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 been experimenting with reducing my dependence on cloud-based AI services. This has been driven by two main factors: cost and trust. Cloud services can be expensive, and I have sensitive information that I don't feel comfortable sending to a cloud service. I've found that using a hybrid approach with local coding models significantly cuts cloud AI spending, but I still don't recommend a completely local-only approach due to the trade-offs in capabilities.

Running my local model on my own machine gives me peace of mind knowing that my data never leaves my PC. There's no trail, no monitoring, and no flags that give third parties access to audit my conversations. While none of the major cloud AI service providers offer zero data retention by default, consumer plans don't offer it at all.

The free plans don't even provide access to the best models. For example, on Anthropic's Claude Pro, there are no zero data retention options, and Google saves Gemini chats for 18 months by default. OpenAI's ChatGPT saves chats indefinitely unless manually deleted. Despite these concerns, local AI models offer full capability from the start, which is a significant advantage over cloud services.

One of my main concerns with cloud services is the uncertainty about how my data will be used and stored. I've shared and dumped more information on AI platforms than I would on search engines or forums, making me more vulnerable to potential adverse events. Additionally, AI services aren't yet profitable, and it's unclear how providers will make money.

Recently, OpenAI started showing ads to Free and Go tier users, which raises concerns about the protection of personal information from advertisers. For me, a local model running on my own PC is the only option where I can trust that my data is secured without needing to rely on a corporation's policy.