With @higgsfield releasing their API/MCP, and my ongoing craze around motion videos, I had an idea.
What did I do next? Ran the idea through @trymurmurhq
Here's the link to the entire foundation: https://t.co/U46dlnaH2n
With @higgsfield releasing their API/MCP, and my ongoing craze around motion videos, I had an idea.
What did I do next? Ran the idea through @trymurmurhq
Here's the link to the entire foundation: https://t.co/U46dlnaH2n
Umm… you could do what?
Use your own GPT plan for token usage instead of paying for the product’s own tokens, by signing in with GPT account?
Oh.
That suddenly makes the Cursor × OpenAI situation a lot more interesting. 👀
Dots, Space ~ Things one would need to ease up running multiple SaaS products under a single umbrella.
Dottie seemed to be quite slow for live demos tho xD
No 5hr limit in Pro 500 demn!
The real question now is GPT 6.1 Sol or Claude Sonnet 5.5?
#DevDay
Every number in the above video comes from a real run.
The video renders straight from the run's own data: memo length, competitors found, stories on the Jira board, total time. Nothing typed in by hand.
The first frame is a real capture of that run's Jira backlog.
@hedrees_ayinde well Astra 6 wipes out an entire week's credits in 2 days, on a single project.
on the other hand Opus 5.5 stays in for almost 5 days, while working on multiple projects.
using different models for different use cases makes more sense currently.