Will a Trained model be upgraded once there will be gpt-4?

I am planning to integrate a model - Curie or Davinci for our business, but the concern is that once GPT-4 (and next) is out, I will have to start from scratch. Frankly this is quite a hurdle as there has to be quite a bit of investment into training a model and then having to start all over again as there will be a possible significant upgrade to a new version.
Anyone has any insights into how this issue can be approached?

Marius
Ps: I asked, of course, chatgpt this, but he has undecided, and said I have to decide myself, but provideed no realistic criteria for thia

1 Like

I’m also asked new bing about your question, it says you should contact open ai for this :slight_smile:


I see, more or less the same eexperience. Currently gpt-e.5 turbo has no training model but only stock chatgpt.
Anyway, I should still work on the connectivity with the other parts of the business (ERP, CRM, PIM etc) and trying to understand the implications of the APIs as well
@mohsen.naragh mohsen.naragh Do you hav similar intentions for the trained model or very diferent?

oh well, my experience on the topic of ai, especially development, is pretty small i-m also coming from the field of 2d and 3d artists and put all of my time into sculpting 3d characters, rendering, texturing, etc but recently I thought maybe stepping up and learning some c++ code along with unreal engine 5.1 so I’ve started learning the software and visual studio 2022 at the same time and sometimes when I ran into an issue I ask to chat Gpt if it can debug the code and explain the issues I have. :slightly_smiling_face:

my line and background very different. bad same scope :slight_smile: