Member-only story
GPT-4 Vs GPT-3!
Welcome back! GPT-4 is the newest update to the very popular GPT-3 model developed by OpenAI! So, let’s take a look at what the new version of GPT is bringing! If you want to check out their entire release log, here is a link to it:
Images
Probably the biggest thing within GPT-4, we now have the ability of sending an image to GPT-4 and getting an output from it! The example that OpenAI used was the following image and asking GPT what the outcome would be when the object drops:
Safety Update
The new version of GPT-4 also updates the safety of the model as well, according to OpenAI, it could be 40% more factual, as well as 82% less likely to respond to a request that is considered disallowed, this is addressing one of the biggest concerns from the previous version of ChatGPT (based on the previous model of GPT).
Increase Text Input
One of the biggest changes within this version of GPT is the increase from input text, GPT-4 can now process up to 25,000 words of text! On top of this, you can also send GPT-4 a link and even have GPT-4 interact with the websites as well!
In my opinion, this is a massive improvement coming from the previous version of GPT-3, I also do believe that the next version of GPT (GPT-5) could probably end up bringing some sort of Dall-E support, possibly outputting images since we can now import images to GPT.
Do you plan on using GPT-4? I would love to hear your thoughts about this!
Thanks So Much!
Feel free to watch my videos on YouTube or connect via Twitter if you have any thoughts or questions:
Thanks so much for your support!
