<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>AI This Week — Pranav Dhawan</title>
    <link>https://pranavdhawan.com/blog/</link>
    <atom:link href="https://pranavdhawan.com/blog/feed.xml" rel="self" type="application/rss+xml"/>
    <description>A weekly, auto-curated digest of AI developments from official lab blogs, arXiv, and GitHub.</description>
    <language>en-us</language>
    <lastBuildDate>Mon, 10 Aug 2026 00:00:00 GMT</lastBuildDate>
    <item>
      <title>Responding to the next frontier of critical cyber capabilities</title>
      <link>https://openai.com/index/responding-next-frontier-critical-cyber-capabilities</link>
      <guid isPermaLink="false">https://openai.com/index/responding-next-frontier-critical-cyber-capabilities</guid>
      <pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate>
      <category>Research</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">OpenAI</source>
      <description>OpenAI is sharing preliminary cybersecurity evaluations for Astra and the steps we’re taking to strengthen safeguards and security controls.</description>
    </item>
    <item>
      <title>Improving Fable 5 S Biology Safeguards</title>
      <link>https://www.anthropic.com/news/improving-fable-5-s-biology-safeguards</link>
      <guid isPermaLink="false">https://www.anthropic.com/news/improving-fable-5-s-biology-safeguards</guid>
      <pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate>
      <category>Product Update</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Anthropic</source>
      <description>Read the full story at the source.</description>
    </item>
    <item>
      <title>TutorMoments: Do AI tutors know when to help and when to hold back?</title>
      <link>https://huggingface.co/blog/allenai/tutormoments</link>
      <guid isPermaLink="false">https://huggingface.co/blog/allenai/tutormoments</guid>
      <pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate>
      <category>Open Source</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Hugging Face</source>
      <description>Read the full story at the source.</description>
    </item>
    <item>
      <title>MirrorWorld: Taming Video Diffusion Models for Mirror Reflection Generation</title>
      <link>https://arxiv.org/abs/2608.07463v1</link>
      <guid isPermaLink="false">https://arxiv.org/abs/2608.07463v1</guid>
      <pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate>
      <category>Research</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">arXiv</source>
      <description>Recent advances in video diffusion models (VDMs) have enabled high-fidelity video synthesis. However, generating mirror reflections remains challenging because the content within a mirror must remain consistent with the…</description>
    </item>
    <item>
      <title>How HSP GRUPPE builds AI capabilities for tax advisory</title>
      <link>https://openai.com/index/hsp-gruppe</link>
      <guid isPermaLink="false">https://openai.com/index/hsp-gruppe</guid>
      <pubDate>Fri, 07 Aug 2026 00:00:00 GMT</pubDate>
      <category>Product Update</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">OpenAI</source>
      <description>Discover how HSP GRUPPE uses ChatGPT Enterprise to boost productivity, improve work quality, and create more capacity for tax advisory and client service.</description>
    </item>
    <item>
      <title>WeatherNext: AI model achieves breakthrough in forecasting cyclones</title>
      <link>https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/</link>
      <guid isPermaLink="false">https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones</guid>
      <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
      <category>Research</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Google DeepMind</source>
      <description>Read the full story at the source.</description>
    </item>
    <item>
      <title>Why Meta Builds Its Own AI Data Centers</title>
      <link>https://about.fb.com/news/2026/08/why-meta-builds-its-own-data-centers/</link>
      <guid isPermaLink="false">https://about.fb.com/news/2026/08/why-meta-builds-its-own-data-centers</guid>
      <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
      <category>Product Update</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Meta</source>
      <description>Developer and creator Tom Shaw sits down with Meta’s VP of Data Centers Rachel Peterson to discuss how Meta is building some of the world's largest computing facilities. The post Why Meta Builds Its Own AI Data Centers…</description>
    </item>
    <item>
      <title>The latest AI news we announced in July 2026</title>
      <link>https://blog.google/innovation-and-ai/technology/ai/google-ai-updates-july-2026/</link>
      <guid isPermaLink="false">https://blog.google/innovation-and-ai/technology/ai/google-ai-updates-july-2026</guid>
      <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
      <category>Product Update</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Google AI</source>
      <description>Here are Google’s latest AI updates from July 2026</description>
    </item>
    <item>
      <title>Introducing Shieldstral.</title>
      <link>https://mistral.ai/news/shieldstral/</link>
      <guid isPermaLink="false">https://mistral.ai/news/shieldstral</guid>
      <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
      <category>Model Release</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Mistral AI</source>
      <description>Shieldstral introduces a 3B open-weights multimodal safety classifier that outperforms models up to 7x its size.</description>
    </item>
    <item>
      <title>Solving the solvent problem</title>
      <link>https://news.mit.edu/2026/solving-solvent-problem-sodium-metal-batteries-0804</link>
      <guid isPermaLink="false">https://news.mit.edu/2026/solving-solvent-problem-sodium-metal-batteries-0804</guid>
      <pubDate>Tue, 04 Aug 2026 00:00:00 GMT</pubDate>
      <category>Research</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">MIT News — AI</source>
      <description>By focusing on electrolytes, MIT scientists are making sodium-metal batteries a more practical energy storage option.</description>
    </item>
    <item>
      <title>FareedKhan-dev/kimi-k3-in-c — trending AI repo</title>
      <link>https://github.com/FareedKhan-dev/kimi-k3-in-c</link>
      <guid isPermaLink="false">https://github.com/fareedkhan-dev/kimi-k3-in-c</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <category>Open Source</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">GitHub</source>
      <description>A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU. (★ 795)</description>
    </item>
    <item>
      <title>From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon</title>
      <link>https://bair.berkeley.edu/blog/2026/07/29/cuda-to-mlx-k-search/</link>
      <guid isPermaLink="false">https://bair.berkeley.edu/blog/2026/07/29/cuda-to-mlx-k-search</guid>
      <pubDate>Wed, 29 Jul 2026 00:00:00 GMT</pubDate>
      <category>Open Source</category>
      <source url="https://pranavdhawan.com/blog/feed.xml">Berkeley AI Research</source>
      <description>Figure 1: CUDA-to-MLX optimization translation map. CUDA optimization knowledge can be translated into architecture-native MLX strategies rather than copied instruction-for-instruction. We face a new epoch in computing…</description>
    </item>
  </channel>
</rss>
