bitcoin:native bouncing off 200W MA
Historically it’s been a great long term support with exception during the FTX blowup.
So unless there’s another unannounced blowup…
You've been asking for this one...
Now in preview: Codex in the ChatGPT mobile app.
Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox.
@sporadica Tbh if anything it��s bullish. Anyone who is a multi model power user knows Grok sucks. They’ve recently made great moves with Cursor to fix this.
So most tech users already knew how bad Grok was, Elon is taking steps to improve (Cursor) and recouping idle costs in interim.
So far day 1 into using @sama 's Codex and biggest difference besides insane usage limits is unlike @cursor_ai it doesn't eat all my RAM and randomly zero out my markdown files.
There 100% is less configurability though and debate out on who has the better harness for output.
@goodalexander Agreed. For the first time in months I’ve actually switched from Cursor to try out Codex. I love Cursor but it’s hard to justify not using 5.5 at these insane subsidies. Gotta love competition.
@banteg I’ve noticed it’s correlated to off peak times. Early morning European hours there’s less demand for compute and models tend to work faster.
Just my anecdotal experience though. Check for yourself.
@levelsio Unfortunately, the best combo I’ve found is getting ear plugs and keeping the window open. It’s not perfect as loud noise still comes through and the ear plugs will occasionally pop out.
@ericzakariasson There’s a really annoying bug, when focused on a markdown document and you switch to another agent, the document will lose all its contents. Consistently have to spend credits to get the agent to restore from transcript if it’s a recent document with no git history.