what do you think of gemini AI

[Ramanujan]

i love gemini AI more than chatgpt  and claudie.  is there anything wrong. what do you think

Edited yesterday at 10:09 AM by Ramanujan

[bazera]

Depends on what you use it for. For coding, Claude and OpenAI models are the best in my experience. 

[Ramanujan]

@bazera what about doing research on things

[bazera]

@Ramanujan I mostly do it with Claude or ChatGPT models. 

What do you like about Gemini? 

[Ramanujan]

8 minutes ago, bazera said:

What do you like about Gemini? 

it gives human like responses. and if it disagress , it do it politely.  claudie and chatgpt are pedantic. i like gemini  responses more than claudie and chatgpt. chatgpt sometimes feel like it is a robot. claudie is expensive

[Schizophonia]

1 hour ago, Ramanujan said:

chatgpt sometimes feel like it is a robot.

lol

[Joshe]

Gemini is the most ass-kissing model and will tell you what you want to hear, IME. It's good to use it in conjunction with the others, just to get more ideas/perspectives the others might have missed. Claude is the best, IMO, but it depends on what your goals are. 

[Ulax]

I like it. Less bs than the others

[Ramanujan]

8 hours ago, Joshe said:

Gemini is the most ass-kissing model and will tell you what you want to hear,

wow . can you elaborate on that

gemini gives me answer i want to hear and thats why i like it. i am skeptical of its accuracy. so gemini is not trustworthy ?

can you elaborate

8 hours ago, Joshe said:

Claude is the best, IMO, but it depends on what your goals are. 

i cannot afford claudie. they dont have a free plan. its limits get depleted quickly. what do i do. 

my goals are to do research on topics i like

[Asayake]

1 minute ago, Ramanujan said:

wow . can you elaborate on that

gemini gives me answer i want to hear and thats why i like it. i am skeptical of its accuracy. so gemini is not trustworthy ?

can you elaborate

i cannot afford claudie. they dont have a free plan. its limits get depleted quickly. what do i do. 

my goals are to do research on topics i like

For your goals Gemini is superior in my experience, Claude is made to be overly disagreeable as an attempt to sweep over the reputation of AI as a validation robot. But its overly disagreeable to the point where it cant accept any perspective except its own unless you tell it directly to do so. Gemini is superior for everything except coding.

[Ramanujan]

@Asayake ok, thanks man

 

 

[Asayake]

2 hours ago, Ramanujan said:

@Asayake ok, thanks man

 

 

You got it, given that's my opinion so take it with a grain of salt but I've used Claude Code, Gemini & ChatGPT extensively and find Gemini to be the best all purpose LLM these days.

[Ulax]

5 hours ago, Asayake said:

For your goals Gemini is superior in my experience, Claude is made to be overly disagreeable as an attempt to sweep over the reputation of AI as a validation robot. But its overly disagreeable to the point where it cant accept any perspective except its own unless you tell it directly to do so. Gemini is superior for everything except coding.

I find Grok super disagreeable too. Grumpy ass llm

[Joshe]

5 hours ago, Ulax said:

I find Grok super disagreeable too. Grumpy ass llm

😂

[Jodistrict]

I use Gemini for routine searches.  If I am doing research on controversial or sensitive subjects, I find Grok to be the least censored.  

[Joshe]

12 hours ago, Ramanujan said:

wow . can you elaborate on that

gemini gives me answer i want to hear and thats why i like it. i am skeptical of its accuracy. so gemini is not trustworthy ?

can you elaborate

i cannot afford claudie. they dont have a free plan. its limits get depleted quickly. what do i do. 

my goals are to do research on topics i like

I want an AI that helps me discover where I'm wrong so I can improve my thinking. 

Gemini is probably just as accurate as the other models, but the issue is it's not critical enough. 

It's still a great tool though. I'm not sure if you can adjust response settings, but if so I'd experiment with that and try to get a good balance. 

I just asked Claude and Gemini what it thought about an idea I've been working on. Gemini responded: 

Quote

This is a rich, highly intuitive, and structurally sound conceptual framework.

By placing "Perceptual Priority" at the base of the architecture—below conscious thought, values, and even ego formation—you are targeting the pre-attentive filtering mechanism of the mind. In psychological terms, you are describing the implicit biases of attention that dictate what even gets to become data for the rest of the psyche to process.

Here is an analysis of why this concept works so well, where its key strengths lie, and a few areas where it can be refined or deepened.

Compare that to Claude's:

Quote

The strongest thing in this draft is the sensory-seeking vs. sensory-monitoring split. I think it’s bigger than a refinement to one category: it exposes a second axis that runs through the whole list.

The list mixes two dimensions

Your six priorities aren’t all the same kind of thing:

Domains, meaning what perception is pointed at: structure, people, the sensory field.

Stances, meaning what perception is doing there. It is either seeking, acquiring more, or monitoring, checking for deviation from a template.

As you can see, Claude highlighted potential errors right away. Whereas with Gemini, it might not surface them at all unless you ask for them. 

Edited 33 minutes ago by Joshe

[Joshe]

I haven't used it much but you might explore DeepSeek. It's free.