<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>The Panda Strike Blog</title>
    <link>https://blog.pandastrike.com/</link>
    <description>Exploring the Pandaverse</description>
    <pubDate>Wed, 13 May 2026 15:54:00 +0000</pubDate>
    <image>
      <url>https://i.snap.as/o5TMPz1g.ico</url>
      <title>The Panda Strike Blog</title>
      <link>https://blog.pandastrike.com/</link>
    </image>
    <item>
      <title>EU DMA Is For Real</title>
      <link>https://blog.pandastrike.com/eu-dma-is-for-real?pk_campaign=rss-feed</link>
      <description>&lt;![CDATA[When the EU’s DMA and its cousin, the DSA, passed in 2022, I was surprised. !--more-- First, it had gone from a proposal to law in less than two years. Second, and more to the point, from what I could tell (not being a lawyer), it looked like a serious attempt to regulate Big Tech. I had always believed it would take much longer for that to happen.&#xA;&#xA;Still, I waited to see how it would play out: perhaps there would be loopholes, or maybe the EU wouldn’t enforce the rules. After all, Big Tech has plenty of cash to spread around.&#xA;&#xA;The first indication that the DMA was for real was when Apple reversed course after initially disabling PWAs in iOS. The second was when the EU charged Apple with violating the DMA by “violating developers’ rights.” A ruling is due in March 2025, and fines can be as high as 10% of Apple’s worldwide revenue. Based on their 2023 revenue, that would be over $38B.&#xA;&#xA;We’ll see how it plays out, but the mere threat of a fine that large, coupled with the no-nonsense process the DMA lays out for enforcement, may coerce Big Tech into compliance. In turn, this could make the DMA the de facto regulatory framework worldwide. Firms would otherwise need special variants of their platforms for each regulatory environment. Perhaps they will do so, but, in many cases, it’s hard to even know what that would mean.&#xA;&#xA;Update: They’ve also charged Microsoft.&#xA;&#xA;Update: I don’t know if the EU is entirely serious, but Apple sure seems to think so.&#xA;&#xA;#EU #DMA]]&gt;</description>
      <content:encoded><![CDATA[<p>When the EU’s DMA and its cousin, the DSA, passed in 2022, I was surprised.  First, it had gone from a proposal to law in less than two years. Second, and more to the point, from what I could tell (not being a lawyer), it looked like a serious attempt to regulate Big Tech. I had always believed it would take much longer for that to happen.</p>

<p>Still, I waited to see how it would play out: perhaps there would be loopholes, or maybe the EU wouldn’t enforce the rules. After all, Big Tech has plenty of cash to spread around.</p>

<p>The first indication that the DMA was for real was when Apple <a href="https://arstechnica.com/gadgets/2024/03/apple-changes-course-will-keep-iphone-eu-web-apps-how-they-are-in-ios-17-4/">reversed course</a> after <a href="https://open-web-advocacy.org/blog/did-apple-just-break-web-apps-in-ios17.4-beta-eu/">initially disabling PWAs</a> in iOS. The second was when the EU <a href="https://arstechnica.com/tech-policy/2024/06/eu-says-apple-violated-app-developers-rights-could-be-fined-10-of-revenue/">charged Apple</a> with violating the DMA by “violating developers’ rights.” A ruling is due in March 2025, and fines can be as high as 10% of Apple’s worldwide revenue. Based on their 2023 revenue, that would be over $38B.</p>

<p>We’ll see how it plays out, but the mere <em>threat</em> of a fine that large, coupled with the no-nonsense process the DMA lays out for enforcement, may coerce Big Tech into compliance. In turn, this could make the DMA the de facto regulatory framework worldwide. Firms would otherwise need special variants of their platforms for each regulatory environment. Perhaps they will do so, but, in many cases, it’s hard to even know what that would mean.</p>

<p><strong>Update:</strong> They’ve also <a href="https://apnews.com/article/microsoft-teams-eu-european-union-antitrust-26d11ada00f504d537d1b054dd6f6bbf">charged Microsoft</a>.</p>

<p><strong>Update:</strong> I don’t know if the EU is entirely serious, but <a href="https://arstechnica.com/gadgets/2024/07/report-apple-approves-epic-games-store-on-ios-in-europe/">Apple sure seems to think so</a>.</p>

<p><a href="https://blog.pandastrike.com/tag:EU" class="hashtag"><span>#</span><span class="p-category">EU</span></a> <a href="https://blog.pandastrike.com/tag:DMA" class="hashtag"><span>#</span><span class="p-category">DMA</span></a></p>
]]></content:encoded>
      <guid>https://blog.pandastrike.com/eu-dma-is-for-real</guid>
      <pubDate>Wed, 10 Jul 2024 20:21:49 +0000</pubDate>
    </item>
    <item>
      <title>CISA Open Source Safety Analysis</title>
      <link>https://blog.pandastrike.com/cisa-open-source-safety-analysis?pk_campaign=rss-feed</link>
      <description>&lt;![CDATA[The US cybersecurity agency analyzed popular open-source projects and found that many widely used projects still contain a lot of unsafe code.&#xA;&#xA;!--more--&#xA;&#xA;Projects still relying on unsafe code include: Linux, Chrome (Chromium), Firefox (Gecko), Node, PHP, Python, Redis, Postgres, MySQL, OpenSSH, and OpenSSL.&#xA;&#xA;On the other hand, this is nevertheless a huge improvement over a decade ago. Projects that rely mostly on safe code include Ruby and Go. There are others, but they mostly make the list because they’re built on something else, which is often unsafe (ex: WordPress on PHP).&#xA;&#xA;For our work, we rely on Node and Linux. Linux will never be safe, but someone may rewrite the majority of Linux in Rust (Runix?!).&#xA;&#xA;Alternatives to Node already exist, but one relies on V8, which is not memory-safe (it’s part of Chromium, in case you’re wondering why it’s not on the CISA list). Bun relies on Zig, which is less safe than Rust:&#xA;&#xA;  Zig removes some of the most egregious footguns from \[C\], has better defaults, makes some good practices more ergonomic, and benefits from a fresh start in the standard library (eg using slices everywhere). But it does not nearly approach the level of systematic prevention of memory unsafety that rust achieves. It is still trivial to violate memory safety in zig.&#xA;&#xA;Of course, we also rely heavily on Amazon’s stack, and determining its safety is non-trivial. Fortunately, there is evidence that Amazon takes memory safety into account. For example, DynamoDB is based on Java, a memory-safe language. Of course, Java itself is typically implemented in C, which is not.&#xA;&#xA;#CISA #Safety #Security]]&gt;</description>
      <content:encoded><![CDATA[<p>The US cybersecurity agency analyzed popular open-source projects and found that <a href="https://www.cisa.gov/sites/default/files/2024-06/joint-guidance-exploring-memory-safety-in-critical-open-source-projects-508c.pdf">many widely used projects still contain a lot of unsafe code</a>.</p>



<p>Projects still relying on unsafe code include: Linux, Chrome (Chromium), Firefox (Gecko), Node, PHP, Python, Redis, Postgres, MySQL, OpenSSH, and OpenSSL.</p>

<p>On the other hand, this is nevertheless a huge improvement over a decade ago. Projects that rely mostly on safe code include Ruby and Go. There are others, but they mostly make the list because they’re built on something else, which is often unsafe (ex: WordPress on PHP).</p>

<p>For our work, we rely on Node and Linux. Linux will never be safe, but someone may rewrite the majority of Linux in Rust (Runix?!).</p>

<p>Alternatives to Node already exist, but one relies on V8, which <a href="https://v8.dev/blog/sandbox">is not memory-safe</a> (it’s part of Chromium, in case you’re wondering why it’s not on the CISA list). Bun relies on Zig, which is <a href="https://www.scattered-thoughts.net/writing/how-safe-is-zig/">less safe than Rust</a>:</p>

<blockquote><p>Zig removes some of the most egregious footguns from [C], has better defaults, makes some good practices more ergonomic, and benefits from a fresh start in the standard library (eg using slices everywhere). But it does not nearly approach the level of systematic prevention of memory unsafety that rust achieves. It is still trivial to violate memory safety in zig.</p></blockquote>

<p>Of course, we also rely heavily on Amazon’s stack, and determining its safety is non-trivial. Fortunately, there is evidence that Amazon takes memory safety into account. For example, DynamoDB is based on Java, a memory-safe language. Of course, Java itself is typically implemented in C, which is not.</p>

<p><a href="https://blog.pandastrike.com/tag:CISA" class="hashtag"><span>#</span><span class="p-category">CISA</span></a> <a href="https://blog.pandastrike.com/tag:Safety" class="hashtag"><span>#</span><span class="p-category">Safety</span></a> <a href="https://blog.pandastrike.com/tag:Security" class="hashtag"><span>#</span><span class="p-category">Security</span></a></p>
]]></content:encoded>
      <guid>https://blog.pandastrike.com/cisa-open-source-safety-analysis</guid>
      <pubDate>Mon, 08 Jul 2024 19:00:00 +0000</pubDate>
    </item>
    <item>
      <title>RIAA Sues AI Music Companies</title>
      <link>https://blog.pandastrike.com/riaa-sues-ai-music-companies?pk_campaign=rss-feed</link>
      <description>&lt;![CDATA[The RIAA joins the New York Times in claiming that training LLMs does not constitute “fair use.”&#xA;&#xA;!--more--&#xA;&#xA;I’m not a fan of either organization, but these lawsuits seem to have merit and, in any event, raise legitimate questions about the relationship between AI and fair use. Cory Doctorow seems concerned that limiting fair use is at odds with stronger regulation and organizing, but that doesn’t follow: we can do both.&#xA;&#xA;#RIAA #AI]]&gt;</description>
      <content:encoded><![CDATA[<p>The <a href="https://www.theverge.com/2024/6/24/24184710/riaa-ai-lawsuit-suno-udio-copyright-umg-sony-warner">RIAA joins the New York Times</a> in claiming that training LLMs does not constitute “fair use.”</p>



<p>I’m not a fan of either organization, but these lawsuits seem to have merit and, in any event, raise legitimate questions about the relationship between AI and fair use. Cory Doctorow <a href="https://pluralistic.net/2024/06/21/off-the-menu/">seems concerned</a> that limiting fair use is at odds with stronger regulation and organizing, but that doesn’t follow: we can do both.</p>

<p><a href="https://blog.pandastrike.com/tag:RIAA" class="hashtag"><span>#</span><span class="p-category">RIAA</span></a> <a href="https://blog.pandastrike.com/tag:AI" class="hashtag"><span>#</span><span class="p-category">AI</span></a></p>
]]></content:encoded>
      <guid>https://blog.pandastrike.com/riaa-sues-ai-music-companies</guid>
      <pubDate>Fri, 05 Jul 2024 07:00:00 +0000</pubDate>
    </item>
    <item>
      <title>InterOp 2024</title>
      <link>https://blog.pandastrike.com/interop-2024?pk_campaign=rss-feed</link>
      <description>&lt;![CDATA[InterOp has led to a tremendous improvement in browser interoperability, particularly with newer features. So it’s great that they’ve set new goals for 2024. !--more-- Of particular interest, they’re beginning to turn their attention toward WASM:&#xA;&#xA;  As well as focus areas, Interop 2024 will also feature a new investigation into improving the integration of WebAssembly testing into web-platform-tests. This will open up the possibility of including WASM features in future Interop projects.&#xA;&#xA;#InterOp #WASM #OpenWeb]]&gt;</description>
      <content:encoded><![CDATA[<p>InterOp has led to a tremendous improvement in browser interoperability, particularly with newer features. So it’s great that they’ve <a href="https://hacks.mozilla.org/2024/02/announcing-interop-2024/">set new goals for 2024</a>.  Of particular interest, they’re beginning to turn their attention toward WASM:</p>

<blockquote><p><em>As well as focus areas, Interop 2024 will also feature a new investigation into improving the integration of WebAssembly testing into web-platform-tests. This will open up the possibility of including WASM features in future Interop projects.</em></p></blockquote>

<p><a href="https://blog.pandastrike.com/tag:InterOp" class="hashtag"><span>#</span><span class="p-category">InterOp</span></a> <a href="https://blog.pandastrike.com/tag:WASM" class="hashtag"><span>#</span><span class="p-category">WASM</span></a> <a href="https://blog.pandastrike.com/tag:OpenWeb" class="hashtag"><span>#</span><span class="p-category">OpenWeb</span></a></p>
]]></content:encoded>
      <guid>https://blog.pandastrike.com/interop-2024</guid>
      <pubDate>Wed, 03 Jul 2024 07:00:00 +0000</pubDate>
    </item>
  </channel>
</rss>