如何将OpenAI账户验证错误反馈给工程师?
我试图联系OpenAI的某位工作人员,关于他们的手机验证流程中似乎存在的一个bug。
这个问题与我正常的账户登录或双重身份验证(2FA)无关。我可以成功登录ChatGPT,并且我的基于身份验证器的双重身份验证也能正常工作。
问题发生在OpenAI的手机验证流程中(这似乎现在是使用Codex所必需的),网址为 https://auth.openai.com/phone-verification。
在手机验证中,我需要通过WhatsApp接收验证码。
然而,我在WhatsApp中没有收到任何验证码。
此外,我从未将WhatsApp设置为身份验证方式,并且我所在的国家并不是OpenAI列出的支持WhatsApp验证的国家。我的电话号码也不与OpenAI列出的支持WhatsApp验证的国家相关联。值得一提的是,作为备选方案也没有提供短信或电话选项。
因此,我无法验证我的手机号码,也无法访问Codex,尽管我可以成功使用我的ChatGPT账户进行身份验证。我已经使用Codex大约12个月,所以这不是一个新账户。我之前也在Plus $200计划中。
我已联系OpenAI支持并提供了问题的详细信息。支持人员为我处理了退款并取消了我的订阅,但我仍然无法再访问Codex。
我不想创建第二个账户,以免触发任何欺诈检测,并且我有正在进行的工作线程,不想失去上下文等信息。
如果需要,我可以分享完整的屏幕录制、时间戳、重现步骤,甚至是HAR文件。
在这里发布是希望OpenAI的某位工作人员能看到,或者如果有人知道如何将真正的bug提交给工程团队。
提前感谢 :)
查看原文
I'm trying to reach someone at OpenAI regarding what appears to be a bug in their phone verification flow.<p>The issue is not related to my normal account login or 2FA. I can successfully log in to ChatGPT, and my authenticator-based 2FA works as expected.<p>The problem occurs during OpenAI's phone verification flow (which is seemingly now required to use codex) at
https://auth.openai.com/phone-verification<p>For phone verification, I'm required to receive a verification code via WhatsApp.<p>However I don't receive any verification codes in WhatsApp.<p>Furthermore, I've never setup WhatsApp as an authentication method, and I'm not located in a country that OpenAI lists as supporting WhatsApp verification. My phone number is also not associated with a country that OpenAI lists as supporting WhatsApp verification. For what it's worth no SMS or call options are given as a fallback either.<p>So as a result, I can't verify my phone number and cannot access Codex, despite being able to authenticate successfully with my ChatGPT account. I've been using Codex for ~12 months, so this is not a new account. I was also on the Plus $200 plan.<p>I have contacted OpenAI Support and provided details of the issue. The support agent issued me a refund and cancelled my subscription, but I still can't access codex anymore.<p>I don't want to create a second account at the risk of tripping any fraud things, and I have ongoing work threads that I don't want to lose context etc on too.<p>I have a full screen recording I can share, timestamps, reproduction steps and even a HAR file if needed.<p>Posting here in case someone from OpenAI sees this, or if anyone knows how to get a genuine bug in front of the engineering team.<p>Thanks in advance :)