Rambowang

[Rambowang @xhwjy]

@lunaaalunaluna @jinanni2024 由不得他们，地缘政治和安全问题，没有谈判空间，就看导弹有多少

[Rambowang @xhwjy]

@rl_env 🤡

[Rambowang @xhwjy]

@xKaka1456 I choose 2 and 5. We shouldn’t assume that compute capacity can be freed up or expanded. Options 1, 3, and 4 all violate this fundamental constraint.

[Rambowang @xhwjy]

@thsottiaux Go！

[Rambowang @xhwjy]

@thsottiaux @maria_rcks I still believe in ChatGPT and its ability to deliver newer, more powerful models. Today’s challenges are only temporary. What matters most is never losing the drive to innovate and explore the frontiers of AI. Keep pushing forward, We believe in you!

[Rambowang @xhwjy]

@linkfosta @jinanni2024 @grok 放屁，这根本不是经济的事情，你让夏威夷独立看看！

[Rambowang @xhwjy]

@mylifcc 今天消失了，这下更难受了吧😆

[Rambowang @xhwjy]

@SnozakiSakura 要不借你截图？周均50亿消耗😆

[Rambowang @xhwjy]

@MagicPower21M 因为在先蒸某些傻屌们的亲妈，打着中文在那鬼叫

[Rambowang @xhwjy]

@NFT_Chen 给我们超了他们还吃什么？

[Rambowang @xhwjy]

@thsottiaux Of all the updates over these 9 days, this is the one I’m most eager to test! Since Day 7’s context update, I keep getting “stream disconnected before completion” errors in Codex chats and have to restart the app to recover. Please fix this！

[Rambowang @xhwjy]

@laopai1990 Yes, that's exactly the dilemma they're facing right now

[Rambowang @xhwjy]

@t20000622yy 828k，也就是1m

[Rambowang @xhwjy]

@loveAIwang 这个修改让我今天报错了至少5次，原本我就设定的1m上下文（828k）

[Rambowang @xhwjy]

@thsottiaux Tibo, I suggest OpenAI follow Anthropic: offer more granular model tiers with matching subscription and API pricing. Let users choose compute based on their needs instead of treating everyone the same. This could improve resource allocation and ease OpenAI’s compute constraints.

[Rambowang @xhwjy]

@thsottiaux My Codex threw errors 4–5 times today, which has never happened to me before. I suspect this may be related to recent official changes to the context window. I had previously manually increased mine to 828K tokens.

[Rambowang @xhwjy]

@shengtang135754 八小时？上午用完，下午划水，晚上继续上班到凌晨？我们要更自由的使用，更大的量，周末能相对轻松一些自己能跑的dots。但和OpenAI收紧算力的目的是违背的。

[Rambowang @xhwjy]

@thsottiaux I need uninterrupted workdays. A 5-hour cap would disrupt my workflow, even with 2x weekly limits, leaving unused weekend capacity. Keep the no-window model and consider flexible rate limits or optional background queues. You previously ruled out restoring the cap.

[Rambowang @xhwjy]

@thsottiaux We enjoy freedom from 5-hour limits, not watching the clock. As a Pro $200 user, I’ve accepted lower quotas, but 5-hour caps would disrupt my work. I don’t want extra accounts for weekdays, then waste quota or sacrifice my weekends.