Google today released Gemini 1.5. This new model outperforms ChatGPT and Claud with 1 million token context window — the largest ever seen in natural processing models.
But at home, I use bard (or Gemini now). Also Gemini does not have a api I can go into.
Comparing the two with about half a year of usage –
Gemini has been giving better answers overall. One extremely good use case is writing JavaScript and node. It frequently gives me modern JS, and often is correct in one or two prompts. Where Chatgpt, I have to modify the prompt a few times to request esmodules over commonjs and some other technical details. It freaks me out how bad some of the Chatgpt JS code examples are and worried to see my juniors use it.
Another is that Gemini also gets updated with the latest data. Ive been doing a lot of hobby research on geography and time periods of the geography, and Chatgpt frequently has to get fact checked. One major wtf was when the area I was researching had a “Springfield”, and Chatgpt gave me Simpsons storefronts.
Chatgpt has been just better for creativity. Those hallucinations are hilarious, and have been great for placeholder content or to help me flesh something out.
Chatgpt (or maybe OpenAI) generates better art. Gemini just released that option, so I haven’t done much comparison.
I honestly don’t care who wins. I’m just trying to do cool shit with AI while we’re still in a wild West of it.
I honestly don’t care who wins. I’m just trying to do cool shit with AI while we’re still in a wild West of it.
I have had this thought since the beginning I started using ChatGPT, I’m honestly so glad there is a lot of competition to make this something that won’t be locked down and capitalized by a single company… Yet.
At work, we have chatgpt for code AI support.
But at home, I use bard (or Gemini now). Also Gemini does not have a api I can go into.
Comparing the two with about half a year of usage –
Gemini has been giving better answers overall. One extremely good use case is writing JavaScript and node. It frequently gives me modern JS, and often is correct in one or two prompts. Where Chatgpt, I have to modify the prompt a few times to request esmodules over commonjs and some other technical details. It freaks me out how bad some of the Chatgpt JS code examples are and worried to see my juniors use it.
Another is that Gemini also gets updated with the latest data. Ive been doing a lot of hobby research on geography and time periods of the geography, and Chatgpt frequently has to get fact checked. One major wtf was when the area I was researching had a “Springfield”, and Chatgpt gave me Simpsons storefronts.
Chatgpt has been just better for creativity. Those hallucinations are hilarious, and have been great for placeholder content or to help me flesh something out.
Chatgpt (or maybe OpenAI) generates better art. Gemini just released that option, so I haven’t done much comparison.
I honestly don’t care who wins. I’m just trying to do cool shit with AI while we’re still in a wild West of it.
For art both are a joke compared to midjourney
How do you run bard at home?
By visiting bard.google.com
I have had this thought since the beginning I started using ChatGPT, I’m honestly so glad there is a lot of competition to make this something that won’t be locked down and capitalized by a single company… Yet.