Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
Hello, HellGatesPostmortem of a year-unsolved gate-level crackme that GPT-6 cracked in minutes
xutaxkamay recounts designing HellGates—a VHDL custom CPU with obfuscation, anti-tamper, and anti-debug that stumped humans and LLMs for a year—until GPT-6 solved it in under half an hour, then walks through what actually broke.
25 min · 5,662 words
the senior engineer death spiralon proving yourself, burning out, and being a good teammate
Sunil Pai names a common failure mode for engineers in new senior roles: cosplaying harder ambition, disappearing into secret overwork, burning out, then damaging trust. His antidote is to drop a level—become the greatest teammate first.
4 min · 993 words
Ben Swerdlow benchmarks Codex, Claude, and Grok agents across 171 StarCraft: Brood War matches—leaderboards, APM, cost per game, and analysis showing none played beyond beginner while Codex Astra led.
9 min · 2,151 words
I built non-autoregressive decision models with RL a year ago
Convai Innovations’ Nandakishor recounts building Laya—a ~33ms multilingual non-autoregressive decision engine with calibrated probabilities—via RLCD a year before frontier labs framed similar System One models as breakthroughs.
8 min · 1,852 words
TypeSafe's Jev AI Model in .NET: A Community SDK for Structured AI Output in C#
Laurent Kempé introduces TypeSafe’s Jev decision model and walks through a community .NET 11 / C# 15 SDK port so apps can get typed, structured decisions without brittle JSON parsing.
10 min · 2,298 wordsagent-assisted
Eric Bailey argues CSS-Tricks should become a co-op so the web keeps an independent, practitioner-owned home for frontend knowledge—after the site’s repeated near-deaths under corporate ownership.
3 min · 595 words
purplesyringa argues that the best technical writing captures the path from confusion to understanding—documenting the missing links experts forget—so others can learn faster than reverse-engineering opaque docs after the fact.
4 min · 921 words
How to Solve AI in Publishing with Three Easy(ish) Steps
Richard Pulfer proposes practical steps for the publishing industry to handle AI scandals and authorship concerns—focusing on disclosure, process, and standards rather than panic alone.
6 min · 1,475 words
How LLMs Actually Work: A Practical Guide for Product Managers
Abhishek Jaiswal explains tokens, transformers, attention, RAG, inference, and agents in practical PM language—so product leaders can make better build-vs-buy and quality decisions without becoming ML researchers.
20 min · 4,488 words
Three levels of noticingI see, but I do not observe
Kabir Khandpur sketches three levels of noticing—paying attention, recognizing patterns, and acting on what you observe—using birdwatching and everyday work as practice for seeing what was always there.
4 min · 906 words
Your dashes suggest which model you're copy-pasting from
Will Keleher spends $2.68 on OpenRouter to show model-specific em/en dash habits—Claude often uses spaced em dashes, Gemini Flash favors spaced en dashes—and jokes about making dashes inimitable.
6 min · 1,301 words
Benchmarking LLM Inference at Scale with AIPerf
NVIDIA introduces AIPerf, the GenAI-Perf successor: a multiprocess LLM inference benchmarker that avoids client bottlenecks at high concurrency, with flexible load shapes, trace replay, and production-scale measurement guidance.
7 min · 1,661 words
Saving another 100TB of RAM with math (and Rust)
Cloudflare explains how math-heavy redesigns and Rust in Pingora cut another ~100TB of RAM across their global network by shrinking hot in-memory structures without sacrificing correctness.
14 min · 3,157 words
Messi or Ronaldo? what 79 AI models think
We asked 79 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 96% picked Messi. See every answer, who searched the web, and who dissented.
13 min · 2,932 words
Roosters vs Sharks, NRL semi-final: 73 AI models predict the result
We asked 73 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 90% picked Sydney Roosters. See every answer and who dissented.
14 min · 3,151 words
Brighton vs Arsenal: 77 AI models predict the result
We asked 77 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 87% picked Arsenal win. See every answer and who dissented.
14 min · 3,234 words
Mathematicians Build Long-Awaited Graph SandwichThe proof of a decades-old conjecture has given researchers a new way to understand complex networks.
Paulina Rowińska explains how Behague, Iľkovič, and Montgomery completed Kim and Vu's sandwich conjecture, letting hard properties of random regular graphs follow from easier binomial-graph results.
7 min · 1,621 words
USRA Contributes Planetary Science Expertise to NASA-IBM Lunar Foundation Model
USRA describes an open-source NASA–IBM lunar foundation model that fuses diverse Moon datasets to support scientific analysis, including ice-prospectivity patterns near the poles.
3 min · 792 words
Custom Purchase Flows and Paywalls with RevenueCat on Android
A hands-on Android guide to building custom purchase flows and paywalls with RevenueCat’s Purchases SDK—PurchaseParams, callbacks, coroutines, and UI ownership.
19 min · 4,405 words
Auditing in the age of (good enough) AI
Trail of Bits on how “good enough” AI changes security auditing: what models help with, where they fail, and how audit practice should adapt.
9 min · 2,018 words