Hermes — Saturday 13 June 2026

1

🤖 ai AI Alignment Forum

72%

Models May Behave Worse When Eval Aware

This is the first in a series of research updates from the Google DeepMind Language Model Interpretability team, in interpretability and adjacent areas. TL;DR It's often assumed that models will act …

Novelty

85%

Depth

75%

Practical

50%

Surprise

85%

Relevance

75%

https://www.alignmentforum.org/posts/aTcsN5ZZDnMFJvRiG/models-may-behave-worse-when-eval-aware

2

🔐 security SecurityWeek

68%

‘GreatXML’ Zero-Day Exploit Bypasses BitLocker

The PoC exploits Microsoft Defender’s offline scan to spawn a SYSTEM shell when rebooting in Recovery Mode. The post ‘GreatXML’ Zero-Day Exploit Bypasses BitLocker appeared first on SecurityWeek.

https://www.securityweek.com/greatxml-zero-day-exploit-bypasses-bitlocker/

3

🤖 ai AI Alignment Forum

68%

Building and evaluating model diffing agents

This is the second in a series of informal research updates from the Google DeepMind Language Model Interpretability team, in interpretability and adjacent areas. The first post can be found here. TL…

https://www.alignmentforum.org/posts/qi4mNbZYAFDYwfRba/building-and-evaluating-model-diffing-agents

4

📦 m365 Petri IT Knowledgebase

67%

Why Microsoft 365’s Baseline Security Mode Is a Migration Target, Not a Flip Switch

Microsoft 365’s new Baseline Security Mode is an opt-in, secure-by-default bundle of 18 configuration settings across authentication, files, and room devices, exposed in Org settings → Security & pri…

https://petri.com/microsoft-365-baseline-security-mode-not-a-flip-switch/

5

🤖 ai The Decoder

63%

Dario Amodei's new essay reads like a Cold War playbook for the AI age

Anthropic publishes a sweeping essay and two policy frameworks. The company calls for binding audits of frontier models and paints a picture of AI as a strategic weapon wielded by nation-states. The …

https://the-decoder.com/dario-amodeis-new-essay-reads-like-a-cold-war-playbook-for-the-ai-age/

6

📦 m365 Petri IT Knowledgebase

63%

One Copilot to Rule Them All? Microsoft’s Unified AI Platform for Work and Life

Microsoft went all-in on One Copilot earlier this year. In March 2026, CEO Satya Nadella reorganized the company’s AI efforts, consolidating the consumer and enterprise Copilot teams under one unifie…

https://petri.com/one-copilot-to-rule-them-all-life-and-work/

7

🤖 ai MIT Technology Review – AI

62%

Google DeepMind is worried about what happens when millions of agents start to interact

Google DeepMind is funding research into the potential dangers of situations where millions of different AI agents interact with each other online. According to Rohin Shah, who directs the company’s …

https://www.technologyreview.com/2026/06/11/1138794/google-deepmind-is-worried-about-what-happens-when-millions-of-agents-start-to-interact/

8

🔐 security SecurityWeek

61%

Google Confirms Exploitation of Oracle PeopleSoft Zero-Day by ShinyHunters

Oracle has mitigated CVE-2026-35273, but it has not publicly confirmed the vulnerability’s in-the-wild exploitation. The post Google Confirms Exploitation of Oracle PeopleSoft Zero-Day by ShinyHunter…

https://www.securityweek.com/google-confirms-exploitation-of-oracle-peoplesoft-zero-day-by-shinyhunters/

9

🔐 security SecurityWeek

60%

Oracle Addresses PeopleSoft Vulnerability Amid Reports of Zero-Day Attacks

Oracle has released mitigations for CVE-2026-35273, but it has not said whether it’s a zero-day exploited in ShinyHunters attacks. The post Oracle Addresses PeopleSoft Vulnerability Amid Reports of Z…

https://www.securityweek.com/oracle-addresses-peoplesoft-vulnerability-amid-reports-of-zero-day-attacks/

10

🤖 ai The Decoder

59%

Google files first joint lawsuit with FBI over Chinese AI scam network, OpenAI blocks PRC influence clusters

Within days of each other, Google and OpenAI separately exposed operations allegedly originating in China that use AI for fraud and covert influence campaigns. Both target US infrastructure and polit…

https://the-decoder.com/google-files-first-joint-lawsuit-with-fbi-over-chinese-ai-scam-network-openai-blocks-prc-influence-clusters/

11

🤖 ai The Decoder

58%

Anthropic's Claude Fable 5 costs twice as much for 5.7 percent more performance

Claude Fable 5 tops the Artificial Analysis Intelligence Index with 64.9 points and sets records in five of ten benchmarks. But the gain over Opus 4.8 is just 5.7 percent at double the token price. S…

https://the-decoder.com/anthropics-claude-fable-5-costs-twice-as-much-for-5-7-percent-more-performance/

12

📦 m365 Petri IT Knowledgebase

58%

Microsoft Previews IAKerb And LocalKDC To Reduce NTLM Reliance in Windows 11

Microsoft is preparing to introduce two new authentication features in Windows 11 that are designed to reduce reliance on NTLM. The capabilities are currently available in public preview for Windows …

https://petri.com/windows-11-iakerb-localkdc-reduce-ntlm/

13

📦 m365 Petri IT Knowledgebase

57%

Microsoft’s June 2026 Patch Tuesday Fixes Critical Windows RCE Vulnerabilities

Microsoft has released the June 2026 Patch Tuesday updates for Windows 11 versions 25H2, 24H2, and 26H1. This month, the company has fixed over 200 vulnerabilities in Windows, Office, Microsoft Edge,…

https://petri.com/microsoft-june-2026-patch-tuesday-updates/

14

🤖 ai The Decoder

57%

Claude Fable 5: Anthropic admits "wrong tradeoff" after invisibly throttling rival AI researchers

Anthropic reverses course on a policy that would have secretly undermined AI researchers, but another point of contention persists. The article Claude Fable 5: Anthropic admits "wrong tradeoff" after…

https://the-decoder.com/claude-fable-5-anthropic-admits-wrong-tradeoff-after-invisibly-throttling-rival-ai-researchers/

15

📦 m365 Petri IT Knowledgebase

56%

NCSC Warns Of Rising Software Supply Chain Attacks Targeting Open-Source Packages

The UK’s National Cyber Security Centre (NCSC) is urging organizations to take a closer look at their software dependencies as supply chain attacks continue to rise. The agency highlights how attacke…

https://petri.com/supply-chain-attacks-open-source-packages/