I rather have a product that is only good at one single thing than mid for everything else especially when the developer experience for me is much more consistent than using gemini and chatgpt to the point that I only have chatgpt for productivity reasons and also sometimes making better prompts to claude (when I don't use claude to make a better prompt). After realizing that Anthropic is discounting token usages for claude code they should have made that more explicit and also the API key (but hindsight is 20/20) they should already have been blocking third party apps or just have you make another API key that has no discount but even then this could have pissed off developers.
You’re asking two different LLMs to help you talk more better to another LLM?