I’ve created 14 MyGPTs so far.
Some I know for sure that I’m the only one using them, but others I don’t. Even though I have one or two that have more than 100 conversations, I’m not sure they’re not all mine . That’s one issue, I don’t think it’s possible to know if anyone other than ourselves is using the GPTs.
I often criticise the behaviour of GPTs because they’re not too constant.
I also try and use many GPTs created by others and some are awesome.
But yesterday I wasn’t able to accomplish what I needed with some custom GPTs so I tried the same with plain ChatGPT-4.
Even though it’s not specialised, I was able to get some pretty good responses, mainly text comparison, rewriting and generate text to download. I couldn’t do that with my own custom GPTs or others.
So I must ask: is it worth it all the hours spent building GPTs?
I think a massive possible benefit for making numerous specialised GPTs is that you can easily use them within a single chat using the @ {the GPT you want to use}
If you’re just doing general tasks I normally use good ol’ normal GPT-4, but having one specialised for it would probably help.
I’m trying to get used to using it too, but it’ll be great when I have an ethics expert I can just pop in and ask what they think, then maybe one that knows the specifics of a project I’m working on, and another that can help me code it, etc.
I have been creating custom chatGPTs for academics ,job , finances etc…and it seems that these custom GPTs perform worse than just using the normal GPT, when it comes to responses they are shallow or too broad and unspecific, overall they just seem to be lower performance than the normal version.
Is anybody having a similar experience? My aim is to have a custom version of Chatgpt that improves on a specific topic as I use it, em I just better off using one chat per topic on the normal GPT? Open to your suggestions.
My point exactly. I’ve been using my custom GPTs (I’ve deleted some that I don’t think were worth the maintenance) for academic purposes, mostly. I try to configure their behaviour so I can get what I need without prompting and some have been quite effective.
I also believe in the value of the specific knowledge that I provide them. It enables me to get answers within the scope of the subjects I need.
But I’ve been slowing down the investment, because most of the times plain ChatGPT 4 gets me where I need.
They are hugely useful in my opinion. I have quite a few of my consulting clients coming to me asking me to build really simple one for them, and some are after quite complex ones. It basically comes down to this for me:
Do you have a great prompt that gets great results; but it’s annoying to have to copy and paste it every time you want to use it? - enshrine it in a GPT
Have you maxed out your custom instructions on standard ChatGPT? - put them into a GPT
Have you built a multistep process to get an output from ChatGPT that get great results and you want to give ChatGPT specific examples of what you are looking for in the output? - Put it into a GPT
Do you want to learn something that requires a curriculum of some sort? - Put it into a GPT. A great example is that I am learning Japanese; but I just want to talk, not do Kanji, Hiragana, Katakana etc (cause I’m painfully lazy). I built a CustomGPT that has conversations with me based solely on the vocab from JLPT N5 and 4. I’m good with N5 which is the basic level; but I want it to push me to N4, so it forces me to have conversations using vocab I am not yet across.
I downloaded the vocabulary requirements for both levels and put them in the knowledge base and told it never to stray outside of this.
I could go to 100 with this list; but yes - they are insanely useful!
. That’s one issue, I don’t think it’s possible to know if anyone other than ourselves is using the GPTs.
