asmi brings nvidia-smi to Apple Silicon

// 45d agoOPENSOURCE RELEASE

asmi brings nvidia-smi to Apple Silicon

asmi is an open-source Rust daemon for Apple Silicon Macs that exposes hardware telemetry, model inventory, process control, MLX server lifecycle management, RDMA topology, and MCP tools over a JSON API. It aims to be the missing `nvidia-smi` plus `systemctl` layer for local LLM rigs, from one Mac to multi-node clusters.

// ANALYSIS

This is the right abstraction for Apple Silicon inference: a lightweight control plane instead of another opinionated runtime. The useful part is not just visibility, but orchestration across the messy stack people already run.

–The daemon collapses a bunch of ad hoc macOS commands into one API surface, which matters once you move from a single laptop to a cluster.
–RDMA and Thunderbolt topology support is the differentiator here; that is the painful part most tooling ignores.
–MLX server lifecycle management makes it more than monitoring, since it can recover, reload, and manage serving state.
–The MCP server is a smart distribution play: it turns infrastructure state into agent-accessible tools.
–The product is strongest for local-LLM operators, not general Mac users, which keeps the scope coherent.

// TAGS

asmiopen-sourceinferencegpuapimcpautomation

DISCOVERED

45d ago

2026-04-24

PUBLISHED

45d ago

2026-04-24

RELEVANCE

9/ 10

AUTHOR

Street-Buyer-2428

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

NEWS36m ago

The rise of terminal-native agentic coding tools introduces major security vectors that developers must proactively defend against.

The rise of terminal-native agentic coding tools, highlighted by Anthropic's Claude Code, presents a dual-use challenge for developers. While these tools greatly accelerate software development, they also serve as a potential 10x entry vector for attackers. The underlying security risk stems from their ability to execute terminal commands, manage files, and interact with the local filesystem, which requires developers to implement strict safety practices and sandboxing to prevent unauthorized execution or compromise.

NEWS59m ago

New Claude checkpoints Fable, Fruitcake leak

New internal model checkpoints from Anthropic, labeled "Claude Fable 5" and "Claude Fruitcake EAP," have reportedly been detected in active testing. This development highlights Anthropic's efforts to bridge the capability gap between its public models and its rumored internal powerhouses like Mythos Preview, indicating that new commercial or early-access versions of their AI may be on the horizon.

RESEARCH1h ago

Autonomous AI agents cut labor costs 94%

A research paper analyzing Perplexity production data shows that autonomous AI agents significantly expand task complexity while reducing labor costs by up to 94 percent. The authors propose an economic framework where agents lower marginal execution costs, shifting human effort toward verification and strategy.