XIO

xWire

Posts from every feed QUEST has found across the web — blogs, news, podcasts, channels and more.

XDA - AI tools

The world’s best source for computing news, reviews, editorials guides, and more.

My local LLM didn't replace my $20 Claude subscription, it replaced everything I was never comfortable sending to the cloud

Over the past couple of months, I've experimented with the idea of reducing my dependence on cloud-based AI services for various reasons. For one, they can be expensive. For another, my personal and professional workflows involve a lot of information that I couldn't reasonably trust a cloud service with.

I remapped the Copilot key on my keyboard to open my local model instead

Microsoft Copilot hasn't got the best rep. What was designed to be a convenient feature to make AI more accessible for those less tech-savvy, it actually turned into something many (myself included) loathed. What's more is my current laptop, the trusty Lenovo Yoga Slim 7x Gen 9, a Copilot+ PC with a Snapdragon X Elite CPU, is primed for running local models. With PowerToys and a little tinkering, I managed to make this dedicated Copilot button on my keyboard more useful by loading a local LLM.

The Lucy AI Studio brings AI agents, local models, and private storage together

Everyone’s using AI these days, but despite powerful desktop hardware being readily available, almost all of that AI usage happens remotely in faraway data centers. This is a potential privacy issue, since you’re sharing personal data with someone else’s AI, and you’re at the mercy of the reliability and security of something beyond your control.

I gave Claude's cheapest new model and a local Qwen the same coding jobs, and one of them is clearly the better deal

Anthropic just made its smallest model a lot cheaper this month, so the timing might be a little awkward for anyone who just started running free models locally. I've been relying on Qwen 3.5 for anything coding-adjacent, so paying for a cloud model on top of that only makes sense if it does something a local one can't. Coding tasks seemed like the best way to test it since that's where small models tend to struggle.

I tried DeepSeek's agent harness, and it's not the Claude Code clone I expected

Agentic coding harnesses get a lot of launch hype, and when I saw DeepSeek Harness turn up in my feed, I knew I wanted to put it to the test. It runs DeepSeek V4.1 Flash by default and is the company's version of Claude Code, Pi, OpenCode, or any other agentic coding tool you could name. If using LLMs for coding has taught me anything, it's that they're only as good as the surrounding tooling, so I figured out a task, created two identical git repos, and let DeepSeek Harness and Claude Code…

My Proxmox server now runs a full dev environment for AI agents, and it's surprisingly easy to manage

I've never been the best coder, whether in college with Assembly or C++, or in any other language I've tried to pick up since. That's changing now that I can use agentic coding tools like Claude Code to explain things as I go, and I've wired them into local models to cut down on token costs. But I still have to set things up on every new machine, and that gets tiresome.

I tried replacing Photoshop, Lightroom, and Acrobat with free vibe-coded alternatives, and they're surprisingly good

Vibe coding is scary, don't get me wrong. As a computer science major, there's probably no one more terrified of what the job market will look like by the time I graduate, especially with how good AI coding tools are getting. However, I also see massive potential in these tools, and I see them opening up opportunities that would've been unimaginable just a few years ago.

I replaced Google search with a local LLM for a week, and it changed how I find answers

Searching with Google used to be the answer to everything. But these days, the answer sits somewhere below five sponsored results and an AI summary that's probably not accurate. Getting to the answer takes way more scrolling than the question actually deserves, and given user behavior, this usually means just settling for the easiest one, which might not be adequate.

I've spent over $3,000 on ChatGPT, Claude, and Gemini, and only one was worth the money

Something I remember particularly well about the early days of generative AI is how quickly people developed an aversion to the idea of paying for yet another subscription. I could understand why! We were already paying monthly for everything from streaming services to cloud storage, and suddenly, using a chatbot just to draft a quick email or (unfortunately) write an essay was going to cost another $20 a month.

I ruined my PC on purpose and gave an AI agent one hour to fix it

I've watched AI coding agents rewrite half a codebase in an afternoon, so I figured fixing a broken Windows PC would be easy. It's the kind of job most of us get roped into anyway, usually over the phone, usually for a relative, and usually while someone reads error messages aloud. So I broke a Windows 11 PC on purpose. Not my own, thankfully, but a Hyper-V virtual machine on my workstation, wrecked in eight different ways. Then I handed Claude Code a single prompt, an SSH connection, and one…

I gave a local LLM read access to my files, and it finds things better than Windows Search ever has

Most people remember files by what was in them rather than what they're titled, which is a problem when the thing you use to search needs the literal words. You'll have a rough idea of the note and remember putting it somewhere, but Windows Search wants the exact term you used six months ago. Type the wrong keywords and you might get a list of files that technically match, and probably just a bunch of web search results.

I moved my local models from Ollama to llama.cpp, and the reason I stayed has nothing to do with speed

Ever since I started hosting local AI models, Ollama has been my default way to run them. For the longest time, it gave me absolutely no room to complain, and I'd wager for most beginners that would be the case. It's not lost on me that local AI inference does seem to have a pressing friction problem, and that a clean GUI like Ollama's doesn't make you work too hard and is therefore less likely to push you back to cloud alternatives by taking care of the problems for you.

I stopped paying for Otter after setting up local Whisper dictation, and it understands my accent better

As a journalist, I spend a lot of time taking notes. Whether I'm outlining an article or gathering pointers for an in-depth guide, typing thousands of words a day gets tedious real quick. I've previously dabbled in Otter for transcription. It worked fine, but as soon as I started talking fast, it failed to capture my accent. So, I started looking for an alternative.