Valery Meleshkin 🕊💙💛

🇯🇵 We're launching @langfuse Cloud Japan data region today. Happy about committing further to this market. It's been rewarding to spend time here & partner with our amazing Japanese community (@langfusejp) and customers. Very excited to keep pushing. What else can we do for you? + still in Tokyo until Friday, DMs open!

Langfuse at Google Cloud Next Want to get to know the team, get to know the product and latest features, and chat about roadmap? Find us at the @ClickHouseDB booth #2917 @marcklingen, @AkioNuernberger, and @lotte_verheyden will be there

@softwarevlogger @gamer3dx хорошая игра. женщинам тоже можно играть за бородатого папу.

@gamer3dx Какой мужчина, а

Pragmata что-ли по фуллпрайсу купить...

@almaleksia говорят был рекордно сухой апрель. так что это неплохо.

Над Берлином нависла эта черная дыра и не двигается

Hegel is an open source property-based testing library for every language, based on, and brought to you by some of the folks behind Hypothesis -- the most widely-used, and arguably the best, property-based testing tool in the world today. We're serious when we say "every language": we released Hegel-rust a couple of weeks ago, and Hegel-go is on available on github today.

PostgresBench is live. An open and reproducible benchmark for Postgres services. Public results, transparent methodology, and open to contributions. We tested 5 providers to start, including Postgres by ClickHouse. Check the results and submit your own 👉 clickhou.se/3NRzx9X

In depth blog on how to write table providers for @ApacheDataFusio : datafusion.apache.org/blog/2026/03/3…

llm-as-a-judge evaluators now support categorical scores. instead of only numeric outputs, define categories like `correct`, `needs_follow_up`, or `blocked` and use them for filtering, dashboards, and routing. langfuse.com/changelog/2026…

@damirca 😟тебе же нравились красивые машины

Выбрал панораму, подогрев руля, камеру 360, ассистент умный, диски, цвет, heads-up display, музыку добавил хорошую и вот. Это пиздец, надо ждать 3 года пока появятся трехлетние такие за 40-50к.

Ran autoresearch on the Langfuse skill, it's the closest demo of the alignment problem I've personally experienced. It still improved the skill though :)

🇪🇺 🏆✨ Tears of joy: EU Parliament buries #ChatControl - on April 4, tech giants must stop indiscriminate scanning of our chats! A historic day for democracy and privacy! THANK YOU for your incredible resistance! ❤️ 🔥 But beware: they will keep trying... ⚔️

🇪🇺They’ve forced a re-vote: tomorrow at 11:00 the European Parliament votes again on #ChatControl mass scanning. Liberals are divided, Socialists&ECR voted with EPP: mepwatch.eu/10/vote.html?v… A quick phone tomorrow call might still change minds: 👉 fightchatcontrol.eu

🔥 Still looking to hire 2-3 Senior SWEs in the US for @ClickHouseDB DM me your CV + impressive work you've done

@mgill25 I see profiling is an experiment that either proves or disproves a hypothesis. To form a hypothesis one needs to reason first 🤷

Found this gem of a comment on HN: "Nobody should listen to anyone who speaks about performance in terms of reasoning about a system instead of profiling it."

Using Claude Code, I deployed another distributed systems algorithm visualization. Interactive single-page Paxos consensus tutorial game. Players try (and fail) to break the safety invariant by killing and delaying nodes. What is your high score?

langfuse just got much faster. read about all changes and how to enable the upgrades here: langfuse.com/docs/v4

@MDereviannykh @EgorBarsukov С одной стороны впечатляет. с другой: они это делают настолько поздно в пайплайне, что креативный контроль вышел из чата. Из всего что показали только на Starfield можно смотреть без моментального отторжения.

@EgorBarsukov Делает. Текущие requirements граф движков приводят к костам на создание ААА игры под 200-400 млн долларов в среднем 💁

Как я уже намекал ранее, геймдев будет все больше уходить от fine grained методов создания миров и итоговой картинке, к более генеративной. В особенности геймдев т.к. у неё фактически фин. кризис Конечно же причины не только в $. Скорее сложность разрешение фунд. проблем графики

btw, for anyone that needs to hear it the 3D printed 6+1DoF arm files are open source now. you can literally print this stiff beast on your bambu p1s, zero milling involved!! super stiff and smooth for precise teleoperation and deploying policies. github.com/robertorobotic…

I've built a new JavaScript runtime that runs inside the BEAM. Every JS runtime is a GenServer with its own OS thread. No JSON anywhere — JS objects map to BEAM terms natively through a lock-free queue. What makes it different from running Node/Deno/Bun alongside Elixir: → JS runtimes live in supervision trees. They crash, restart, recover state — standard OTP → fetch() goes through :httpc. WebSocket through :gun. crypto.subtle through :crypto. BroadcastChannel through :pg — works across a cluster → The DOM is lexbor (C library). JS renders into it, Elixir reads it directly — no serialization, no re-parsing → Workers are BEAM processes. They get preemptive scheduling for free → TypeScript toolchain (OXC) and npm client built in — no Node.js on the machine at all Full control over the JS layer: parse ASTs, bundle imports, transform TypeScript, minify — all from Elixir via OXC NIFs. Use cases: — SSR with Preact/React into native DOM, Elixir reads the tree — Sandboxed user-defined business rules with memory limits, timeouts, and a controlled API surface — Parallel Workers that compute and broadcast via distributed process groups — Evaluating or bundling TypeScript without any external toolchain — Running npm packages inside the BEAM Still a research project in early beta. Covered with tests including Web Platform Tests ports, but expect rough edges. github.com/elixir-volt/qu…

monte-carlo path tracing performs iterative denoising diffusion model does it as well the authors bridge the two: maps MC noise → diffusion noise + correct timestep so diffusion model can finish rendering as a learned denoiser