@thsottiaux I thought you reset 🤦🏾♂️. I wasn’t ready. Don’t scare me like that.
If you like your new look after the haircut then I’m sure everyone is looking forward to one though haha 😬.
A new look celebration 🎉
@thsottiaux Allow for bi-directional syncing of projects/workspaces across devices (files and chats). Currently have a janky setup with syncthing to keep my work synced between a laptop and desktop. But having that inbuilt would be cool.
@thsottiaux Usage limits.
Far more tenable than alternatives. Work’s something best done with focus. Interruptions of limits making you wait sucked. And codex was better at letting me get to a flow state without the system blocking me. And without a 5 hour; self pacing weekly was just prime.
@thsottiaux The 5 hour limits are a shackle on actual work though. Please don’t :(
Maybe a 12 hour limit?
5 is…too restrictive.
I really dont want to have to work after work hours. And sometimes thats what the 5 hour limit would effectively make me do.
@sama@OpenAI In cursor when a chat gets too long it prompts you to summarize the current chat and start a new one.
I try to be conscious when using larger models when feasible and use smaller ones to summarize. Could you guys make that a click of a button on ChatGPT please.Ty❤️
@sama The more I think about it the more I’m into it.
Tranched subscription plans. Modularize it so users can choose what they want; (eg 5$ for unlimited 4o, 5$ for 200 o3-mini high, 5$ for 10 deep research etc.) Yeah I’d be down. And with this you could integrate plus plans with IDE’s
In the conversion from speech to text it would be nice if it can identify speakers. And actually annotate both questions since it has caught both (hence its response) but only annotated one of them.
I noticed an interesting capability in advanced voice mode. Me and a friend tested it a couple times when we talk over it and ask it completely different questions at the same time. And it answered the both of us. (Doesn’t always work, it can mishear sometimes) @sama@OpenAI
Although the interesting thing was; it wouldn’t have both questions annotated in text after the call.
The model obviously caught both sentences (hence the response) but only converted one from speech to text.
@chessinogame So...this like. Never happened huh? Feelsbadman. I wish this app got more dev support. The userbase has shrunk to next-to-nothing because of this...
I just remembered that one guy who started a Twitter bot that tracks Elon’s private jet…was buying out Twitter. Just to…like. It can’t be. Right? Like. That would be crazy…right??? 😳
@elonmusk Maybe Tesla should focus on getting the CYBRTRCK out there first 😅. Still waiting fam…I reckon adding another R&D intensive project when existing ones haven’t scaled production yet would draw you too thin.
Some guy: I’m done with you Microsoft. I’m moving on.
Microsoft: I’ll never let go…. Although jokes aside. Does he still get stock based compensation for this takeover? Is this circumventing insider trading? Real loophole found???