When Open Models Learn Advanced Cyber Exploitation
from AI & Tech Daily with Jesse Owen Hosted by Jesse Owen
Japan's AI Safety Institute has found that an open-weight model could complete an advanced cyber-exploitation task, complicating easy assumptions about model access and safeguards. Jesse also looks at the new US federal AI task force, Anthropic's $100 million enterprise engineering academy, Google's first Project Suncatcher test in orbit, direct Iceberg and Parquet queries from Aurora PostgreSQL, GitHub's new security-advisory tools, AMD's space-grade Versal AI package, and the Gemini Skills rollout replacing Gems. Full transcript: https://www.owenonthenet.com/podcast/2026-10-05/ Sources: https://aisi.go.jp/activity/activity_notes/261002_2 https://apnews.com/article/b8689ea07de9102a52bd1cd2049b5901 https://www.anthropic.com/news/claude-frontier-academy https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-prototype https://aws.amazon.com/about-aws/whats-new/2026/09/aurora-postgresql-query-apache-iceberg-and-parquet https://github.blog/changelog/2026-10-02-confidential-comments-on-repository-security-advisories https://github.blog/changelog/2026-10-02-repository-security-advisory-comments-api-in-public-preview https://newsroom.amd.com/news/amd-sampling-versal-ai-core-adaptive-soc https://blog.google/products-and-platforms/products/gemini/automate-tasks-with-skills
Embed this episode
What this episode covers
Japan's AI Safety Institute has found that an open-weight model could complete an advanced cyber-exploitation task, complicating easy assumptions about model access and safeguards. Jesse also looks at the new US federal AI task force, Anthropic's $100 million enterprise engineering academy, Google's first Project Suncatcher test in orbit, direct Iceberg and Parquet queries from Aurora PostgreSQL, GitHub's new security-advisory tools, AMD's space-grade Versal AI package, and the Gemini Skills rollout replacing Gems.
No transcript yet.