Agus

Building something.

[Agus @AgusMierda31558]

@thsottiaux @CtrlAltDwayne Dont do that. Dont give us hope

[Agus @AgusMierda31558]

@thsottiaux @tldraw @sama You forgot to tell him "make no mistakes" tho

[Agus @AgusMierda31558]

@wtkprol @thsottiaux Because it doesnt matter how cheap the tokens/dollar are for the Pro plans some people just can't justify spending $100 on inference.

I get your point. paying $20 and expecting more is not realistic. However, my point is that those are the accounts where openAI can make money.

[Agus @AgusMierda31558]

@wtkprol @thsottiaux The plus users are literally the ones paying substantially more per token than Pro x20. If theres a tier that its subsidized is the Pro ones. And also, they can barely squeeze those tokens with the 5h window limit.

They are the ones paying for your reasonable 100$ plan

[Agus @AgusMierda31558]

@thsottiaux So no reset? :')

[Agus @AgusMierda31558]

@thsottiaux Say no more. I just used my banked reset ffs

[Agus @AgusMierda31558]

@thsottiaux @FelipeFr1702 The fact that every competitors reset brings OpenAI and your name into memory is a big win in terms of marketing tbh

[Agus @AgusMierda31558]

@thsottiaux @itssotsot If only we were billed by turns or skills or plugins the panel would be useful

[Agus @AgusMierda31558]

@thsottiaux But I dont get it. Was there a problem with the rollout? Was it planned? Is it because a infrastructure limitation?

[Agus @AgusMierda31558]

@thsottiaux @avirajkhare00

[Agus @AgusMierda31558]

@melvindvivas Luna high whole day so... like... 2 simple prompts?

[Agus @AgusMierda31558]

@thsottiaux ...no?

[Agus @AgusMierda31558]

@thsottiaux You might convince me with a little reset to be fair. I need to move from opencode and can't be bothered to do it manually. I need tokens for that

[Agus @AgusMierda31558]

@thsottiaux An enhance analitics panel we we could see not only turns but tokens by models. I swear depending on the day my usage drops faster or slower and I dont know if its because of the cache hits, number of tokens of... 'something else' *wink,wink*

[Agus @AgusMierda31558]

@thsottiaux A finetuned model focused on vibe-coding / orchestration. It would focus on talking in plain english to the user and token optimization when it comes to delegating tasks. I hate tinkering with my prompt to get there only for a new model to come out and behave differently.

[Agus @AgusMierda31558]

@thsottiaux I keep on using Opencode, why should i switch to Codex?

[Agus @AgusMierda31558]

@thsottiaux Essentially, more control over my usage. That (or something like that) would allow to schedule my usage based not only in my availability but also in the tasks scopes.

[Agus @AgusMierda31558]

@thsottiaux ...would be to maybe have a customizable session limit.

Say I have a long task in a setup with all permissions so the agent can run for a while.

I dont know how much weekly usage that will take. I might want to set a session limit of "20% of my weekly".