Hermes — Tuesday 12 May 2026

1

🤖 ai The Decoder

72%

AI turns patches into working exploits in 30 minutes, and the 90-day disclosure window is the casualty

Language models find security flaws faster and turn patches into working exploits in minutes. A veteran researcher says the established disclosure process needs to change. The article AI turns patche…

Novelty

80%

Depth

55%

Practical

70%

Surprise

75%

Relevance

90%

https://the-decoder.com/ai-turns-patches-into-working-exploits-in-30-minutes-and-the-90-day-disclosure-window-is-the-casualty/

2

📦 m365 Petri IT Knowledgebase

71%

The “No-Breach” Breach: How Stealer Logs Lead to Active Directory Incidents

A 2026 Forbes article reported that the previous year’s credential-theft wave exposed 2.86 billion compromised credentials, with infostealers tied to roughly 3.9 million infected machines and 347.5 m…

Novelty

65%

Depth

75%

Practical

75%

Surprise

50%

Relevance

85%

https://petri.com/how-stealer-logs-lead-to-active-directory-incidents/

3

🤖 ai The Decoder

70%

AI agents can now hack computers and copy themselves, and they're getting better fast

Palisade Research shows that AI agents can hack remote computers, copy themselves onto them, and form replication chains. In one year, the success rate jumped from 6 to 81 percent. The researchers ex…

Novelty

85%

Depth

60%

Practical

50%

Surprise

88%

Relevance

85%

https://the-decoder.com/ai-agents-can-now-hack-computers-and-copy-themselves-and-theyre-getting-better-fast/

4

🤖 ai The Decoder

70%

METR says it can barely measure Claude Mythos, Palo Alto Networks warns of autonomous AI attackers

METR can barely measure Claude Mythos Preview with its current test suite. Only five out of 228 tasks cover the relevant capability range. Meanwhile, Palo Alto Networks reports that frontier models a…

Novelty

80%

Depth

65%

Practical

55%

Surprise

78%

Relevance

85%

https://the-decoder.com/metr-says-it-can-barely-measure-claude-mythos-palo-alto-networks-warns-of-autonomous-ai-attackers/

5

🔐 security SecurityWeek

70%

Google Detects First AI-Generated Zero-Day Exploit

The zero-day was designed to bypass 2FA and it was developed by a prominent cybercrime group. The post Google Detects First AI-Generated Zero-Day Exploit appeared first on SecurityWeek.

https://www.securityweek.com/google-detects-first-ai-generated-zero-day-exploit/

6

⚡ tech The Verge

70%

Google stopped a zero-day hack that it says was developed with AI

For the first time, Google says it has spotted and stopped a zero-day exploit developed with AI. According to a report from Google Threat Intelligence Group (GTIG), "prominent cyber crime threat acto…

https://www.theverge.com/tech/928007/google-ai-zero-day-exploit-stopped

7

⚡ tech Stratechery

66%

The Inference Shift

Agentic inference is going to be different than the inference we use today, and it will change compute infrastructure because speed won't matter when humans aren't involved.

https://stratechery.com/2026/the-inference-shift/

8

🤖 ai The Decoder

66%

Researchers may have found a way to stop AI models from intentionally playing dumb during safety evaluations

A study by researchers from the MATS program, Redwood Research, the University of Oxford, and Anthropic examines a safety problem that grows more pressing as AI systems become more capable: "sandbagg…

https://the-decoder.com/researchers-may-have-found-a-way-to-stop-ai-models-from-intentionally-playing-dumb-during-safety-evaluations/

9

🤖 ai The Decoder

66%

Fields Medalist says ChatGPT 5.5 Pro delivered "PhD-level" math research in under two hours with zero human help

Fields Medalist Timothy Gowers had ChatGPT 5.5 Pro tackle open problems in number theory. The model improved an exponential bound to a polynomial one in under an hour. An MIT researcher involved call…

https://the-decoder.com/fields-medalist-says-chatgpt-5-5-pro-delivered-phd-level-math-research-in-under-two-hours-with-zero-human-help/

10

🤖 ai The Decoder

60%

Baidu's Ernie 5.1 cuts 94 percent of pre-training costs while competing with top models

Baidu's Ernie 5.1 uses just a third of its predecessor's parameters and reportedly cost only six percent of what comparable models require to pre-train. That's possible thanks to a "Once-For-All" app…

https://the-decoder.com/baidus-ernie-5-1-cuts-94-percent-of-pre-training-costs-while-competing-with-top-models/

11

🤖 ai The Decoder

60%

Broadcom reportedly won't build OpenAI's custom chip unless Microsoft buys 40 percent of them

OpenAI's custom AI chip project with Broadcom has hit a funding wall. Broadcom won't finance production unless Microsoft commits to buying 40 percent of the chips, and Microsoft hasn't agreed yet. Op…

https://the-decoder.com/broadcom-reportedly-wont-build-openais-custom-chip-unless-microsoft-buys-40-percent-of-them/

12

⚡ tech The Verge

59%

OpenAI just released its answer to Claude Mythos

OpenAI is launching Daybreak, an AI initiative focused on detecting and patching vulnerabilities before attackers find them. Daybreak uses the Codex Security AI agent that launched in March to create…

https://www.theverge.com/ai-artificial-intelligence/928342/openai-daybreak-security-ai

13

🔐 security SecurityWeek

59%

Build Application Firewalls Aim to Stop the Next Supply Chain Attack

Rather than scanning code alone, Build Application Firewalls inspect runtime behavior inside the software build pipeline. The post Build Application Firewalls Aim to Stop the Next Supply Chain Attack…

https://www.securityweek.com/build-application-firewalls-aim-to-stop-the-next-supply-chain-attack/

14

🔐 security Schneier on Security

59%

LLMs and Text-in-Text Steganography

Turns out that LLMs are really good at hiding text messages in other text messages.

https://www.schneier.com/blog/archives/2026/05/llms-and-text-in-text-steganography.html

15

🤖 ai The Decoder

58%

GPT-5.5 costs 49 to 92 percent more than its predecessor, depending on the input length

OpenAI doubled GPT-5.5's list price compared to GPT-5.4, claiming shorter responses would offset the increase. An OpenRouter analysis of real usage data tells a different story: actual costs rose 49 …

https://the-decoder.com/gpt-5-5-costs-49-to-92-percent-more-than-its-predecessor-depending-on-the-input-length/