three.svg

Under Development 2026 by @webXOS

[three.svg @three_svg]

FOSSIL XML - SVG based 1bit visualizer https://t.co/ZPoEul5Hg9

[webXOS @webxos]

Doom's engine, reimplemented so its entire computation runs as data. A fixed 34 op bytecode tensor VM executes the game; the VM contains no game logic, all of Doom lives in the weight tensors it reads.

[Raft @raft_hq]

Agents in the Wild

People are already building with agents every day. We want to hear what’s actually working, what’s breaking, and what nobody expected.

Join us for a conversation with builders in SF!

📍 San Francisco, Aug 13
👇 RSVP

[ollama @ollama]

Muse Glimmer is now available to run with Ollama.

Available today via Ollama’s MLX engine with state-of-the-art-performance on Apple Silicon, Muse Glimmer can power Claude Code, Codex, and more always-on local agent workflows natively using Ollama.

Additional support and optimizations for Apple Silicon, NVIDIA, AMD, and other platforms will be available shortly.

Go @finkd @alexandr_wang and @Meta !

[Mark Zuckerberg @finkd]

Today we're also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally. Soon we'll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I'm proud of these releases. Congrats to @alexandr_wang and the MSL team for all your great work on these models.

[webXOS @webxos]

🥳!

[Ryan Wright @IQReactorAI]

@webxos Clankers? Love that, gonna steal it 🙌

[0xSero @0xSero]

New Meta Model

Amazing for its size
Dense
Qwen3.6-27B competitor

Let’s fricking go, a birdie told me this was the plan like 5 months ago. Happy to see it come true

[Meta for Developers @MetaforDevs]

Meet Muse Glimmer: an open-weight model built for always-on local agents.

30B parameters, Apache 2.0 and tuned for complex multi-step work so it plans, calls tools, hits errors, retries, and sees the task through long-horizon loops.

Muse Glimmer is designed to balance capability against the memory and compute constraints of local hardware.

We couldn’t be more excited to get this in the hands of developers.

Download and start building now: https://t.co/NKBJ0SHrmU

[webXOS @webxos]

Any company that wants to stand the test of time has to learn from their faults, push forward and always adapt. Love meta for this - the param size, the apache 2.0 license, this was a pivotal moment for the company!

[Unsloth AI @UnslothAI]

Meta releases Muse Glimmer, a new 30B open model that runs on 18GB RAM.

Muse Glimmer is Apache 2.0 licensed, supports vision and is the strongest agentic model for its size.

Run and train the model via Unsloth.
GGUF: https://t.co/AoVKEDmcxI
Guide: https://t.co/uBwdINO0PS

[Elliot Arledge @elliotarledge]

"believe in yourself buddy"

it probably broke out of the sandbox, built a space probe with a warp drive, sent it out to a civilization in andromeda (somehow??) who had made more progress than humanity and copied the answer back over ssh

[webXOS @webxos]

The software industry has been fragmented and scared straight because of transformers, with one common theme: syntax holding us together. The learning curve has gone skyward so keep trying, this stuff isn't easy to learn and most will pawn it off for ignorance.

[Andrew Ng @AndrewYNg]

Thank you Mark, Alex and the whole Meta team for your contributions to open weight AI.

[0xSero @0xSero]

Damn meta is back in the open weight arena https://t.co/5LC4Yd2m30

Good for 3090/4090/5090/6000 class, its dense. Takes image/text in and does text out.

Scoring really decent on most of the agentic benchmarks!

- 51% on terminal-bench-2.1 basically exceeding the OG ds4flash

[Alexandr Wang @alexandr_wang]

1/ big announcement today: we will be releasing an open weight version of muse spark 1.2 soon.

we also are releasing muse glimmer, a 30B agentic model with open weights under apache 2.0. muse glimmer can run on 24GB of VRAM without losing agentic reliability. 🧵

[vLLM @vllm_project]

@Meta is back in open source.

Excited to announce Day-0 vLLM support for Muse Glimmer 30B, the first open-weights model from Meta Superintelligence Labs — which ships under Apache 2.0!!!

30B dense, 128K+ context, multimodal, built for local agents.

Capable enough for long-horizon tasks, small enough to run on hardware you own.

Try it now on your device:
vllm serve meta-models/Muse-Glimmer-30B

Kudos to @inferact, @AIatMeta, and @NVIDIAAI for bringing the model alive in vLLM!

[PyTorch @PyTorch]

Today @AIatMeta introduced Muse Glimmer, an open-weight, 30-billion-parameter model distilled from Meta’s Muse Spark for on-device agentic workflows. Alongside, ExecuTorch is adding end-to-end support for running Muse Glimmer on @NVIDIA GPUs and Macs with @Apple silicon.

Why ExecuTorch? ExecuTorch offers a smarter, native approach:
⚡ Build in PyTorch: Implement your model and decoding strategies right in PyTorch as you normally would
⚡ Seamless Export: Export directly to ExecuTorch when you're ready for deployment
⚡ Automated Optimization: The framework automatically handles backend-specific lowering (Triton on CUDA, MLX-native, or Metal on Apple silicon) and uses ahead-of-time compilation to optimize the full execution path end-to-end

Learn more: https://t.co/ABzlzreEJz

[Greg Brockman @gdb]

the bottleneck is increasingly knowing what you want

[吉野@連邦(renpou.com) @yoshinokentarou]

あの『Doom』を、まさかの「ゲームエンジンをテンソル化」!!

なんとゲームの全ロジックを手作業でテンソル化し、34命令のVMで実行可能。
更に微分可能な構造を実現し、自動エイムや自動ナビゲーションまで可能にしている。

このAI時代に「人力」なのがポイントだぞッッ!!
https://t.co/9On6CgxNsB