ContentPosts from @preetypenny..
Discovery IconThat's all from @preetypenny — explore more posts below...
Link
@faun shared a link, 2 hours ago

So you want to parse a PDF?

Out of 3,977 real-world PDFs, 0.5% broke during xref pointer parsing. Not a huge number—unless you're the one parsing them. The top culprit? Junk data before the start pointer. Classic. Other file weirdness: broken xref tables, bad object offsets, and inconsistent xref chains...

Link
@faun shared a link, 2 hours ago

Anthropic says OpenAI engineers using Claude Code ahead of GPT-5 launch

Anthropic just shut the door on OpenAI, yanking access to theClaude Code APIafter spotting ChatGPT engineers poking around—likely prepping forGPT-5. Claude Codeisn’t just an internal toy. It’s a serious coding co-pilot, used in the wild by devs who want answers without babysitting a model. Market ..

Link
@faun shared a link, 2 hours ago

Scaling Netflix's threat detection pipelines without streaming

Netflix’s “Psycho Pattern” stitched togetherSpark, Kafka, and Airflowinto a relentless micro-batch pipeline. It tracked high watermarks for near-real-time threat detection—fast enough, sharp enough. Then came the Flink switch. Lower latency? Sure. But it missed the mark. Signal quality stayed flat...

Link
@faun shared a link, 2 hours ago

GitHub Copilot crosses 20M all-time users

GitHub Copilot just crossed20 million users. Five million joined last quarter alone. Enterprise usage? Up75%quarter-over-quarter. It’s now in the hands of90% of the Fortune 100, according to Microsoft. Here’s the kicker: Copilot’s AI coding biz is now bigger than all of GitHub’s revenue when Micros..

Link
@faun shared a link, 2 hours ago

My Functional Programming Awakening: Patterns I'd Been Using All Along

A dev takes functional programming from Python class to JavaScript land—with surprising wins. The usual suspects show up:closures,function composition, and some spicyparser combinators. But the real magic? Swapping out side-effect soup forpure functions,Result-based error handling, andhigher-order f..

Link
@faun shared a link, 2 hours ago

The many, many, many JavaScript runtimes of the last decade

JavaScript runtimes aren’t just multiplying—they’re splintering. Big engines likeV8,JavaScriptCore,QuickJS,Hermes, andSpiderMonkeynow sit at the core of purpose-built runtimes everywhere: cloud, edge, mobile, IoT, even smart TVs. Platforms likeCloudflare Workers,Deno Deploy,Bun,LLRT, andNativeScrip..

Link
@faun shared a link, 2 hours ago

2025 Stack Overflow Developer Survey

Visual Studio and VS Code continue to reign supreme, fending off AI IDEs in the Stack Overflow 2025 Developer Survey. AI-generated devs noted as time-consuming and lacking trust, while Microsoft tools still dominate in agentic AI with GitHub and ChatGPT. More to discover, as always, Stack Overflow D..

Link
@faun shared a link, 2 hours ago

GenAI vs. Agentic AI: What Developers Need to Know

Docker’s getting serious about agent-based AI. It just rolled out tools tailor-made for building modular, goal-chasing LLM systems. Model Runnerlets devs spin up LLMs locally—zero cloud, zero wait.Offloadtaps cloud GPUs when local ones tap out. And theMCP Gatewaypipes in external tools without duct..

Link
@faun shared a link, 2 hours ago

Building a Redis Clone from Scratch – In-Memory KV Store with TCP

A solo dev just spun up a public build of aRedis-style key-value store in Java—lean, thread-safe, and backed by a custom TCP server. Right now it handlesGET,SET, andDELETEover a socket-level protocol. No HTTP. No bloat. At its core: aConcurrentHashMapdoing the heavy lifting. Fast, in-memory, and de..

Link
@faun shared a link, 2 hours ago

I used NotebookLM to learn a new programming language, and it actually worked

A CS student taught themselvesSwiftusingNotebookLM, Google’s AI that sticks to sources you feed it. They pulled in handpicked docs, YouTube transcripts, and visual mind maps—all dropped into a custom notebook. No generic guesses. No hallucinated trivia. Just clean, source-grounded answers on syntax ..