PJM plans to study SoftBank’s Ohio AI project under state agreement approach

How Do You Threat-Model an Application That Uses Autonomous AI Agents?

You spent years building a threat model you trust. You know your trust boundaries, you've mapped your data flows, and your team can recite STRIDE in their sleep.

• Autonomous AI agents act on untrusted input, unlike traditional software.
• Prompt injection is the new SQL injection risk for AI agents.
• Threat modeling must include agent, tools, and memory as assets with trust boundaries.

3h ago ·dev.to

I finally cut AI context bloat by 80% and the fix was embarrassingly simple

A few months ago I watched an agent do something deeply cursed. It had already fetched the data. It had already parsed the data. It had already stored the useful fields.

• AI agents previously included entire raw tool responses in subsequent model calls
• Solution was to store full results in structured state outside the model
• This change improved agent quality and efficiency significantly

3h ago ·dev.to

Artificial intelligence created an attack for Chrome in six days

Researchers at security firm XBOW used an AI model to build a full-chain Google Chrome exploit combining three unknown bugs in six days.

• Researchers at XBOW created Chrome exploit in six days using AI
• AI performed source code analysis, testing, and initial exploit creation
• Exploit affects Windows users, no user interaction needed

3h ago ·dev.to

Thrive TRM’s AI Capabilities Turn Years of Executive Search History Into Faster, Smarter Hires

The executive talent platform built for private equity and venture capital, in-house teams, and search firms is now available with secure AI that compounds intelligence with every search.

3h ago ·financialpost.com