I Also Like Gemini’s Web App more than ChatGPT
So I’m a big fan of Google’s Gemini. It seems to give me better results than OpenAI’s GPT. There have been a lot of comparisons lately between GPT and Gemini and how GPT is better.
Although I don’t think these comparisons are correct. They are not testing which is better, they are testing which is more comprehensive. GPT-4 is extremely comprehensive, it just likes to talk, talk, and talk. And so it’s going to win in comprehensiveness. But it just doesn’t feel natural.
Gemini, on the other hand, gives much more natural responses. So although it’s not as comprehensive I do like its responses for being more natural and more like what a person would actually say.
But this only applies to the API. And let me tell you I have some… strong opinions on people that say the API and the web interface for these AIs are the same. This was most notable with GPT 3.5. GPT 3.5 is not the same as the free version of ChatGPT. ChatGPT is significantly more detailed.
This always puzzled me. It says GPT 3.5 there, it should behave identically. But it doesn’t. It could be that ChatGPT has some additional prompts to make GPT 3.5 give a much more comprehensive answer but I have no idea what prompts they are giving it. GPT-4 does behave much more like ChatGPT but I don’t have a pro ChatGPT account to test it.
Which brings me to Google’s Gemini web app versus ChatGPT. Gemini does something that ChatGPT does not: it allows you to give it photos. And let me tell you, it has been a lot of fun giving Gemini pictures of random Mexican dishes.









