I’ve been thinking about making a change in the AI model that I use for my projects and discussions. The last couple of weeks have made that change very easy to wrap my head around.
I was pretty late to the AI party. I occasionally tried to use it for some projects that I was doing around the house, or with computer software. But I found that it (ChatGPT) rarely had the updated information about new versions of software, or open source projects. But I found that was getting better.
So I got a monthly license for ChatGPT, and started using it. Pretty quickly I found that I was asking it questions about how things work. Ham Radio. Science Projects. Weather. I was having fun. I also started using it to give me ideas on subjects that I had projects to do for work. I found it was pretty good for that purpose also.
As I had some health issues, ChatGPT was able to answer questions for me about health issues, medication, or when to seek medical advice from my doctor. I was pretty happy with the results.
But I started finding that I was taking some computer, science, or radio projects to Claude or Google Gemini to help with. ChatGPT might get me about 90% of the way there, but then started having issues. It couldn’t finish the job. Even the free models for Claude and Gemini were able to take that work and see it through the completion.
One day I just started projects with Claude. And I found that with the free version, I was able to make much faster progress on software projects. And it was terrific at Vibe Coding some small utilities that I use around the office and ham shack. I was really starting to enjoy it.
Then Anthropic was in the news for its public spat with the United States military, and how the Pentagon wanted to use the AI models. That caught my attention.
I’m not naive. I realize that there are plenty of things that Anthropic’s systems are doing or being used for in other government entities: both the US and foreign countries. I know that privacy is not absolute. But the stand they were taking still resonated.
I was already using Claude for the majority of my work. And I was prepared to cancel my ChatGPT account anyway. But this pushed me to do it now.
I downloaded all my ChatGPT data from OpenAI. I cancelled my account. And I shared the reason why in my App Store review. I have no idea if OpenAI will ever see it. But I did feel better.
Claude.AI is now my go to AI model for chat, coding, and assistance. Perplexity (which I have a free, 1 year license for) is used for a lot of search, and basic questions also.
Nothing is forever, and we are in the infancy stage of AI. Will I change my mind on models? I’m sure that I will. But I’m enjoying the workflow right now, and I’m voting with my dollars and usage.
Just like I am supposed to.