Symbio, a local AI assistant that learns from user corrections and fine-tunes itself without cloud or subscriptions, was released on GitHub by developer HuyEdits. The system runs on Apple Silicon usin…
TurboFieldfare, an independent Swift and Metal runtime, reports running Google's Gemma 4 26B A4B Mixture-of-Experts model with a resident memory footprint of roughly 1.9–2.1GB on Apple Silicon by stre…
Fast Company reports that free desktop apps like LM Studio and Jan.ai let users run small open-source AI models such as Llama 3.2 3B, Mistral 7B, and Phi-3 entirely offline on Mac or Windows PCs, ensu…
On July 30, an open-source project ran Kimi K3, a 2.78-trillion-parameter model that requires 1.42 TB of storage, on a MacBook Pro with 64 GB of RAM at 0.3 tokens per second, demonstrating that memory…
Researchers at UC Berkeley Sky Lab extended the K-Search evolutionary kernel optimization framework with a CUDA-to-MLX translation layer, enabling automatic transfer of GPU kernel expertise from NVIDI…
Google added voice-driven writing and editing to its Gemini macOS app on July 29th, giving users an Fn-key shortcut called Speak to Window that can place dictated text, rewrites, file summaries and ge…
A developer has released TurboFieldfare, an open-source Swift and Metal engine that runs Google's Gemma 4 26B-A4B instruction-tuned model in about 2 GB of RAM on any Apple Silicon Mac, including 8 GB …
Researchers at UC Berkeley Sky Lab extended the K-Search evolutionary kernel optimization framework with a structured CUDA-to-MLX translation layer, enabling automatic transfer of decades of CUDA kern…
AutoClaw, a new desktop AI assistant, combines office automation, browser automation, and IM workflows into one app, allowing users to generate documents, analyze data, manage social media content, an…
Developer Arthur Ficial released apfel-clip, a free macOS app that uses Apple's on-device AI model to transform clipboard content via a hotkey (⌘⇧V) without sending data to the cloud. The app detects …
Epilude released Model 4.1, a refined on-device cleanup model for Local Mode on Mac, that reduces critical errors from 4 to 3 and passes 79 of 90 internal benchmark scenarios, up from 78. The model is…
An engineer investigating speculative decoding speedup found that the algorithm's performance ceiling is determined by acceptance rate and cost ratio, not just hardware overhead. Testing on Apple Sili…
Palmier Pro, an open-source native macOS video editor built with Swift and Metal, has been released on GitHub, exposing its GPU-accelerated timeline pipeline for programmatic and AI-driven control. Th…
A user reports that threadfork, a local AI notetaker for Apple Silicon, enables secure meeting transcription without sending data to third-party servers, addressing privacy concerns for legal and cons…
A new LLM inference runtime called BaseRT achieves up to 6.4x faster local inference on Apple Silicon than llama.cpp by writing directly to Apple's Metal GPU API, skipping intermediate frameworks like…
ExecuTorch's experimental MLX delegate, released in May 2026, enables PyTorch models to run on Apple Silicon GPUs. Testing Qwen3-0.6B on an M1 Max, MLX INT4 achieved 188.9 tokens/s, 4.52x faster than …
Rondine, a tool that detects hardware and configures inference engines, enables running GLM-5.2, a 744B-parameter Mixture-of-Experts coding model requiring approximately 245GB of memory, locally on a …
Rondine, an open-source hardware-aware control plane, detects the host machine, recommends fitting models, builds an optimized serving plan, downloads weights, and starts an OpenAI-compatible local se…
Threadfork, a new AI meeting notetaker that runs entirely locally on Apple Silicon Macs, has been launched by developer Navneet Nair. The tool captures microphone and system audio without a meeting bo…
PyTorch 2.13, released July 8, brings FlexAttention to Apple Silicon with up to 12x speedup on sparse attention patterns, such as a 32,768-token sequence with a 256-token sliding window (35ms vs 431ms…