Claude trying to let me know that the design updates i made to the UI will not be incorporated, including Glass blur, and that we need to re-write the entire app with Tauri/Rust/WebGL and it would take weeks + all the bugs fixed will emerge again with new ones. Claude advising me to stick to GTK/Python we used and ship.
> informs claude that we'll use the Three.js, we'll rewrite in Tauri/Rust and use that WebGL buddy
i was fighting claude code who is using opus 5 to find a solution for my usage limit widget. Opus completely advised to give up the project stating that there is no way to get the numbers accurate/calibrated to the real usage limits... it took me lot of tokens going back and forth to make it follow the "claude login" instead of "claude token" connection
right now modifying the complete ui/ux generated by claude code, had to give it some new guidelines to follow. I pushed the first stable version to gitea, now working on pushing the second one with a complete upgraded design. More to come
Who maintains it or fixes it when it breaks?
> the ai that created the code or contributors (github).
What is the actual cost to support it?
> ai subscription or free with open source llm.
What happens when the LLM goes prague or the model co (OpenAI or Anthropic etc) raise prices because more compute is necessary.
> switch to open source llm, cheap or free.
yes, i was shocked the other day to see that doubao is popular that qwen, minimax... and the others. They have another brand app for the international market called Dola (ex cici) but i think it is banned in NA
1. Doubaoโ ByteDance โ 382 million MAU
2. Qwen (Tongyi Qianwen)โ Alibaba โ 167 million MAU
3. DeepSeekโ DeepSeek AI โ 130 million MAU
4. Quarkโ Alibaba โ 180 million MAU
5. Yuanbaoโ Tencent โ 150 million MAU
6. ERNIE Botโ Baidu โ 220 million MAU
7. Kimiโ Moonshot AI โ 90 million MAU
8. StepFunโ StepFun โ ~50 million MAU*
9. MiniMaxโ MiniMax โ ~40 million MAU*
10. Baichuanโ Baichuan AI โ ~30 million MAU*
*Estimated based on tier positioning