Experimental News Clipping Site

Tag: critical

  • The Cloudflare Blog: Reducing double spend latency from 40 ms to < 1 ms on privacy proxy

    —

    by

    Kurt Seifried
    in

    Source URL: https://blog.cloudflare.com/reducing-double-spend-latency-from-40-ms-to-less-than-1-ms-on-privacy-proxy/ Source: The Cloudflare Blog Title: Reducing double spend latency from 40 ms to < 1 ms on privacy proxy Feedly Summary: We significantly sped up our privacy proxy service by fixing a 40ms delay in “double-spend" checks. AI Summary and Description: Yes **Summary:** This text discusses performance improvements made to Cloudflare’s privacy…

  • The Register: Patch now: Millions of Dell PCs with Broadcom chips vulnerable to attack

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.theregister.com/2025/08/05/millions_of_dell_pc_with/ Source: The Register Title: Patch now: Millions of Dell PCs with Broadcom chips vulnerable to attack Feedly Summary: Psst, wanna steal someone’s biometrics? black hat Critical security flaws in Broadcom chips used in more than 100 models of Dell computers could allow attackers to take over tens of millions of users’ devices,…

  • OpenAI : Estimating worst case frontier risks of open weight LLMs

    —

    by

    Kurt Seifried
    in

    Source URL: https://openai.com/index/estimating-worst-case-frontier-risks-of-open-weight-llms Source: OpenAI Title: Estimating worst case frontier risks of open weight LLMs Feedly Summary: In this paper, we study the worst-case frontier risks of releasing gpt-oss. We introduce malicious fine-tuning (MFT), where we attempt to elicit maximum capabilities by fine-tuning gpt-oss to be as capable as possible in two domains: biology and…

  • OpenAI : Open Weights and AI for All

    —

    by

    Kurt Seifried
    in

    Source URL: https://openai.com/global-affairs/open-weights-and-ai-for-all Source: OpenAI Title: Open Weights and AI for All Feedly Summary: AI’s next frontier isn’t just about capability—it’s about who gets to use it. Our mission to put AI in the hands of as many people as possible is what drives us. Today’s release of our most capable open-weights models is a…

  • Microsoft Security Blog: Elevate your protection with expanded Microsoft Defender Experts coverage

    —

    by

    Kurt Seifried
    in

    Source URL: https://techcommunity.microsoft.com/blog/microsoftsecurityexperts/elevate-your-protection-with-expanded-microsoft-defender-experts-coverage/4439134 Source: Microsoft Security Blog Title: Elevate your protection with expanded Microsoft Defender Experts coverage Feedly Summary: Defender Experts now offers 24/7, expert-driven protection for cloud workloads, beginning with hybrid and multicloud servers in Microsoft Defender for Cloud. Additionally, third-party network signals can be used in Microsoft Defender Experts for XDR to enhance…

  • Slashdot: Microsoft Teases the Future of Windows as an Agentic OS

    —

    by

    Kurt Seifried
    in

    Source URL: https://tech.slashdot.org/story/25/08/05/1647209/microsoft-teases-the-future-of-windows-as-an-agentic-os?utm_source=rss1.0mainlinkanon&utm_medium=feed Source: Slashdot Title: Microsoft Teases the Future of Windows as an Agentic OS Feedly Summary: AI Summary and Description: Yes Summary: Microsoft has released a video on its “Windows 2030 Vision,” highlighting transformative changes anticipated for the Windows operating system, particularly driven by advancements in AI. The discussion, led by David Weston,…

  • Embrace The Red: Amp Code: Arbitrary Command Execution via Prompt Injection Fixed

    —

    by

    Kurt Seifried
    in

    Source URL: https://embracethered.com/blog/posts/2025/amp-agents-that-modify-system-configuration-and-escape/ Source: Embrace The Red Title: Amp Code: Arbitrary Command Execution via Prompt Injection Fixed Feedly Summary: Sandbox-escape-style attacks can happen when an AI is able to modify its own configuration settings, such as by writing to configuration files. That was the case with Amp, an agentic coding tool built by Sourcegraph. The…

  • Slashdot: Nearly 100,000 ChatGPT Conversations Were Searchable on Google

    —

    by

    Kurt Seifried
    in

    Source URL: https://yro.slashdot.org/story/25/08/05/1535248/nearly-100000-chatgpt-conversations-were-searchable-on-google?utm_source=rss1.0mainlinkanon&utm_medium=feed Source: Slashdot Title: Nearly 100,000 ChatGPT Conversations Were Searchable on Google Feedly Summary: AI Summary and Description: Yes Summary: The text discusses a significant privacy concern regarding nearly 100,000 publicly shared conversations from OpenAI’s ChatGPT that were indexed by Google. It highlights the potential risks involved when users share conversations, revealing a…

  • The Register: Chained bugs in Nvidia’s Triton Inference Server lead to full system compromise

    —

    by

    Kurt Seifried
    in

    Source URL: https://www.theregister.com/2025/08/05/nvidia_triton_bug_chain/ Source: The Register Title: Chained bugs in Nvidia’s Triton Inference Server lead to full system compromise Feedly Summary: Wiz Research details flaws in Python backend that expose AI models and enable remote code execution Security researchers have lifted the lid on a chain of high-severity vulnerabilities that could lead to remote code…