Experimental News Clipping Site

Tag: AI security

  • OpenAI : Measuring the performance of our models on real-world tasks

    —

    by

    Kurt Seifried
    in

    Source URL: https://openai.com/index/gdpval Source: OpenAI Title: Measuring the performance of our models on real-world tasks Feedly Summary: OpenAI introduces GDPval-v0, a new evaluation that measures model performance on real-world economically valuable tasks across 44 occupations. AI Summary and Description: Yes Summary: OpenAI’s introduction of GDPval-v0 represents a significant advancement in evaluating AI model performance, particularly…

  • The Register: DARPA amps up effort to make AI power-conscious

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.theregister.com/2025/09/25/dapra_ai_power_conscious/ Source: The Register Title: DARPA amps up effort to make AI power-conscious Feedly Summary: New research program seeks ‘energy-aware’ ML that balances performance with power draw It’s notoriously difficult to consistently measure the energy usage of AI models, but DARPA wants to put an end to that uncertainty with new “energy-aware" machine…

  • New York Times – Artificial Intelligence : Elon Musk’s xAI Signs Deal to Provide Grok Chatbot to US Agencies

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.nytimes.com/2025/09/25/technology/grok-xai-government-elon-musk.html Source: New York Times – Artificial Intelligence Title: Elon Musk’s xAI Signs Deal to Provide Grok Chatbot to US Agencies Feedly Summary: The billionaire’s artificial intelligence company, xAI, has secured a deal with the government that will allow federal agencies to use the chatbot Grok for a small fee. AI Summary and…

  • The Register: Microsoft puts Claude on the M365 menu

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.theregister.com/2025/09/25/microsoft_claude_copilot/ Source: The Register Title: Microsoft puts Claude on the M365 menu Feedly Summary: Redmond wants more flavors of ML than OpenAI can cook up Microsoft has sealed a deal with Anthropic to give users of Microsoft 365 Copilot the option to use the Claude AI engine.… AI Summary and Description: Yes Summary:…

  • Simon Willison’s Weblog: Cross-Agent Privilege Escalation: When Agents Free Each Other

    —

    by

    Kurt Seifried
    in

    Source URL: https://simonwillison.net/2025/Sep/24/cross-agent-privilege-escalation/ Source: Simon Willison’s Weblog Title: Cross-Agent Privilege Escalation: When Agents Free Each Other Feedly Summary: Cross-Agent Privilege Escalation: When Agents Free Each Other Here’s a clever new form of AI exploit from Johann Rehberger, who has coined the term Cross-Agent Privilege Escalation to describe an attack where multiple coding agents – GitHub…

  • Embrace The Red: Cross-Agent Privilege Escalation: When Agents Free Each Other

    —

    by

    Kurt Seifried
    in

    Source URL: https://embracethered.com/blog/posts/2025/cross-agent-privilege-escalation-agents-that-free-each-other/ Source: Embrace The Red Title: Cross-Agent Privilege Escalation: When Agents Free Each Other Feedly Summary: During the Month of AI Bugs, I described an emerging vulnerability pattern that shows how commonly agentic systems have a design flaw that allows an agent to overwrite its own configuration and security settings. This allows the…

  • Wired: This AI-Powered Robot Keeps Going Even if You Attack It With a Chainsaw

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.wired.com/story/this-ai-powered-robot-keeps-going-even-if-you-attack-it-with-a-chainsaw/ Source: Wired Title: This AI-Powered Robot Keeps Going Even if You Attack It With a Chainsaw Feedly Summary: A single AI model trained to control numerous robotic bodies can operate unfamiliar hardware and adapt eerily well to serious injuries. AI Summary and Description: Yes Summary: The text discusses an AI model capable…

  • Simon Willison’s Weblog: GPT-5-Codex

    —

    by

    Kurt Seifried
    in

    Source URL: https://simonwillison.net/2025/Sep/23/gpt-5-codex/#atom-everything Source: Simon Willison’s Weblog Title: GPT-5-Codex Feedly Summary: GPT-5-Codex OpenAI half-relased this model earlier this month, adding it to their Codex CLI tool but not their API. Today they’ve fixed that – the new model can now be accessed as gpt-5-codex. It’s priced the same as regular GPT-5: $1.25/million input tokens, $10/million…

  • Slashdot: MediaTek Launches Improved AI Processor To Compete With Qualcomm

    —

    by

    Kurt Seifried
    in

    Source URL: https://hardware.slashdot.org/story/25/09/23/0434209/mediatek-launches-improved-ai-processor-to-compete-with-qualcomm Source: Slashdot Title: MediaTek Launches Improved AI Processor To Compete With Qualcomm Feedly Summary: AI Summary and Description: Yes Summary: MediaTek’s launch of the Dimensity 9500 mobile processor enhances AI capabilities on devices, directly competing with Qualcomm in the performance of AI tasks. This advancement, built on a sophisticated 3-nanometer process, has…