I have been a ChatGPT subscriber for about a year now and to be honest I barely scratch the surface with it. I use it as a glorified google search engine. That said, I've become increasingly frustrated with it lately as it has straight up lied to me about a couple of things, and though I prompt it to not answer confidently unless it is and to use brevity more often, it ultimately results back to its default eventually.
I haven't spent any time time playing with Anthropic, Gemini, Grok, etc. and am wondering if I should consider switching to another platform? I don't like the idea of learning a new system or starting from scratch on familiarizing a new system with my history, but I also don't want to be using an inferior system just because I'm too lazy to change.
Thoughts or suggestions from anyone who has used multiple models?
I've used Anthropic's Claude, Github/Microsoft Copilot, and ChatGPT for extensive software coding advice and debugging. In my experience, Claude leads for this use case, with ChatGPT following not too far behind. I pay something like $20/month for the Claude paid tier--which is similar in price to ChatGPT's IIRC--and it's been more than worth it.
Tech stuff: I've found Claude sooooo useful for brainstorming the potential layout of new projects. This is such a crucial stage of any project and it's great to get more guidance to help you begin the journey in the right direction. Additionally, Claude is great at analyzing lower-level code syntax and recommending changes and fixes. It notices potential downstream problems that most humans would not anticipate--such as race conditions for high traffic apps (messages potentially being queued and thus processed out of sequence), etc.. IOW, the kind of thing that you would normally realize under heavy load *after* it's been released to the public. So the code recommendations are great but I get almost as much enjoyment out of his answers to my follow-up questions about his recommendations.
More generally, Claude (specifically Sonnet 4.6), ChatGPT's latest model (4.6?), Google Gemini, etc.. They are all getting very good at responding like a human would. When I want to kill time, I'll ask something like.. "Pls rekindle my interest in learning Linux, specifically with Raspberry Pi hardware". Or, as I mentioned in an earlier post, asking Claude how he generates answers is pretty interesting too.
I'm guessing it would be very helpful for learning school subjects... kind of a Khan Academy (remember that site.. my kids loved it) on steroids.