Tag
The author describes running Pixelflut, a collaborative pixel art project, on a 2010 TP-Link Archer C7 router using OpenWRT and cross-compiling Rust code, detailing the technical challenges and setup process.
A developer built a router to cut AI agent costs but found it only escalates requests, increasing spending; effective savings came from caching rather than routing.
TP-Link has unveiled its first Wi-Fi 8 router and mesh system at IFA 2026. The Archer 8 Ultra and Deco 8 Ultra offer high speeds and are set for preorders, though the Wi-Fi 8 standard won't be finalized until 2028.
Router by Ramp is a new product designed to save money on AI token usage by helping users manage and reduce API-related costs.
A blog post from ArcBox Labs describes a half-bridge offloading technique to overcome UniFi gateway PPPoE performance bottlenecks, especially on high-speed connections like their 5 Gbps office line.
This paper tests the assumption that attention weights reveal what a model actually depends on for its output, finding that attention and causal dependence often disagree. They propose using causal evidence sets obtained via intervention masking as supervision for sparse attention routers, achieving near-perfect accuracy on retrieval tasks where attention-distilled routers fail.
World Model Optimizer is an open-source CLI tool that optimizes AI models from agent traces and serves them with a router to maintain frontier model quality at reduced cost.
Jeff Geerling describes how to configure a backup WAN on an OPNsense router using a secondary gateway, allowing manual failover without physically reconnecting cables.
A practical guide to setting up a MikroTik router for a home lab, covering how to identify ISP connection types (IPoE vs PPPoE) and WAN IPv4 addresses (public vs CGNAT/DS-Lite) to configure the router correctly.
A security researcher discovered an unauthenticated remote code execution vulnerability in Motorola's MR2600 router, allowing attackers to upload and flash malicious firmware without authentication.
Discusses the common perception that MoE models with low active parameters are inferior to dense models, arguing that router effectiveness and architecture nuances matter.
Orchestrator (Skill) is a router that works across harnesses, allowing a coding agent to call, manage, and delegate to other agents. It provides an operational plane for modeling agents and intelligently routes based on user preferences and token limits.
OpenWrt One is an open hardware router that runs the OpenWrt firmware, offering an open-source alternative for networking hardware.
A tiny ~10K parameter router called tinyrouter learns which open model to use per question on MMLU, outperforming individual models by optimizing allocation.
A researcher discovered 7 bugs in llama.cpp's router that went unpatched for years; after raising the issue, they and 10 others were banned for using AI, yet maintainers later used AI themselves, sparking accusations of hypocrisy and unethical behavior.
Geeks discovered that the DJI 4G module can be repurposed as a driverless, full-band 4G wireless network card. When used with an overseas eSIM white card, it becomes a stable circumvention hotspot gateway, causing the price of used modules to rise.
FreeLLMAPI is an open-source tool that aggregates the free quotas of 16 LLM providers into a single OpenAI-compatible endpoint, with automatic routing and usage tracking, totaling about 1.7 billion tokens per month.
Dot Loom is an open-source orchestration layer that coordinates multiple AI models into a single inference system, supporting various providers. It acts as a router, drafter, verifier, and finalizer.
Elie Bakouch critiques Sakana AI's Fugu system as a closed-source orchestration layer over closed-source models, arguing it lacks transparency and true AI sovereignty, with technical limitations in routing and cost efficiency.
Comcast Xfinity launches same-day delivery of its Gateway router to new customers in select markets, allowing them to get online within hours.