A developer published a video and written guide showing how to run local AI models on a 16GB M1 MacBook Pro using the mlx-serve command-line utility, which is optimized for Apple's Metal architecture.…
ImageGen, a native macOS app from developer ph1lb4, runs Alibaba's Qwen-Image 2.1 image model fully on-device on Apple Silicon, requiring about 33 GB of unified memory and roughly 35 GB of free disk, …
A developer benchmarked TensorFold, a speculative-decoding inference engine, on a MacBook Pro with an M5 Max and 128 GB of unified memory, finding it decoded Qwen3.8-27B at 220 tokens per second on a …
Apple's new Mac Studio with the M5 Ultra chip begins shipping today, and in benchmark testing it posted a Geekbench 6 CPU multi-core score of 37,870 and a Metal GPU score of 233,865, beating the M3 Ul…
A developer published a reusable Claude Code prompt and companion bootstrap script for provisioning a new Apple Silicon MacBook Pro as a terminal-centric, AI-native engineering workstation. The setup …
Users running Qwen 3.8 on high-reasoning settings report their hardware hitting 90°C, with one user noting an Asus Ascent running an Nvidia GB10 chipset shut down once from heat and that Nvidia chips …
Apple's $899 M6 Mac mini caps at 32GB of memory and 153 to 170 GB/s of bandwidth with no Thunderbolt 5, making it a poor local-LLM machine compared with the $1,699 M5 Pro Mac mini, which offers 307 GB…
Apple confirmed that macOS 27 Golden Gate will launch on September 14, alongside iOS 27, iPadOS 27, and other operating systems. The update introduces a new Siri app with revamped Siri AI, an AI-power…
Argonaut Labs AI's Deltafin fork runs the full, never-pruned 2.8-trillion-parameter Kimi K3 model on a MacBook Pro at 0.2901 tokens per second (3.447 seconds per token), streaming from four SSDs, achi…
SidePulse, a new hardware accessory from Inteliwear, displays AI agent status through RGB LEDs that fit into a MacBook Pro's SD card slot or connect via USB-C, priced at $69 for the Pro model with 8 L…
WARP, an embeddable inference engine written in C, now runs the full 2.78-trillion-parameter Kimi K3 model on a 64 GB MacBook Pro at about 0.6 tokens per second, and the 313-billion-parameter GLM-5.3-…
Apple has unveiled its M6 chip, built on a 2nm process by TSMC, featuring a 12-core CPU, dual 16-core neural engine, and a redesigned GPU with up to 12 cores and memory bandwidth of approximately 200 …
LM Studio's local LLM performance hinges on whether users run MLX on Apple Silicon Macs or GGUF on Windows GPU laptops, with Apple's unified memory enabling larger models while discrete GPUs like the …
Open-source models are catching up to closed-source ones faster, with Qwen3.8-27B matching Claude Opus 4.5 on coding benchmarks nine months after release, and a Claude-level model could run locally on…
An engineer rebuilt the execution backbone of their development workflow onto local large language models (LLMs) to eliminate recurring cloud costs. By purchasing an NVIDIA DGX Spark and combining it …
A developer built SSH Config Manager, a native macOS app that edits ~/.ssh/config without wrecking formatting, saves tunnels as presets, and opens tunnels in-process. The app embeds a keyword catalog …
Brent Ozar, a Microsoft SQL Server expert and podcast host, tested whether a fully local AI setup could answer an Office Hours question, using the MiniMax H3 image-to-video model and a local Qwen3.6-3…
A former Privacy International staffer replaced dozens of mobile apps with private Progressive Web Apps (PWAs) built using Anthropic's Claude Code, citing 20 years of privacy law violations by mobile …
DeepSeekV4SSD, an experimental app from developer yanun0323, streams routed experts from SSD to run all 284 billion parameters of DeepSeek-V4-Flash-0731 on an M-series Mac with about 30 GB of memory, …
Apple's M6 MacBook Pro is expected to ship in fall 2026, according to Bloomberg's Mark Gurman, with the M6 chip built on TSMC's 2nm process and featuring a redesigned GPU with up to 12 cores, updated …