Zhong鐘

AI for Enterprise | Building Full Stack Architect Agent

[Zhong鐘 @yangzhong01]

@thsottiaux @alexgetmancom Tibo, kindly. Alex is at the center of universe

[Zhong鐘 @yangzhong01]

@thsottiaux Offer off peak hour rates. We will reschedule our days for you

[Zhong鐘 @yangzhong01]

@thsottiaux https://t.co/XVKr4Y2WWt

[Zhong鐘 @yangzhong01]

Intelligent UI built this for me, but it doesn't work. I need a Reset from Tibo

[Zhong鐘 @yangzhong01]

Intelligent UI built this for me, but it doesn't work. I need a Reset from Tibo

[Zhong鐘 @yangzhong01]

@matyaslustig @marclou You're definitely onto something here!

[Zhong鐘 @yangzhong01]

@TokenGremlin Now you know why they are cutting prices

[Zhong鐘 @yangzhong01]

@Davidwuuu92 @thsottiaux Check the old X post for a few months ago, this used to be a HUGE issue. It used to take up over 50% of request volumes.

[Zhong鐘 @yangzhong01]

@JustinasRoland0 @thsottiaux https://t.co/C9OSZ2pipS

[Zhong鐘 @yangzhong01]

Your OpenAI subscription used to buy you $3.99 worth of chips (left). Today, that exact same bag is worth $1.99 (right). Are you getting less "value"?

[Zhong鐘 @yangzhong01]

Your OpenAI subscription used to buy you $3.99 worth of chips (left). Today, that exact same bag is worth $1.99 (right). Are you getting less "value"?

[Zhong鐘 @yangzhong01]

@embw_l0x @Selphares1 @thsottiaux now you are making shit up

[Zhong鐘 @yangzhong01]

@matthewmillerai What is USAGE? Usage is the number of tokens you can use each month, NOT how much those tokens cost. Your get X tokens before, now you still get X tokens. OpenAI cut the cost of tokens by 50%. So your X tokens' value is less, but usage stays the same

[Zhong鐘 @yangzhong01]

@embw_l0x @Selphares1 @thsottiaux You are confused, OpenAI give use certain number of tokens each month, NOT certain amount of dollars per month. When they maintain the number of tokens but cut the cost of the tokens, we still get the same number of tokens. The equivalent of dollar of those same tokens got cut

[Zhong鐘 @yangzhong01]

@JohnSmitheuxc @AlchainHust 理解一下，不用谢

[Zhong鐘 @yangzhong01]

@JohnSmitheuxc @AlchainHust 你这是用GPT臆造的回复？Opus5.5好像跟GPT关系不大。

[Zhong鐘 @yangzhong01]

@PawelHuryn This is very good information, but it also can be misleading. For example, two companies may provide the same number of API calls, but company A's API price is 10x of company B's price. This chart will make it look like you are getting a better deal from company A.

[Zhong鐘 @yangzhong01]

@giadotai are you Gia or Gia Dot

[Zhong鐘 @yangzhong01]

@giadotai Reset in 6 hours

[Zhong鐘 @yangzhong01]

@KatieMiller Are you saying "Trust me bro, OpenAI isn't to be trusted" ?

[Zhong鐘 @yangzhong01]

@MaaSonder @thsottiaux You just need a model that creates great frontend then "most" people on X will agree you have the best model. Nothing else matters. That's a fact

[Zhong鐘 @yangzhong01]

@LufzzLiz 您这问题又点大，Opus 5.5 输入只有4K忽略不计。9.8亿都是反复在读缓存。同一个项目文件和历史对话被反复送进API评估了无数次。输出相对很少。相反，Codex读了大量文件，大大超出context window，可以想象缓存命中不会那么高。两个模型做完全不同事情，没有可比性