What is the point of letting it be "creative" if it isn't going to be correct? Not impressed at all.

For example, I asked: "Can I avoid the $250 fee for a Chase Sapphire credit card?"<br><br>Obviously, it should answer "no" the same way I would answer, but there are some steps you can do to get the fee waived.<br><br>But it gets "creative" and says you can do things like "Delete your cookies" and "Use a VPN" to get the fee waived.<br><br>I mean, as a credit card professional who has been doing this for 25 years, let me tell you "Delete your cookies" has nothing to do with a credit card fee. Likewise, there is no correlation between a VPN and a credit card fee. That is a "kitchen sink" type of answer, and even then, a poor one.<br><br>I would rather have GPT come out and say "I don't have any information on this topic" then say "Here's some random information that is completely useless to you". You are welcome to downvote me, I don't care. I love that this technology is here, but I can't wrap my head around the value of a response that is completely inaccurate and misleading.