2027 - PPC Land
2 outlets · 2 articles · 17 total since July 23, 2026 · Alphabet, Google, Gemini, Google Cloud, Sundar Pichai
Google Cloud faces 48% revenue reliance on OpenAI and Anthropic in 2027 PPC Land
Read at pcauthority.com.au →First reported by impli.meMore coverage: europeanbusinessreview.com
Houston nonprofit wins Meta grant to study AI glasses for disabilities
1 outlet · 1 article · 3 total since July 29, 2026 · Meta, wearable technology, alterpoint, AI Glasses Impact Grant, United States
A grant from Meta will help Easter Seals Greater Houston determine whether wearable AI smart glasses improve accessibility for those with disabilities. The local nonprofit is one organization that received a share of $2 million that Meta allocated in late July. "Meta is proud to…
Get the top AI and tech news in your inbox
One email each evening with the day's biggest stories. No account needed, unsubscribe anytime.
How this was computed
Articles in the Skillenai news index are grouped into event clusters at enrichment time (hybrid embedding/BM25/entity candidate search plus an LLM join-vs-new judge). Stories are ranked by the authority of the articles published on 2026-07-29 (UTC) — the sum of their outlet and author authority — so a multi-day story only ranks on the strength of that day's coverage. Article and outlet counts are likewise scoped to the day. The highlighted link is the day's highest-authority article in each cluster.