Zac Zuo

Hi everyone! TADA is one of the most interesting open-source voice releases I’ve seen in a while. The big idea is simple but brilliant: it aligns text and audio one-to-one, so the model never has to juggle that huge mismatch between text tokens and acoustic frames. That single change unlocks the three things people actually care about in TTS: way better speed, much longer context, and basically...

Hi everyone! VENTUNO Q emerging as one of the first really serious Qualcomm x Arduino platforms hits the edge AI and robotics wave perfectly. Combining up to 40 TOPS of local AI compute for vision, LLMs, and multimodal models with a dedicated STM32 real-time controller handles motors, sensors, and deterministic responses all on one board. Delivering a setup like this creates a massive advantage...

Hi everyone! Phi-4-Reasoning-Vision-15B is Microsoft"s new 15B open-weight model that makes multimodal reasoning feel much more efficient. It was trained on 200B multimodal tokens, handles high-res screens well, and stays direct on simpler tasks while switching into deeper reasoning when needed. Looks especially strong for math, science, and computer-use agents. Weights on HF.

Hi everyone! With no Phone (4) this year, the Phone (4a) Pro looks like the phone carrying Nothing in 2026. And it"s probably the most un-Nothing Nothing phone they've ever shipped :) Gone is the full transparent back. Instead, you get a slim full-metal aluminum unibody that feels properly premium. The iconic Glyph is now a bigger, brighter Matrix, and the transparent camera module is basically...

Hi everyone! AI2’s new Olmo Hybrid is the first hybrid 7B that clearly beats a pure transformer baseline (Olmo 3) in a fair fight. Same size as Olmo 3, trains at the same speed, but matches its accuracy with half the data and crushes long-context evals. The 3:1 RNN+attention mix just works. Super clean weights on HF. And you can even run the full model 100% locally in your browser on WebGPU!

"Book a demo" is essentially asking a high-intent buyer to endure friction just to see if your product is even a fit. I see a very clear split forming: AI is for instant product discovery and qualification, while human reps are for complex negotiation, security reviews, and relationship building. Applying the traditional Sales-Led Growth motion to early discovery is just too slow now. Love what...

Hi everyone! Have you seen this? https://www.youtube.com/watch?v=Ykiuz1ZdGBc When I watched the Unitree G1 doing full kung-fu and extreme parkour on Spring Festival Gala, my jaw literally dropped. Last year’s G1 was already impressive, but now Unitree and BIGAI have open-sourced the core control framework behind it: OmniXtreme. I immediately dug into the repo, and the moment I saw the extremely...

Hi everyone! The official Codex app for Windows is now in the Microsoft Store and it's built exactly for how most of us actually work. Previously you could run Codex through PowerShell or the VS Code extension, but this is the native desktop version we've been missing — secure sandbox, real PowerShell support, parallel agents with clean isolation, smooth diff review, and one-click editor...

Hi everyone! Step 3.5 Flash has been out for a few weeks and has quickly become one of the strongest open models for real agentic workflows. 196B sparse MoE with just 11B active per token, MTP-3 giving up to 350 tok/s on coding, solid 74.4% SWE-bench, and clean long-context handling. The OpenClaw support is seamless. Over the last couple days Step 3.5 Flash has even been #1 in daily OpenClaw...

Literally the best app to experience the latest @Qwen3 local AI models on your phone! 🚀