Hermes — Thursday 30 April 2026

1

🔐 security Schneier on Security

73%

Claude Mythos Has Found 271 Zero-Days in Firefox

That’s a lot. No, it’s an extraordinary number: Since February, the Firefox team has been working around the clock using frontier AI models to find and fix latent security vulnerabilities in the brow…

Novelty

90%

Depth

65%

Practical

50%

Surprise

90%

Relevance

85%

https://www.schneier.com/blog/archives/2026/04/claude-mythos-has-found-271-zero-days-in-firefox.html

2

🤖 ai The Decoder

64%

OpenAI lands on AWS one day after Microsoft deal restructuring

Microsoft and OpenAI dissolve their exclusivity deal. One day later, AWS rolls out three new OpenAI offerings on its Bedrock platform, including a jointly built agent service. The article OpenAI land…

https://the-decoder.com/openai-lands-on-aws-one-day-after-microsoft-deal-restructuring/

3

🤖 ai The Decoder

62%

With Nemotron 3 Nano Omni, Nvidia reveals what really goes into a modern multimodal model

Nvidia releases Nemotron 3 Nano Omni, an open multimodal model for text, image, video and audio. Not only the performance is exciting, but also a look at the training data: it comes from Qwen, GPT-OS…

https://the-decoder.com/with-nemotron-3-nano-omni-nvidia-reveals-what-really-goes-into-a-modern-multimodal-model/

4

📦 m365 Petri IT Knowledgebase

60%

Microsoft to Block Legacy TLS Connections for POP and IMAP in Exchange Online

Microsoft is strengthening security in Exchange Online by retiring support for outdated email encryption methods. Beginning in July 2026, organizations still using legacy TLS versions for POP3 and IM…

https://petri.com/exchange-online-block-legacy-tls-pop-imap/

5

🤖 ai AI Alignment Forum

59%

Research Sabotage in ML Codebases

One of the main hopes for AI safety is using AIs to automate AI safety research. However, if models are misaligned, then they may sabotage the safety research. For example, misaligned AIs may try to:…

https://www.alignmentforum.org/posts/LByP4qsF8a4g7Pz3p/research-sabotage-in-ml-codebases

6

🔐 security SecurityWeek

55%

Fresh LiteLLM Vulnerability Exploited Shortly After Disclosure

The vulnerability allows attackers to read data from a LiteLLM proxy’s database and potentially modify it. The post Fresh LiteLLM Vulnerability Exploited Shortly After Disclosure appeared first on Se…

https://www.securityweek.com/fresh-litellm-vulnerability-exploited-shortly-after-disclosure/

7

🔐 security SecurityWeek

55%

Critical GitHub Vulnerability Exposed Millions of Repositories

The remote code execution flaw CVE-2026-3854 was found to impact GitHub.com and GitHub Enterprise Server. The post Critical GitHub Vulnerability Exposed Millions of Repositories appeared first on Sec…

https://www.securityweek.com/critical-github-vulnerability-exposed-millions-of-repositories/

8

📦 m365 Petri IT Knowledgebase

53%

AI PCs Gain Momentum as Enterprises Move Beyond the Cloud

Most enterprises are already testing or deploying AI PCs as they shift AI workloads beyond the cloud to enable faster, more secure processing directly on devices. As autonomous, agent-driven systems …

https://petri.com/ai-pcs-enterprise-adoption-beyond-cloud/

9

⚡ tech Stratechery

53%

Intel Earnings, Intel’s Differentiation?, Whither Terafab

Intel's earnings were very impressive, but the chief driver was a structural shift in demand for CPUs for AI. Plus, what is going on with Terafab?

https://stratechery.com/2026/intel-earnings-intels-differentiation-whither-terafab/

10

🔐 security Microsoft Security Blog

50%

8 best practices for CISOs conducting risk reviews

Embracing strong proactive security is something we can all do to mitigate our increased exposure to security threats. The post 8 best practices for CISOs conducting risk reviews appeared first on Mi…

https://www.microsoft.com/en-us/security/blog/2026/04/29/8-best-practices-for-cisos-conducting-risk-reviews/

11

🔐 security SecurityWeek

50%

Hundreds of Internet-Facing VNC Servers Expose ICS/OT

Forescout has identified tens of thousands of exposed RDP and VNC servers that can be mapped to specific industries. The post Hundreds of Internet-Facing VNC Servers Expose ICS/OT appeared first on S…

https://www.securityweek.com/hundreds-of-internet-facing-vnc-servers-expose-ics-ot/

12

🤖 ai AI Alignment Forum

50%

The other paper that killed deep learning theory

Yesterday, I wrote about the state of deep learning theory circa 2016,[1] as well as the bombshell 2016 paper by Zhang et al. that arguably signaled its demise. Today, I cover the aftermath, and the …

https://www.alignmentforum.org/posts/zcGmdQHX66NhC69v6/the-other-paper-that-killed-deep-learning-theory

13

🔐 security SecurityWeek

49%

Checkmarx Confirms Data Stolen in Supply Chain Attack

The hackers exfiltrated the data from Checkmarx’s GitHub environment on March 30, a week after publishing malicious code. The post Checkmarx Confirms Data Stolen in Supply Chain Attack appeared first…

https://www.securityweek.com/checkmarx-confirms-data-stolen-in-supply-chain-attack/

14

🤖 ai The Decoder

49%

Google Gemini now generates full documents, spreadsheets, and presentations directly inside the chat

Google Gemini now creates documents directly in the chat, from PDFs and Word files to Excel spreadsheets. The article Google Gemini now generates full documents, spreadsheets, and presentations direc…

https://the-decoder.com/google-gemini-now-generates-full-documents-spreadsheets-and-presentations-directly-inside-the-chat/

15

🤖 ai The Decoder

48%

OpenAI researchers explain why math is the road to AGI

AI models have jumped from grade-school arithmetic to olympiad-level and research mathematics in only two years. In the OpenAI Podcast, OpenAI researchers Sebastian Bubeck and Ernest Ryu explain why …

https://the-decoder.com/openai-researchers-explain-why-math-is-the-road-to-agi/