local AI

A developer has released a project, Strata on GitHub, claiming to run the Qwen 3.8 Flash Next 125B-parameter model on a single consumer RTX 4090 graphics card at around 100 tokens per second. If the performance figures hold up under independent testing, it would make a very large language model practical on high-end home hardware without expensive datacenter equipment. The claim has drawn strong attention from the technical community, where commenters are expected to scrutinize the benchmarks, memory requirements and any quality trade-offs involved.

Mac users are sharing a method to turn off Apple Intelligence on macOS 27 and recover the disk space its local models consume. The approach, documented in a public GitHub repository, walks through removing the AI features and freeing several gigabytes of storage. Discussion centres on how much space the on-device models take and whether the trade-off is worth it for users who do not use the AI features.

A developer has released an open-source Claude Code skill that reads local transcripts and splits usage across the 5-hour and 7-day rate-limit windows by project and by skill. It also pinpoints the exact times the rate limits were hit, helping users see which projects consume their Anthropic quota and when they run out of allowance.

Salvatore Sanfilippo, the creator of Redis, has released ds4, a tool for running large language models locally on personal hardware. The project, available at dwarfstar.sh, is drawing attention among developers interested in local AI inference, with early reaction focused on the novelty of the well-known open source engineer moving into the LLM tooling space.

The Philadelphia Inquirer has built Scrape, an artificial intelligence tool designed to surface hyperlocal news coverage. The project, highlighted by the Lenfest Institute, aims to help the newspaper identify and report on community-level stories across Philadelphia that might otherwise go uncovered. It is drawing attention as an example of how local newsrooms are experimenting with AI to expand coverage amid industry downsizing.

A developer has launched Jevstiller, a tool that distills Jev into a local AI model, claiming a built-in disagreement bound that limits how far the distilled model can diverge from its source. The project was shared on Hacker News under the Show HN format, drawing moderate attention from the community. Details on the guarantee behind the disagreement bound are published on the project's site.

A developer has released Corral, an open-source tool aimed at killing every command that an AI coding agent starts. Shared on Hacker News as a Show HN project, it targets a growing concern as more developers hand command execution to autonomous agents and need a way to stop runaway processes quickly. Discussion centres on how to control and supervise agent-spawned commands on local machines.

A developer has shown how to use an iPhone as a second GPU alongside a MacBook, reporting that prefill times for running the Qwen 3.8 27B model are 29 to 44 percent faster. The experiment exploits Apple's unified connectivity to share compute between devices for local AI inference. Commenters are debating how the setup works, its practicality, and what it means for running large models on consumer hardware.

A reported claim suggests that restrictions from the Federal Communications Commission on robots could speed up the move toward locally run AI systems in robotics. If devices face limits on cloud connectivity or radio use, manufacturers may opt for on-device processing instead of remote services. The potential impact on the robotics industry is being discussed.

Developers have shared a new MCP (Model Context Protocol) server for scraping Google Maps, letting AI assistants query business listings, locations and reviews directly. The tool was posted on Hacker News under the Show HN label, where it drew modest attention. Discussion centres on its usefulness for lead generation and local data workflows, and on the legality of scraping Google Maps.

A company offered residents of a French town more than €80,000 in exchange for allowing the construction of an AI data center in their community. The inhabitants refused the offer, and the company has now taken legal action against the town. The case is drawing attention in France as debates grow over the local impact of AI infrastructure projects, from energy consumption to land use.

A developer has released a small open-source plugin for Claude Code that breaks down Anthropic's 5-hour and 7-day usage limits by individual project and skill. It works entirely from local transcripts, without sending data anywhere. The tool is being shared among developers interested in productivity tooling and in understanding how their AI coding allowance is actually consumed across different work.

An open-source tool has been released that lets Mac users delete around 12GB of data stored by Apple Intelligence on macOS. The tool targets the large local models and cached files Apple downloads to power its AI features, which many users do not use or want. Coverage is highlighting how much space the AI system quietly takes up and giving users a way to reclaim it.

Hackaday has published a new article looking at running AI directly on gaming PCs, focusing on the hardware side of the topic. The piece examines how consumer graphics cards and local machines can be used for AI workloads, a subject of growing interest among hobbyists and PC enthusiasts discussing local AI setups.

An open-source tool has been released that allows Mac users to delete around 12GB of local Apple Intelligence data stored on macOS. The tool, covered by The Verge, highlights how much space Apple's AI features occupy on users' machines and gives people a way to reclaim storage if they do not use the AI capabilities. It is drawing attention among Mac users and tech watchers discussing the storage footprint of on-device AI.

RemoveMacAI, an open-source utility, deletes Apple's built-in Apple Intelligence features from Macs, which reportedly have no off switch since macOS 27. By removing the local AI models, the tool frees up storage and system resources on machines running the features. It is drawing attention from Mac users looking for a way to fully disable Apple's AI software.

Gizmodo reports a hands-on look at what is described as the first 'wireless GPU', a device said to use Wi-Fi instead of local hardware or paid cloud subscriptions to handle AI computing tasks. Coverage frames it as a potential challenge to subscription-based AI services, though detailed performance claims remain limited to the outlet's initial impressions.

Executives from Anthropic, OpenAI, Google and Meta are set to testify at a New York City Council hearing examining the risks of artificial intelligence. The session puts some of the most prominent AI developers before local lawmakers as cities weigh how to regulate the fast-moving technology, and is drawing attention as a rare joint appearance by rivals on the question of AI safety.

An open-source tool has been highlighted for allowing users to run artificial intelligence models directly on their own machines, without relying on cloud services. Coverage in software media points to growing interest in local AI setups, driven by privacy concerns, cost savings, and independence from major providers, though specific details about the tool remain limited.

A 19-year-old founder has raised $11 million for Ghost, a startup building a $3,499 desktop computer designed to run personal AI locally. The funding round, reported by TechCrunch, puts a spotlight on the young entrepreneur and the growing market for dedicated hardware that lets users run AI assistants privately at home.

Beijing, Shenzhen, Hangzhou, and Chongqing have each released a wave of AI industrial policies for 2025-2026, framed in Western coverage as 'catching the AI wave.' But analysts argue the documents reveal a different logic: the policies function as a debt resolution strategy, with AI investment positioned as the main path to economic recovery.

The New York City Council has held a debate focused on the potentially catastrophic risks posed by artificial intelligence, according to Spectrum News NY1. The discussion brings questions about AI safety and regulation to the local government level, as lawmakers weigh what role cities should play in addressing threats from rapidly advancing AI technology.