Claudy With A Chance Of Meatballs Edition
<div><br />
</div>
<h2>Top Story</h2>
<div>
<ul>
<li>[url=https://www.theregister.com/ai-and-ml/2026/08/27/hugging-face-offers-399-robot-duck-to-help-you-quack-the-ai-code/5293011]Hugging Face has announced a new robot duck.[/url] (The Register)<br />
<br />
$13 billion for this?</li>
</ul>
<div><br />
</div></div>
<h2>Tech News</h2>
<div>
<ul>
<li>[url=https://wccftech.com/intels-24-core-core-ultra-7-270k-plus-dips-below-200-for-the-first-time-since-launch-pressuring-amds-x3d-lineup/]Intel's 270K CPU - an honestly pretty good processor that you definitely shouldn't buy - is now down to $200, putting it at the launch price of its slower sibling the 250K.[/url] (WCCFTech)<br />
<br />
The 270K and its entire platform will be replaced by the end of the year, and its bargain price is dwarfed by the cost of memory right now.<br />
<br />
<br />
</li>
<li>[url=https://arstechnica.com/security/2026/08/claude-codex-and-hermes-installed-unowned-code-inside-corporate-networks/]AI coding tools like Anthropic's Claude and OpenAI's Codex installed unowned code inside corporate networks.[/url] (Ars Technica)<br />
<br />
Text files. These fuckers are talking about text files. Not programs of any kind.<br />
<br />
Now there is a real problem in that the text files tell other AI agents to do dumb shit, and the AI agents promptly do it, because they are being run by dumb people.<br />
<br />
Also the first comment on that article calls it out as AI slop itself, which I agree with. Ars Technica has been caught with their pants down <span style="font-style: italic;">and </span>on fire that way before. <br />
<br />
<br />
</li>
<li>[url=https://arstechnica.com/tech-policy/2026/08/new-twitter-launches-says-musks-x-gave-up-the-name/]A new Twitter has launched, claiming that X abandoned the trademark.[/url] (Ars Technica)<br />
<br />
Betting you can outlast a trillion-dollar company in court is certainly a choice.</li>
</ul>
<div><br />
</div></div>
<div><br />
</div>
<h2>Noodling</h2><img src="http://ai.mee.nu/images/MenoLang.png" width="100%" />
<div>Albeit a different kind of noodling. First project I've been working on was building out a practical version of the database I created a proof-of-concept for in Python years ago. LMDB, MessagePack, LuaJIT, and ZeroMQ did the heavy lifting, and it was actually <span style="font-style: italic;">fast </span>- 1GB per second of reads on a small VM - until I started on the indexes and then it all caved in.<br />
<br />
Now I'm working on replacing LuaJIT in there with a capability-based language that is simpler and more secure, again using a model I tinkered with years ago.<br />
<br />
It's AI... Antislop? What I'm doing is the exact opposite of Node.js at every level, and using AI assistance to pull it together.<br />
<br />
</div>
<div><br />
</div>
<h2>Musical Interlude</h2>
<div>[ytlite=GyWhDPbkh2c]</div>
<div><br />
</div>
<div><br />
</div>
<div><br />
</div>
<div><br />
</div>
<div><span style="font-style: italic;">Disclaimer: Database is did.</span></div>