Home

Inside JetPride: How JetBrains Employees Built an LGBTQIA+ Community

What makes a workplace feel welcoming? Sometimes it’s finding colleagues who share your interests. Sometimes it’s discovering people with similar experiences. Sometimes it’s knowing you can talk openly about your life, ask questions, share perspectives, and be yourself without giving…

dotInsights | June 2026

Did you know? The var keyword isn’t a keyword! It’s one of several “contextual” keywords in C#, and it only has special meaning when used to declare a variable. Try defining a class called var and see what happens to…

puppet-checkmk dusted off 🧹🦊

VoxPupuli’s puppet-checkmk [1][2] had become invisible on the Puppet Forge. [3] The Forge has a simple filter: No updates for a long time? Hide it! Since we hadn’t put out a release in a while, we got dropped off the…

How One Person Companies Make Money

How do One Person Companies make money? You charge ¥100 per AI copywriting project. Take 100 projects—you’ve still only earned ¥10,000. Why? Because you’re trading time for money. The breakthrough point is productization: turn your service into a product, build…

puppet-checkmk entstaubt 🧹🦊

VoxPupulis puppet-checkmk [1][2] war in der Puppet-Forge unsichtbar geworden. [3] Die Forge hat einen schlichten Filter: >FIXNUM Zeit kein Update erhalten? Ausblenden! Wei wir schon lange kein Release mehr veröffentlicht haben, sind wir aus der Forge gefallen. Wir haben das…

Best Python AI Frameworks in 2026

Whether you’re building chatbots, training computer vision models, or analyzing business data, choosing the right AI framework can make or break your project. Python has become the dominant language for AI and machine learning development, and the ecosystem of frameworks…

MPS 2026.1 Release Candidate Arrives

Before we finalize the 2026.1 release, we are presenting a release candidate build so you can preview the new features ahead of time. Download MPS 2026.1 Release Candidate now: DOWNLOAD MPS 2026.1 RC 1 The release candidate brings a single,…

Contribute to the State of PHP Survey

Together with The PHP Foundation, we’re embarking on an exciting expedition – the first annual State of PHP census that aims to tally the elePHPant population across the world. Come join us! Take the survey With the help of the…

The Benefits Of Cognitive Inclusion In UX Research

In the summer of 2024, I became co-chair of a working group of expert researchers who came together to determine how best to perform accessibility testing with people with cognitive disabilities. This was work I did for Fable, where I…

Micro Context Switching

NB, These are the mental rambling of an aged software engineer and AI-sceptic. Writing computer code is a little like writing a well considered letter In both cases we plan what we want to say, we have to consider that…

有人在拆 Transformer:Memory Caching 與 CTM 各拆走了一半

這篇要談的兩篇研究——Google 的 Memory Caching(RNNs with Growing Memory)和 Sakana AI 的 Continuous Thought Machine(CTM)——常被包裝成「Transformer 殺手」。不是。它們是兩篇研究論文,不是產品,也不是要取代 Transformer。把它們放在一起讀,真正的故事只有一句: Transformer 的 self-attention 把記憶(在上下文裡 recall)和計算(思考發生在 forward pass)綁在同一個機制裡,代價是 O(L²)。這兩篇各拆走一半。 Memory Caching 拆記憶那一半,CTM 拆計算那一半。理解了這個軸,後面所有細節都會歸位。 一個先講清楚的規矩:本文只採用原論文能支持的宣稱。二手文章裡那些「在 SWE-bench / GPQA 上如何如何」的數字,凡是回不到原論文的,一律不寫。這兩篇論文本身都沒有報告 SWE-bench 結果——把二手整理的 agent 數字寫成論文結論,是這個題目最常見的造假。 一、成本牆:融在一起的代價 先講為什麼有人想拆。 self-attention 可以理解成一種可微分的關聯記憶:每個…

Modern C++ Support in CLion: What’s New

Modern C++ makes advanced high-performance techniques more accessible, with features like compile-time computation, zero-overhead abstractions, and expressive template code. But as your codebase grows, your ability to use these techniques productively depends heavily on how well your tooling understands them.…