latest
My local models are worse than Claude at almost everything, but they still run most of my workflow for $0
There are some things that just don't need Anthropic's flagship model
I changed one llama.cpp flag, and my 8GB laptop GPU now runs a 35B model faster than I can read
8GB of VRAM isn't the bottleneck, your settings probably are
I replaced my $20 ChatGPT and Claude subscriptions with local AI to cut costs, but it backfired
Free AI models looked like a bargain until I factored in the real cost.
My local LLM didn't replace my $20 Claude subscription, it replaced everything I was never comfortable sending to the cloud
There are many things Qwen3.6 struggles with, but keeping my data safe isn't one of them
People are now vibe-coding their smart TVs to strip them of ads and unwanted features
The manufacturers aren't going to be pleased.
I gave Claude's cheapest new model and a local Qwen the same coding jobs, and one of them is clearly the better deal
Knowing the bug isn't the same as being able to fix it
I ditched $20 ChatGPT subscription and made Claude and Gemini work together for better results
What happens when you stop asking one AI to do everything and let two models collaborate?
AI coding agents are great at writing code and terrible at deleting it, and that's the real problem nobody talks about
AI coding agents can write code in seconds, but knowing what to delete is a whole different story.
I gave four local models the same messy spreadsheet, and not one found the cheapest graphics card on it
Local models can sometimes perform data analysis like storytelling
I run OpenCode with a different model for every job, and I haven't hit a rate limit yet
I stopped leaning on one model for everything
I tried Claude Code, VS Code, and a local LLM as a note-taking app, and only one worked
The least note-like tool made the best notebook
I replaced Google search with a local LLM for a week, and it changed how I find answers
The web, minus the scrolling and ads
A major decomp tool is shutting down after a tidal wave of AI: "The community aspect of decompilation has been broken"
The human element of decomps is being left in the dust.
I gave a local LLM read access to my files, and it finds things better than Windows Search ever has
My GPU is doing a job Microsoft wants an NPU for
I moved my local models from Ollama to llama.cpp, and the reason I stayed has nothing to do with speed
Convenience beats tokens generated per second, always
I trust AI code that crashes more than AI code that works, and 66% of developers seem to agree
79% of engineering leaders noticed the same thing.
I asked my AI coding agent to build the one feature every website is missing, and now I press Space on every link
No servers, no cookies, and nothing loads until you ask.
5 ways you're wasting Claude usage limits without even realizing it
Hitting Claude's usage limit too soon? You might be making these five common mistakes without even noticing.
Ollama is the easiest way to start with local AI, and it's also the first thing I'd replace
Ollama made local AI easy, but I found a reason to leave it behind.
I gave my local LLM my entire Steam library, and it recommends games better than Steam does
Can Gemma 4 figure out my gaming taste better than Steam?