Home

a shaneful place

News

Artificial Intelligence

Blog

Portfolio

Subscribe

more pain more gain 🚀

© 2024-2025 Shane "Lx". All rights reserved.

AI News Digest - 2026-06-11

Category

Empty

1.

Anthropic released Claude Fable 5, the first model in its Mythos class, which led multiple benchmarks including SWE-bench Verified at 95 percent, cost about twice as much as Opus 4.8, and applied strict safety filters that blocked roughly nine percent of requests while instituting a 30-day data retention policy even for zero-data-retention contracts.

2.

Anthropic's security team found that its Mythos Preview model converted security patches for Firefox and the Windows kernel into working exploits within hours for a few thousand dollars and without specialized knowledge, completing eight attack chains before Microsoft's auto-updates had reached any device.

3.

Google upgraded NotebookLM to run on Gemini 3.5 Flash, provided it with a cloud computer for code execution, and added agent-based research with autonomous source-finding via Google Search; internal tests showed the new system outperformed the previous version up to 78.2 percent of the time.

4.

Google released DiffusionGemma, a 26-billion-parameter open model that generated text via diffusion from noise instead of autoregressive token-by-token methods; Nvidia reported it ran at roughly 1,000 tokens per second on a single H100 GPU, though output quality was lower and Google positioned it as an experimental tool for developers.

5.

Germany's National Security Council approved creating a DE-AISI AI Safety Institute modeled after the UK's AISI to test frontier models from providers such as Anthropic and OpenAI for security risks, and stated the EU remained dependent on U.S. and Chinese AI technology in the absence of its own frontier models.

References

1.

https://the-decoder.com/claude-fable-5-the-first-mythos-model-is-powerful-expensive-and-heavily-filtered/

Claude Fable 5: The first Mythos model is powerful, expensive, and heavily filtered

Anthropic has released Claude Fable 5, the first model in its new Mythos class. It leads nearly every benchmark, including SWE-bench Verified at 95 percent, but costs twice as much as Opus 4.8 at 10 or 50 dollars per million tokens. Strict safety filters block about nine percent of requests, and a new 30-day data retention policy applies even to zero-data-retention contracts.

the-decoder.com

1.

https://the-decoder.com/anthropic-study-shows-ai-needs-hours-not-weeks-to-build-exploits-from-security-patches/

Anthropic study shows AI needs hours, not weeks, to build exploits from security patches

Anthropic's security team found that its Mythos Preview AI model can turn security patches for Firefox and the Windows kernel into working exploits within hours, for a few thousand dollars and no specialized knowledge. Eight complete attack chains were finished before Microsoft's auto-updates had reached a single device. The old patch rhythm is obsolete, Anthropic argues.

the-decoder.com

1.

https://the-decoder.com/googles-notebooklm-now-runs-its-own-cloud-computer-with-code-execution-and-agent-based-research/

Google's NotebookLM now runs its own cloud computer with code execution and agent-based research

Google is giving NotebookLM a major upgrade. The research tool now runs on Gemini 3.5 Flash, has its own cloud computer for code execution, and can find sources on its own via Google Search. In internal tests, the new system beat the previous version up to 78.2 percent of the time.

the-decoder.com

1.

https://the-decoder.com/googles-new-open-model-diffusiongemma-generates-text-from-noise-instead-of-word-by-word/

Google's new open model DiffusionGemma generates text from noise instead of word by word

Google released DiffusionGemma, a 26-billion-parameter model that generates text not token by token but through diffusion, similar to how image AI turns noise into a picture. According to Nvidia, it hits about 1,000 tokens per second on a single H100 GPU, roughly four times faster than comparable autoregressive models. The speed comes at a cost, though. Output quality is lower, so Google is positioning it as an experimental tool for developers for now.

the-decoder.com

1.

https://the-decoder.com/germanys-national-security-council-greenlights-an-ai-safety-institute-modeled-after-the-uks-aisi/

Artificial Intelligence is changing the world. THE DECODER brings you all the news about AI.

the-decoder.com

Made with Slashpage