I finally found a local LLM I want to use every day (and it's not for coding)
…Related I replaced my local LLM with a model half its size and got better results — And it wasn't about the parameters I switched from a 20B model to a 9B…
Search
All sources
xda-developers.com
56
amd.com
9
phoronix.com
8
cnet.com
7
developer.nvidia.com
7
windowscentral.com
4
huggingface.co
4
wccftech.com
3
androidauthority.com
3
techcrunch.com
3
pcworld.com
3
androidpolice.com
2
Paper page - SleepWalk: A Three-Tier Benchmark for Stress-Testing Instruction-Guided Vision-Language Navigation
…AI-generated summary Vision-Language Models (VLMs) have advanced rapidly in multimodal perception and language understanding , yet it remains unclear whether they can reliably ground language into spatially coherent, plausibly executable actions…
Google's $100 AI Ultra is still overpriced, and here's what I use instead
…Then the hammer dropped. AI labs have moved to a "compute-based" usage model, which factors in complexity, tool calling, and how long your chats are into the usage cap. It's…
I replaced my entire AI prompt library with Fabric, and now I can't imagine going back
…Anthropic, and others D Only locally hosted open-source models Correct! Fabric supports multiple AI backends, including OpenAI, Anthropic (Claude), Google Gemini, Ollama for local models, and more. This flexibility means users…
Top stories
LM Studio now lets you use your iPhone to talk to local models on your Mac - 9to5Mac
5d agoI ran Gemma 4 and Qwen 3.5 for the same local tasks, and one pulled miles ahead
6d agoI ditched Claude for Obsidian and a local LLM, and miss it less than I expected to
1w agoI ditched LM Studio for an open-source alternative — and my local model is doing things it couldn't before
1w agoLinux May Drop Old Network Drivers Now That AI-Driven Bug Reports Are Causing A Burden
Linux May Drop Old Network Drivers Now That AI-Driven Bug Reports Are Causing A Burden Written by Michael Larabel in AI on 21 April 2026 at 03:45 PM EDT. 65…
AMD Ryzen™ AI Processors for Intelligent Computing
…Who they’re for Developers and AI enthusiasts Veteran creators and engineers Customers exploring local AI at scale Example workloads Running and fine-tuning AI models locally AI-assisted video, 3D, and…
Build AI Systems That Scale at Advancing AI
…personal AI workflows locally on a Ryzen AI PC How to connect local agents to frontier models deployed on AMD Instinct GPUs How to route tasks between local and cloud models based…
Discussions and forums
Show HN: Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks
Hi HN, I'm Antoine Zambelli, AI Director at Texas Instruments.I built Forge, an open-source reliability layer for self-hosted LLM tool-calling.What it does:- Adds domain-and-tool-agnostic guardrails (retry nudges, step e…
660
240
The Oats Protocol – Open Agent Tools for Local Coding Agents
Recently I was using functiongemma and watched it load and run local source code as a tool call without any training/tuning. A couple days later I got Qwen35 in Open-WebUI to use the "native" tool-calling. With Open-WebU…
5
Stop traumatizing AI into loops and turn hallucinations into an honest "I don't know!" by being NICE to them (Proof of Concept, Research, I don't want to sell anything)
!UPDATE!(20.05.2026) WE HAVE NEW NUMBERS FROM 1.500+ TESTS IT'S WORKING! check my update post https://www.reddit.com/r/LocalLLaMA/s/AyNOehjkYT Or the go straight to the my Github https://github.com/OttoRenner/Gentle-Codi…
Unpopular Opinion: The DGX Spark Forum community of devs is talented AF and will make the crippled hardware a success through their sheer force of will.
There is a lot of disdain for DGX Sparks here on the sub. And I get it. A lot of people say “It could have been great if it had been better memory bandwidth”, “SM-121 is a fake /second-class Blackwell chip” yadda, yadda.…
Show HN: Rotunda - A browser built for agents with simulated typing
Hi HN! Pierce here.Rotunda is a firefox fork primarily intended for agent use, which I’ve been hacking on nights/weekends.There was a [lengthy](https://news.ycombinator.com/item?id=48024859) discussion last week on how e…
13
5
Google Chrome is reportedly auto-installing a massive 4GB AI model without your consent
…The local model is not used by that AI, instead it powers features like “Help me write”. Hanff says that the silent installation of the model could potentially be illegal in several…
Ansys Discovery and AMD Expand Access to High Performance Simulation for More Engineers and Designers
…The local memory capacity that variable graphics memory provides makes that possible, even for the largest and most complex models. With AMD Ryzen™ AI Max+ PRO processors, engineering teams can process large…
Microsoft CEO says its AI data centers consume less water than your local diner
Microsoft CEO says its AI data centers consume less water than your local diner