luminousmen

Hello wonderful people! I’m thrilled to announce that my new book, “Grokking Concurrency,” has officially hit the shelves! You can find it here: manning.com/books/grokking… I could really use your support in spreading the word!

Made this for engineers who aren't sure if they're using AI correctly or just laundering their thinking. Full rant on my Substack: luminousmen.substack.com/p/stop-feeding…

Adding this to my bookmarks bar, just in case: yousentmeaislop.com

Then it breaks. And 9 out of 10 times, it breaks because of memory. Full blog post: luminousmen.substack.com/p/deep-dive-in…

Spark is powerful. It scales. It's fast out of the box. And yeah, the defaults are surprisingly decent - until your dataset grows, your joins get messy, or you start mixing Scala with PySpark and Arrow and some eager ML engineer starts throwing 200MB Pandas UDFs at the cluster.

Sounds about right

👉 [FREE] Join our community of memeologists and get curated, frontline memes delivered directly to your inbox. Subscribe: luminousmen.substack.com/?showWelcome=t…

ONLYFANS could be the most revenue-efficient company on the planet, beating Nvidia, Meta, Tesla, and Amazon - powered by ass, not AI.

Don't give AI agents any more power than you'd hand to a junior engineer with near-root access Link: promptarmor.com/resources/goog…

If you're experimenting with Antigravity or any similar agent-driven development tools, keep the following in mind: - Lock down access to secrets - Audit what capabilities your agents actually have - Treat AI agents like remote developers

Security researchers at PromptArmor have discovered a critical vulnerability in Google Antigravity - Google's new AI-powered IDE that uses Gemini-based agents.

"First thing I did at Bell Labs? Got an alligator for the office." A 4.5-hour interview with Ken Thompson - the guy behind Unix and C 🔗Link: computerhistory.org/blog/a-computi…

In my latest post I break down what actually happens when you hit Run in BigQuery - from Colossus and Capacitor up through slots, shuffles, and the Query Master. luminousmen.substack.com/p/bigquery-exp…

So when you cluster/partition well, you're not "making queries faster". You're giving the engine permission to do nothing on most of your data.

Most people treat BigQuery like a magic SQL endpoint. You write a query, hit Run, wait a few seconds... and a petabyte-sized answer pops out. If it's slow or expensive, the default reaction is: "I need more compute". That's backwards. #dataengineering

So an absolute "never" feels more like a stance than an argument. Yes - in their current form, LLMs aren't intelligent. But shutting down the conversation with a definitive "never" feels shortsighted at best. What do you guys think? 🔗 Link: futurism.com/artificial-int…

The real question is whether that counts as intelligence, or just a very sophisticated simulation of it. From a technical perspective, the architecture can evolve: LLM+world model+memory + planning could realistically form the foundation for something more genuinely intelligent.

The article claims that LLMs will never become "intelligent".