Why is it hard to write peak-speed GPU code? By Wayne Joubert johndcook.com/blog/2025/08/2…

As an SWE, You should at least have a concept of a plan how to build something like an OS, compiler, linker, browser, web app, mobile app, video game, or VM, From scratch.

مثلا لو عايز اعمل compiler هكتشف اني محتاج اذاكر algorithm زي الgraph coloring عشان اعمل assign للvariables لعدد محدود من الregisters, كده بقا عندي سبب لمذاكرة الgraph coloring. لكن لو قولتلي اذاكر الgraph coloring كجزء من رودماب للCP؟ غالبا مش ههتم.

🎓 MIT's Performance Engineering of Software Systems course is fantastic if you are interested in performance. Topics include perf analysis, algorithmic techniques for high perf, instruction-level & caching optimizations, parallel programming, and more. ocw.mit.edu/courses/6-172-…

This example was built with Meson. Apart from the usual project() + executable(), adding CTRE is as simple as:

Conclusion? I don't know. Either he has something, ran out of time, panicked, created the whole project with an LLM or he had a bet with his friends like "wanna bet I can get *two* talks accepted with pure AI slop?". In the latter case, he won.

"A subgroup H of a group G is normal if for every pair of elements a and b such that ab is in H, ba is also in H."

The code for this project is full of emojis, nopped implementations of stuff with the comment "in a real implementation...", the SAT solver just returned SAT, the anti-debug patcher just returned true without doing anything, all sorts of stuff. We knew at the talk this was BS,

If you’re a C or C++ developer, having uv makes life easier

Granovetter’s Threshold Model Volterra-Lotka Heating and Cooling Ising Model Polya’s Urn/Kirman-Folmer Schelling’s Chessboard Sandpile Model Logistic Growth and exponential growth Concentration of measure in the real world Emergence processes of heavy tails Supply and demand

Why haven't quantum computers factored 21 yet? algassert.com/post/2500

x.com/i/article/1964…

Would you hire a developer who uses a constraint solver to solve a hard leetcode problem ? I would. And Hillel Wayne has a few very good examples. (link: 👇)

Rust is not magic. It takes skill and acumen to make good software.

Because the one thing that I really needed was more daily spam in my inbox.

11,000 lines of code is nothing. I reckon slack or teams are not less than 700,000 lines of code.

got a cheap e-ink display, coded up a dashboard, and installed a weather station on the roof. first rain today and seeing the rainfall totals update in real time is pure nerd dopamine

Barbarians at the Gate is a very interesting new paper, with some exciting results showing the potential for AI in systems research. But I think the authors aren't quite asking the hardest problem about where this takes systems as a field. I wrote a new blog post about it.

😊

Mohammad Alaggan, Ph.D. دكتور نبيل العجان وأنا طالب في الكلية كان بيقولنا إنه عمل كده فعلاً وعمل OS لما كان بيدرس المادة. ولما كان بيشرح لنا اننا بندرس المادة غلط كنت بستغرب وبحس إنه في كوكب تاني بس لما اتخرجت نزلت ارض الواقع ولاقيت الشباب اللي اتعلموا بجد عارفين كل ده ودرسوه في الجامعة وبقى