Emanuel L

Building WanderHue - visualize how much of your city (and the world) you've actually explored. 🗺️
Check It Out

[Emanuel L @E_Lopini]

@larsencc The fact that it isnt a default phone assistant and the voice turns are too slow when searching

[Emanuel L @E_Lopini]

@benln Voice mode is goated. Wish it was a bit faster and gave feedback while performing a task and didnt go silent. When we having an assistant from GrokBot as an android personal assistant?

[Emanuel L @E_Lopini]

@blueemi99 Dont use Max silly

[Emanuel L @E_Lopini]

@hqmank Toyota better than Ferrari? Match price to price a 20 dollar sub of ChatGPT has separate usage for chats, Claude charges for chats so 20$ better than Claude but a 100$ claude is indeed better for today before devday than the 100$ chatgpt. Tomorrow may change

[Emanuel L @E_Lopini]

@mreflow You keep looking at max setting and not medium and high

[Emanuel L @E_Lopini]

@zwb44 Now depending on if Haiku5.5 is luna level then you dont need two subscriptions you can finally just pay for 1

[Emanuel L @E_Lopini]

@zwb44 No, its cheaper. Its for easier work. Its too expensive too throw all your work to opus low. Throw easy work to sonnet medium and then if you dont like it go up

[Emanuel L @E_Lopini]

Am I the only one that just noticed /insights in Claude?

[Emanuel L @E_Lopini]

@shownotover 4.7 is known to be more expensive in tokens when a task is completed. You are better off just going back to 4.5 or 4.6. We will see how 4.8 turns out.

[Emanuel L @E_Lopini]

@lydiahallie This is fantastic. Quick question when they reset is there a better prompt than 'Try Again' would 'Continue' be better. Does Try Again restart it from the beginning?

[Emanuel L @E_Lopini]

@darekgusto I am in your same boat. I understand the race on intelligence but Composer 2.5 was too cheap to ignore as a daily driver. Hope 4.7 is successful or they come up with the Composer 3 model

[Emanuel L @E_Lopini]

Has anyone found good insertions for Jev in their AI workflows?
So far model effort, model delegation, and if tests should run is all I got

[Emanuel L @E_Lopini]

@theo Im torn between Luna and Opus 5.5 they serve different purpose in my stack. But ima go big brain better since u forced me to pick

[Emanuel L @E_Lopini]

@ArtificialAnlys Why are we only comparing Max the data already shows that Max isnt the best setting. Show this same analysis with Medium or High and tell us what you find

[Emanuel L @E_Lopini]

@claudeebum I didnt notice only 30% more usage. Thanks for pointing that out. Il ltake what they can offer for the same price

[Emanuel L @E_Lopini]

@alexgetmancom Half price for same performance

[Emanuel L @E_Lopini]

@DanDr1s You must be rich to not use the cheaper pricing on Luna. I would use both Claude as an orchestrator with Luna subagents using Codex CLI.

[Emanuel L @E_Lopini]

@LuminaBench You are dismissing the price difference 1 point difference for half the price. Tokens are scarse