William Emeny

🚀 Just published: MNIST Digit Classifier Train a neural net to 98%+ accuracy in minutes, then deploy it with an interactive web UI. Hands-on ML project — from PyTorch to Flask. 👉 github.com/williamemeny/M… #MachineLearning #PyTorch

Ollama is now updated to run the fastest on Apple silicon, powered by MLX, Apple's machine learning framework. This change unlocks much faster performance to accelerate demanding work on macOS: - Personal assistants like OpenClaw - Coding agents like Claude Code, OpenCode, or Codex

This results in a large speedup of Ollama on all Apple Silicon devices. On Apple’s M5, M5 Pro and M5 Max chips, Ollama leverages the new GPU Neural Accelerators to accelerate both time to first token (TTFT) and generation speed (tokens per second). note: test was conducted on using Alibaba’s Qwen3.5-35B-A3B model quantized to nvfp4 and Ollama’s previous implementation quantized to q4_K_M using Ollama 0.18. Ollama 0.19 will see even higher performance (1851 token/s prefill and 134 token/s decode when running with int4).

Computer use is now in Claude Code. Claude can open your apps, click through your UI, and test what it built, right from the CLI. Now in research preview on Pro and Max plans.

Google open-sourced a time series foundation model. it works with any data without training. unlike traditional models, no dataset-specific training needed. TimesFM forecasts out of the box. trained on 100B real-world time-points across traffic, weather & demand forecasting.

She woke up to find her social media full of posts she never wrote. Then one of them claimed she gave birth at seventeen. You Posted It is a psychological thriller built on one brilliant fear: what if the version of your life online knew more than you did? amzn.to/4sASQ6n

If you like psychological thrillers with podcast culture, buried secrets, and that creeping “something is badly wrong here” feeling: The Missing Episode has a true-crime host, a cold case, dangerous questions, and a story that gets far too close to home. amzn.to/4sCw3Y4

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

Introducing Hark, an AI lab building the most advanced, personal intelligence in the world. We're creating intelligent foundation models paired with next generation hardware designed to serve as a universal interface between humans and machines. hark.com

Today I'm excited to introduce Hark, a new artificial intelligence lab building the most advanced, personal intelligence in the world We've been in stealth for 8 months, assembling one of the greatest AI and hardware teams on the planet I want to explain why I started Hark and what we're focused on I've spent the last 3 years working on the hardest AI challenge imaginable: giving AI a humanoid body. On the digital side, I've been using all the existing LLM chatbots - and I have to say, they feel incredibly dumb to me AGI, in the limit, should feel like a sci-fi movie. It should be able to listen and talk. It should have persistent memory and be highly personalized. It should see and touch the world. But we're far from this today We are crafting a new interface to AGI. Intelligence that lets you offload your mental workload into a system that begins to think like you and sometimes ahead of you We started Hark with one goal: build the world's most advanced personal intelligence - paired with next-generation hardware designed to serve as a universal interface between humans and machines hark.com

Subagents are now available in Codex. You can accelerate your workflow by spinning up specialized agents to: • Keep your main context window clean • Tackle different parts of a task in parallel • Steer individual agents as work unfolds

Today, we're taking Manus out of the cloud and putting it on your desktop. Introducing My Computer, the core feature of the new Manus Desktop app. It’s your AI agent, now on your local machine.

🇦🇺An Australian tech founder with zero biology background sequenced his dog’s tumor DNA, then used ChatGPT and AlphaFold to design a custom mRNA cancer vaccine. A month later, the tumors shrank by half. And this is just the start of AI medicine.

OpenClaw just got a lot cheaper to run. Alibaba Cloud dropped a Coding Plan that gives you 4 frontier models under one API key. Plug it straight into OpenClaw and you're done. Qwen 3.5-Plus. Kimi K2.5. MiniMax M2.5. GLM-5. 18,000 requests a month for just $10. In single subscriptions, you can swap the models and keep building. The interesting part isn't even the price. It's that they built this for developers already inside tools like OpenClaw, Claude Code, and Cline. The top engineers will quietly plug this in this week and say nothing. The rest will find out in 6 months when the cost gap is impossible to ignore. Setup takes 30 seconds. 👇

Today we're showing Helix 02 that can tidy a living room fully autonomously Figure is designed so when you leave the house, your home resets exactly how you like it

I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)

The Qwen 3.5 small model series is now available on Ollama.

The Qwen 3.5 small models are available on Ollama. All models support native tool calling, thinking, and multimodal capabilities in Ollama. 9B: ollama run qwen3.5:9b 4B: ollama run qwen3.5:4b 2B: ollama run qwen3.5:2b 0.8B ollama run qwen3.5:0.8b Model page, including the bigger Qwen 3.5 models: ollama.com/library/qwen3.5

The Qwen 3.5 small model series is now available ollama run qwen3.5:9b ollama run qwen3.5:4b ollama run qwen3.5:2b ollama run qwen3.5:0.8b All models support native tool calling, thinking, and multimodal capabilities in Ollama.

🚀 Introducing the Qwen 3.5 Small Model Series Qwen3.5-0.8B · Qwen3.5-2B · Qwen3.5-4B · Qwen3.5-9B ✨ More intelligence, less compute. These small models are built on the same Qwen3.5 foundation — native multimodal, improved architecture, scaled RL: • 0.8B / 2B → tiny, fast, great for edge device • 4B → a surprisingly strong multimodal base for lightweight agents • 9B → compact, but already closing the gap with much larger models And yes — we’re also releasing the Base models as well. We hope this better supports research, experimentation, and real-world industrial innovation. Hugging Face: huggingface.co/collections/Qw… ModelScope: modelscope.cn/collections/Qw…

With Qwen3-TTS Voice Design, you can shape tone and richness just using text. If you want more consistency: Generate the first segment with Voice Design, then continue with Voice Clone. Hope you enjoy it — would love to hear the fun voice styles you come up with. 🎙️