𝔈𝔥𝔰𝔞𝔫

After almost 6 years, it's done. createlang.rs The journey ehsanmkermani.com/posts/2025-12-…

Portable or fast. GPU programmers have been forced to pick one for years. We don't think you should have to. The key: component boundaries that let you swap platform-specific implementations without touching kernel logic. Part 4 of the Structured Mojo Kernels series explains how: modular.com/blog/structure…

Zero day implementation for amd mi355x... modular.com/blog/day-zero-… Time has changed.

@PeteHegseth “Back to the Stone Age” was also the US slogan during the Vietnam War, which lasted 20 years and ended in defeat. Your original goal was liberating a civilization, not ending it.

The brave Iranian people have risked everything to rid themselves—and the world—of the blight of the Islamic Republic. Threats to send Iran “back to the Stone Age” are inhumane. This regime shows no regard for the destruction of Iran—in fact, it has driven it.

Gemma 4 is live on Modular Cloud, day zero, with the fastest performance on both NVIDIA and AMD. Our MAX inference framework delivers 15% higher throughput vs. vLLM on B200, and we’re the only inference provider to ship @googlegemma 4 on a framework we built ourselves. Two multimodal models live now: Gemma 4 31B (dense, 256K context) and 26B A4B (MoE, only 4B params active per pass). Both SOTA on Modular Cloud: modular.com/blog/day-zero-… Modular Cloud runs on MAX, our inference framework that unifies GPU kernels, graph compilation, and high-performance serving in a single hardware-agnostic stack. New weights to SOTA deployment in days, on two hardware platforms: modular.com/request-demo?u…

Massive win by the team! 🚀🚀🚀 modular.com/blog/day-zero-…

Google Deep Mind's impressive fully-open Gemma 4 is live day-zero on Modular Cloud. Modular provides the fastest performance on NVIDIA Blackwell and AMD MI355X, thanks to MAX and Mojo🔥. The team took this impressive new model to production inference in days.🚀

FA4 on Blackwell: 14 ops, 5 hardware units, 28 dependency edges. One wrong sync = a race condition sanitizers won't catch. We built a constraint solver that derives the pipeline schedule automatically, in Mojo 🔥 Part 1 of our series is out → modular.com/blog/software-…

130 lines instead of 870. That's the difference between our conv2d implementation on Blackwell and CUTLASS's. We broke kernels into three swappable pieces: one for moving data, one for coordinating the pipeline, one for compute. When you need a new kernel, you only change the piece that actually needs to change. Part 3 of our Structured Mojo Kernels series walks through the details: modular.com/blog/structure…

2 days ago we shipped image generation in <1s 🔥 Today, we make that <300ms 🤯 NVIDIA + AMD⚡️ Full demo below ⬇️

Zero celebrities speak out.

0-day book support 🤓

Bobby Green, just watched your this, the way your voice broke when you spoke about the execution of Iranian wrestler Navid Afkari, it broke me. You were right. It’s heartbreaking. Today Saleh Mohamadi, 19-year old wrestler executed for protesting. We need your voice again. 💔

The top three learning resources we're sharing with attendees at our @NVIDIAGTC booth this year: 1. Structured Mojo kernels blog series, 2. Mojo GPU puzzles, and 3. MAX LLM book. Links in thread 🧵 👇

@elonmusk Dear Elon, In these critical days, the people of Iran need internet access. Please help them stay connected as they struggle to reclaim their country from the rule of the mullahs. Stand with the Iranian people. Iran will not forget its friends.

Computer science is gradually returning to the domain of physicists, mathematicians, and electrical engineers as large language models automate much of what we currently call software engineering. The field’s center of gravity is shifting away from manual code writing and toward deeper theoretical thinking, mathematical insight, and systems-level reasoning.

GPU kernel development doesn't have to mean thousands of lines of interleaved pipeline, memory, and compute logic. Structured Mojo Kernels Part 2 walks through how separating those concerns into three components with hard boundaries simplifies the codebase, makes kernels easier to maintain, and keeps the same structure working across NVIDIA and AMD hardware generations. modular.com/blog/structure…

You shouldn't have to choose between peak GPU performance and code you can actually maintain. We built Structured Mojo 🔥 Kernels to fix that. Performance, usability, and portability without the tradeoff. 14k to 7k lines. ~1.8k TFLOPS held. We wrote a 4-part series on how. Part 1 is up modular.com/blog/structure…

Our fave slide: 2026 is the year of Mojo! 1.0 and compiler open sourcing are on the horizon. 🥳

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow. Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes. As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now. It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.