If you use Chat GPT-4, what do you think about Chat GPT-3.5?

For regular tasks like "summarise an article" or "Generate a launch guideline", GPT-3.5 & GPT-4.0 would perform similarly. The differences becomes obvious when you give it complex tasks that require some form of complex logical reasoning where it need to make logical conclusion based on provided context. For developers, a major difference in the API will be the ability to for GPT-4 to process more token/words. GPT-3 supports about ~4000 tokens while GPT-4 has capacity for about ~8k tokens to 32k tokens Cost-wise, GPT-4 is cost more, so for basic tasks you're better off using GPT-3.5.

If you compare them it's like 3.5 is a child but 4.0 is like a teenager. I mean 3.5 will lie to you because of the knowledge gap but 4.0 will lie to you because it wants to impress you 😅 But both are hard-working and talented kids.