I’m thinking about trying ChatGPT Premium and wanted to hear from anyone who’s already using it. How’s your experience been so far?
What do you use it for?
Do you feel it’s worth the monthly cost?
Any tips or creative ways you’ve integrated it into your daily routine or work?
Also, on a side note — I’m interested in learning to speak Mandarin Chinese. Has anyone used ChatGPT to help with that? I’m wondering if it’s good for practicing conversations, vocabulary, or pronunciation tips.
Would love to hear your thoughts, especially from those who’ve been using it consistently. Thanks in advance!
I use it almost daily, for general searches and information, for my work, and for personal projects. Some examples include writing employee reviews based of my already formalized notes, structuring complicated business emails, making shooting schedules/plans for my photography, creative brainstorming around photography, lots more. It has been very helpful and I find it worth it, but you do have to pay attention. It isn’t always right and sometimes you need to remind it or reiterate items (I think you have to note when you want something to be remembered, but I’m still learning). It also sometimes doesn’t pay close attention to details and this seems a bit random. The other day I asked it for food options that were open late after 11 and it gave me choices that closed at 10. Even when I reminded it.
I’ve also been trying Claude. It also works well, but if you get into a really long discussion with it, it will just stop working. I’ve since learned that you need to break these discussions into subtopics, add them back as artifacts within a project for it to reference/remember.
I feel like I’m probably in need of training at this point to really start doing things correctly, but I’ve found it worthwhile despite some shortcomings. I typically type manually and don’t use speech to text or real-time conversation, so I can’t comment on the language learning question other than to say I think it might be a bit slow for intensive learning.
Answering difficult customers can be a lot of creative fun.
Maybe in your business it is not needed, but how would they react when talking face to face, when ChatGPT cannot aid them?
Reminds me of an owner of a print shop, who reacted to a customer review on Google starting with “Dear Mister Blockhead, only idiot of your format…”
I think we are going to lose huge part of humanity in both good and bad way.
I don’t think we are losing anything with this. Some clients are actually assholes and are unreasonable. Better leave it up to AI to answer them and not waste our time.
If you have no plans to go retail, then it is understandable.
I have experience from retail, that some, mostly elderly, but not only, people would go to pay us a visit just to torture us will bullshit questions about the products, while I knew that 90% of such people are never going to buy anything at all.
For them it is kind of entertainment.
Italians are like that. They will keep us on the chat for a long time only to buy maybe 1 thing or not at all. 0 sense of value of time.
The only plus side of having the Italian warehouse is they never complain when things go wrong or delayed. They’re used to the dysfunction of Italy where everyone is delayed and companies dont care about their experience.
The Brits act like you shot their dog if their package is delayed by 1 day like I am the actual courier and doing it on purpose.
It’s worth for as much use as you want to give it. I have been using it extensively for proof reading, email drafting, basic research, even conversations and general advice.
I should just warn that it seems that it eventually gets tailored to your thoughts and your POV, so might suffer from some confirmation-bias. If you use it to verify things or double-check your opinions, be wary.
I was surprised (but not upset) to learn that new conversations in ChatGPT do have access to previous conversations – it’s basically a memory system, where it’ll search your history for related ideas and include them in its context.
You can actually tell it not to do this when starting a new chat, though. The instruction it suggested to me, which I’ve tested, was “In this chat, pretend you know nothing about me. Do not use long-term memory.”
But it’s important to send that instruction before asking it anything else, in a separate message, since otherwise the context-search stuff will still get applied to the content of your question and provided with the message.
Similarly you need a fresh chat for this to work – it’s unable to “forget” any previously-matched context in an existing conversation.
Edit: After interrogating it some more about this feature, I’m less sure it does what I thought. It looks like the memory system will still be applied regardless, and it will just pretend not to see it. So confirmation bias could still slip in that way. There’s a hard-toggle for memory access in ChatGPT’s settings, though.
In a way it is both human and non-human… we cannot forget things on purpose
However digital memory is very persistent.
I guess you could go to the memory settings and delete things you want it to forget, along previous chats
But it is a bit cumbersome
Right. I was imagining it had some tools to adjust its own chat settings, like to disable long-term memory in this chat. I think that would be a pretty interesting feature.
It remembered various things that it asked you and you would see a prompt that it was “updating memory” or something like that. I’ll look for it when I get in front of the computer.
