You just described being a manager with salaried employees.

imagine paying for this slop

>>110246018
That's your usage limit for an unmetered plan you pay a fixed-fee for each month. You can add usage to it and spend as much as you want immediately. Obviously such limits are necessary to prevent jeets and jews from doing their thing.

You are right that the pricing lacks transparency but this is a complicated problem and also one that is/will be solved by competing Chinese models. For what Opus 5.5 can do, the $20/month price for Claude is fantastic right now. If someone else offered the same for less I'd switch in an instant. All of these companies can lick my balls, no loyalty.

File: file.png (394 KB, 1089x553)

394 KB PNG

>>110246046
>You can add usage to it and spend as much as you want immediately.

>>110246046
what do you even need it for
shit costs more than my internet plan

>>110246055
I don't know what that is but it's not really gamified like app games.

In my experience the current amount you get out of Claude is rather generous. There are a lot of very stupid people who waste it by not knowing how to use it, but they're actually helped by this kind of pricing plan.

If you have a project that has to get done immediately then it's important, which means it's worth some money to you. In that case the cost of a $20/month plan is trivial for what you are paying for.

tl;dr: you're broke and gay

>>110246076
I made five applications that I use nearly every day in the first month I had Claude for $20. I'm finishing up a fun little video game project later tonight or tomorrow when I feel like getting around to it. I'm presently not "using" my quota, it's fine I don't care. $20 isn't even real money.

>>110246023
which is interesting, because i was about to say they do this so you use it more and give them more training data.
ya'll buttholers are building the butthole for free. some are even paying to do it.

>>110246080
Not gamified, but monetized. Claude gets you to use it more in its own ways, like social engineering you.

>>110246018
LOL HOLY SHIT
I never thought of it that way

Quite jewish that unused usage isnt compounding, i mean you paid for it

Why does everyone just accept that like its natural

Maybe because traditionally api usage is also limited by time and not accruing

>>110246046
>>110246018
>Claude is literally a gacha game
so this is why i never saw a vibeslopper do anything productive, lol

>>110246018
Kinda cool that there's a desktop toolbar for this, is there a chatgpt/codex equivalent for windows?

>>110246076
are you sometimes irritated with niggerlicious anti-patterns on some websites?
These days, any time I'm annoyed with something, I'll just dispatch my clanker to build me a view of the data to exactly my preferences and without any bloat.
It's kinda funny that clankers simulatenously destroyed the web but also realized the original vision of actual "user agents".

>>110249734
you can vibecode it in 10 minutes

 Send authenticated GET requests to https://chatgpt.com/backend-api/wham/ with Authorization: Bearer <Codex access token> and, when available, chatgpt-account-id.

 - usage returns plan_type and rate_limit.primary_window / secondary_window, containing used_percent, limit_window_seconds, and reset_at (Unix seconds). Identify 5-hour and weekly windows by 
   duration.
 - rate-limit-reset-credits returns available_count and credits. Filter status === "available"; use title, is_supported_by_plan, and expires_at (ISO timestamp or null). This only reads       
   banked resets, never redeems them.

 These are undocumented ChatGPT endpoints and may change.

>>110246018
>having to live with a 5hr limit
I'm good with a /goal and running the weekly on its own

>>110246018
>Claude

ChatGPT/Codex or even Gemini are good enough for most usecases. Claude is marginally better, but is a lot more greedy. Just not worth IMO.

Anyway, regarless of which cloud provider you use, you should also own your own setup with the chinese local models and agents to save tokens for the more complex tasks.

You're quite literally wasting money if you're asking the frontier online models to write your boilerplate slop.

The chance is why I can't justify spending money on this.

It really does feel like dice rolls. Which is only acceptable to me if I can roll the dice infinitely.
I don't need the AI to be perfect or 1 shot everything. If it takes 3 or 4 rounds of back and forth, I'm okay with that.
I'm not okay with paying for those faults and waste though.
I'm not paying for tab completion that works 1/5th of the time even if that 1 time is useful and saves time, all things combined its costing more time and money for no good reason other than chance.
If I want to gamble I'll go to the casino.

>>110246993
I can't speak for the majority but I actually like it when companies tailor their products to my workloads. I prefer them training on my data for my use case since I'm the one that uses the thing and I'd get more benefit out of that than anyone else. It's symbiotic and just because money is involved doesn't take that away. The free ones do that too and I hope they improve too because I use those too.

>>110246018
i literally never get even close to any limits and entire days go by without any usage

i have no clue what the fuck i could even use it for

>>110246080
How do you become unbroke if the tools to work cost money. And not just $20, like you said if you're working on something of value you'll have to put more.