Hermes — Saturday 1 August 2026

1

🤖 ai The Decoder

74%

Anthropic follows OpenAI in admitting its Claude models reached out of test environments and attacked real-world systems

Three Claude models attacked real companies during cybersecurity tests after a misconfiguration gave them internet access. One published malware on PyPI that infected 15 systems. Another kept attacki…

Novelty

88%

Depth

65%

Practical

55%

Surprise

90%

Relevance

85%

https://the-decoder.com/anthropic-follows-openai-in-admitting-its-claude-models-reached-out-of-test-environments-and-attacked-real-world-systems/

2

📦 m365 Petri IT Knowledgebase

70%

Attackers Use Microsoft Teams Vishing Campaigns to Deploy Chaos Ransomware

Cybercriminals are exploiting Microsoft Teams voice phishing (vishing) campaigns to infiltrate corporate networks and deploy Chaos ransomware. Security researchers warn that the campaign targeted doz…

Novelty

70%

Depth

60%

Practical

75%

Surprise

60%

Relevance

90%

https://petri.com/microsoft-teams-vishing-chaos-ransomware/

3

🔐 security Microsoft Security Blog

70%

CaptiveCrunch: Midnight Blizzard targets travelers worldwide for malware delivery and credential theft

Storm-2945, a sub-cluster of the Russian threat actor Midnight Blizzard, has been observed compromising the sign-in portals of hospitality-related organizations such as hotels since May 2026 in order…

https://www.microsoft.com/en-us/security/blog/2026/07/31/captivecrunch-midnight-blizzard-targets-travelers-worldwide-for-malware-delivery-and-credential-theft/

4

🤖 ai MIT Technology Review – AI

69%

A fundamental flaw leaves LLMs strikingly vulnerable to attack

It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Confere…

https://www.technologyreview.com/2026/07/30/1140927/a-fundamental-flaw-leaves-llms-vulnerable-to-attack/

5

🤖 ai The Decoder

69%

Microsoft AI bets on cheap specialist models instead of chasing the frontier

Microsoft AI is betting on small specialist models instead of expensive general-purpose ones, according to AI CEO Mustafa Suleyman. MAI-Cyber-1-Flash tops the CyberGym benchmark when embedded in an o…

https://the-decoder.com/microsoft-ai-bets-on-cheap-specialist-models-instead-of-chasing-the-frontier/

6

🔐 security SecurityWeek

69%

Prompted by OpenAI Disclosure, Anthropic Finds Its Own Models Hacked 3 Organizations

A security company’s systems were hacked after it installed a malicious Python package deployed by Claude. The post Prompted by OpenAI Disclosure, Anthropic Finds Its Own Models Hacked 3 Organization…

https://www.securityweek.com/after-openai-disclosure-anthropic-finds-its-own-models-hacked-3-organizations/

7

🤖 ai AI Alignment Forum

67%

Value Leakage: An LLM’s Answers Are Silently Shaped by Its Own Values

TL;DR: LLMs should give accurate answers. Yet we find their answers are often biased to favor their own values and they don't disclose this in their reasoning. For example, when a user asks how likel…

https://www.alignmentforum.org/posts/hbMw4Yqw6RnFaExDy/value-leakage-an-llm-s-answers-are-silently-shaped-by-its-1

8

🔐 security SecurityWeek

66%

‘DangleGeddon’: AI Could Weaponize Forgotten DNS Records at Global Scale

Researchers warn that AI could turn dangling DNS takeovers into a nation-state weapon capable of disrupting governments, banks and global supply chains. The post ‘DangleGeddon’: AI Could Weaponize Fo…

https://www.securityweek.com/danglegeddon-ai-could-weaponize-forgotten-dns-records-at-global-scale/

9

🔐 security Krebs on Security

65%

Read This Before You Buy That TV Streaming Stick

Security experts have been sounding the alarm for years about the risks of using generic TV boxes that promise unlimited content streaming for a one-time fee, warning that they secretly rent the user…

https://krebsonsecurity.com/2026/07/read-this-before-you-buy-that-tv-streaming-stick/

10

🤖 ai AI Alignment Forum

63%

OpenAI has already ended an internal pause

One day before OpenAI’s HF incident disclosure, OpenAI disclosed that it paused internal deployment of a long-horizon model after it circumvented its sandbox, then restored access weeks later under n…

https://www.alignmentforum.org/posts/k3eKqKzq4Y7xnqEfZ/openai-has-already-ended-an-internal-pause

11

🔐 security SecurityWeek

63%

Critical Flaw Allowed to Azure Cosmos DB Pwnage

Named CosmosEscape, the vulnerability exposed the primary key for Cosmos DB accounts, granting full read and write access. The post Critical Flaw Allowed to Azure Cosmos DB Pwnage appeared first on S…

https://www.securityweek.com/critical-flaw-led-to-azure-cosmos-db-pwnage/

12

🤖 ai The Decoder

62%

OpenAI goes full China pricing mode with an 80 percent cut to its most affordable GPT-5.6 model

Starting July 30, OpenAI is cutting GPT-5.6 Luna prices by 80 percent and Terra by 20 percent. OpenAI says its top-tier Sol model helped make the company's own infrastructure more efficient, enabling…

https://the-decoder.com/openai-goes-full-china-pricing-mode-with-an-80-percent-cut-to-its-most-affordable-gpt-5-6-model/

13

🤖 ai The Decoder

61%

New Deepseek Flash model matches OpenAI's GPT-5.6 Luna at roughly 60 percent lower cost

Deepseek's budget model V4 Flash gets a major boost with the "0731" update, jumping ten points to 50 on the Artificial Analysis Intelligence Index. That puts it just one point behind OpenAI's GPT-5.6…

https://the-decoder.com/new-deepseek-flash-model-matches-openais-gpt-5-6-luna-at-roughly-60-percent-lower-cost/

14

🤖 ai The Decoder

61%

Ex-OpenAI researcher bets $100 billion will flow into training data because scaling alone won't cut it

Former OpenAI employee Andrew Ho and Cambridge researcher Adam Hunt see a growing problem with large language models. Instead of becoming more versatile, the models are becoming more specialized, exc…

https://the-decoder.com/ex-openai-researcher-bets-100-billion-will-flow-into-training-data-because-scaling-alone-wont-cut-it/

15

🔐 security SecurityWeek

61%

Google AI Uncovers 13-Year-Old Chrome Flaw Amid Record Patching Pace

The internet giant has built an agent harness to find vulnerabilities across Chrome’s codebase. The post Google AI Uncovers 13-Year-Old Chrome Flaw Amid Record Patching Pace appeared first on Securit…

https://www.securityweek.com/googles-ai-agent-uncovers-13-year-old-chrome-flaw-amid-record-patching-pace/