Ron Shina

[Ron Shina @capuchango]

@thsottiaux Speed resetting from Fast to Standard after restarting the app / after an update

On mac. Checked now

Version 26.527.31326 (3390)

[Ron Shina @capuchango]

@nickaturley Searching past chats / using past chats as a source (like Google Drive)

[Ron Shina @capuchango]

@trq212 (naming) if a branch was named in /branch [name] the name isn't placed in the rectangle on the right like after a /rename

[Ron Shina @capuchango]

@thsottiaux Search

- searching all threads
- using other threads as a source

[Ron Shina @capuchango]

@ericmitchellai The inability to use past chats as a source

[Ron Shina @capuchango]

@ericmitchellai searching past chats isn't great

[Ron Shina @capuchango]

@ajambrosino searching all threads

[Ron Shina @capuchango]

@simpsoka Touchdown (-like) celebration on big feature completion 🏈

[Ron Shina @capuchango]

@Dimillian better search - not having to remember which thread you should be searching in

[Ron Shina @capuchango]

@thsottiaux (prompt background color) choosing the background color for user prompts - like in chatgpt

[Ron Shina @capuchango]

@athyuttamre @l_mejiaC @OpenAI @ChatGPTapp (accent) maintain the requested accent in English and Spanish

(building context) even if the model isn't "smart", being able to say something like "now think extended" (or do that in the chat interface beforehand) so it has the context to give better answers

[Ron Shina @capuchango]

@thsottiaux search across all threads and all projects

you can literally be looking at the search result and it won't find it (if you didn't already know which thread to look in and used the thread's search)

[Ron Shina @capuchango]

@ericmitchellai @_simonsmith Keep the mic icon even after starting to type a prompt (like in codex)

shouldn't have to remember: "it's chatgpt, you have to start by uttering something or you can't use Speech to Text"

[Ron Shina @capuchango]

@ericmitchellai (speech to text) The microphone icon disappearing once you start typing - "you should have thought of that before moving your fingers"

[Ron Shina @capuchango]

@ericmitchellai - searching isn't great
- missing an option to load a chat without previewing artifacts (which can make a chat so slow it's unusable)

[Ron Shina @capuchango]

@thsottiaux Searching in all threads across all projects - you shouldn't need to know in advance which specific thread contains the result you're looking for

[Ron Shina @capuchango]

@Dimillian (branching) like in ChatGPT

[Ron Shina @capuchango]

@Dimillian (navigation with keyboard shortcuts) between prompts and responses / scrolling with the keyboard

(what's in this thread?) an optional pane, or table of contents, that is clickable and contains excerpts of prompts and when they were sent

(searching across all chats / threads)

[Ron Shina @capuchango]

@thsottiaux I thought this was deliberately opaque - notice how the Codex usage page shows percentages, not token counts.

[Ron Shina @capuchango]

@thsottiaux Kodak might be willing to allow "Codex Moment"