Claude just made a mistake, today is Codex’s turn.
stream disconnected before completion: An error occurred while processing your request. You can retry your request, or contact us through our help center at help.openai.com if the error persists. Please include the request ID aaaaaaaa-bbbb-bbbb-bbbb-dddddddddddd in your message.Every time I get this error message, I think it’s not just a bug, it might be a sign that something behind it is worthwhile.
It takes power to turn one AI model, a server to take that power, and a semiconductor to fill the server. When a request is rolled up, there is a bottleneck somewhere. Today’s mistake may be a very small trace of that bottleneck. We don’t know exactly what the problem is, but when we look at this flow, we think that the power, servers, and semiconductor industries can only grow bigger and bigger in the future.
The price of hard drives and RAMs is not too high these days. The rise in the price of components, which would have been overwhelming as before, is now read differently. I think that the resources that AI swallows are not ending in the power and semiconductor of the data center, but are shaking up the personal parts market.
One error message may be overwhelming, but considering the weight of the infrastructure behind it, it doesn’t seem like it’s just a story that ends with “I’m not lucky today.”
Loading comments.