Same thing happens when I need to copy a private key from my phone to pc, just end up opening notepad and typing the shit out. Opportunity to flex my typing muscles and speed. All the Mavis beacon lessons can't go to waste 😂

That's us why I have the zai sub. Just out of my 5hr limit for now. According to their usage docs already used 144 million tokens today

Herdr makes this possible across different agents. Claude code added this feature recently.

Yeah. I know the flash models are very fast. Experimented with them in opencode go I'm considering the 100$ plan to support my already heavy usage for the month. Would test out with a much lower plan to see if it is reliable before committing to the higher plan.

Was considering this but heard reports from Reddit that the tokens per seconds was really slow

Just ran through two codex max subscriptions weekly limit, a GitHub copilot monthly sub, an open code go monthly sub and a 5hr z ai sub and the week hasn't ended.

I m arriving at the same conclusion. All the talk about scaling to more users is meaningless until it actually happens. Onchain seems to be best option.

They finally fixed their UI design issue with astra .

That was fast. And here I was just testing out grokbot and appreciating how good it was and an open source alternative that does a similar thing. And best of all uses pi the best llm harness in my opinion.

What provider do you use? Direct or opencode?