lol people use gaslighting so freely these days. It’s not gaslighting you, it’s “hallucinating” the answer aka making it up and being confident about it. It’s a very common issues with AI chat bots. Even ChatGPT does it even when I shared the pdf documentation with it.
My experience in using both is that ChatGPT information is less words packed better information and more factual information. While Gemini talks bla bla nonsense and too much.
ChatGPT voice dictation is just insane, absolutely amazing. I have never seen anything like it, and Google voice dictation is just normal stuff, which means bad.
Gemini has a lot of problems and sucks when you compare to ChatGPT. It often gets the prompt wrong or refuses to do it saying it's just a large language model. While other language models handle the same tasks without any problems. I paid for both services so I used their latest models. Gemini lecture you on almost everything. Whether you are researching information or it thinks you wrote something wrong. It provides the least accurate information in my opinion. Why pay $20 a month for Gemini when other LLM are much better. They should have it priced at $10 a month because it's not worth the asking price of the subscription. Thanks for the video, I enjoyed it
Gemini is a novice and the primary focus is for the human centric approach. Which means the more is yet come and will have far lesser hallucinations. There is more advanced update pending for public usage.
Hey, the purpose of gemini and chatgpt is different. Chatgpt is mostly based on its innate data that it has collected , with some browsing capabilities . While gemini is more like an extension to a Google search , in order to get good answers to you need ask questions in such a way that it utilizes it's google searching capabilities . Example , for the phone test you can instead write : I've given you specs for 3 phones , identify each of these phones (search the web , to find exact models of phones) . Compare these phones based on camera, screen , chipset and charging . Tell the strengths and weakness of each phone . (My personal opinion, i use the free web version of gemini and i have also use Chatgpt 4).
Cogent review. Each LLM has its own area of expertise. I source each based on the strengths that I have found. For my needs overall, GPT is the winner.
But why do you need to describe what image you gives, i mean if you find any image on the net and you don't have any clue about the content, and you want to ask that what is on the picture, why do you need to give the answer too, to get the answer.. if you already know the answer you won't ask it...
@@TamasKiss-yk4st Well, that's a fair point but, if you find a image which you don't know anything about , then, you could state the prompt in a different manner like, " Search the web , to find the source of this image , give me the link of the source " something like this. 😀
Gemini is like a fan of something tried to recreate that something as accurate as possible so much that they basically create a doppelganger of it that looks slightly off from the original
I fed them both the same prompt to see what the monthly principal payment would be to pay my mortgage off in five years. Gemini initially gave me a website with a mortgage calculator. ChatGPT noticed I forgot to provide the current interest rate with my prompt and upon providing that information gave me a detailed breakdown. I adjusted my prompt for Gemini and it gave me just the answer in what the amount would be. I preferred ChatGPT’s detailed response. I primarily use Gemini and defer to OpenAI and Anthropic when Gemini acts up.
Who in their right mind would even consider paying for 1 million tokens!? That's useless marketing bs, when as little as 4K tokens is enough for 99.9% of all use cases.
If you don't know what that is then let me share you about it, "The Gemini context window refers to the amount of information an AI model can process at once. A larger context window allows for better understanding and reasoning across various forms of information like text, images, videos, audio, and code. "
And you pay for each token. Google's pricing is 7 $ / 1 Million Tokens for each input with more than 128K tokens. So 10 interactions would cost you 70 $. 💸💸💸
@@Alex-nk8bw Then have gemini advanced, they have gemini 1.5 pro on there now. And you're not required to use 1 million tokens all at one place. 1 Million is just for convenience.
My daughter is a microbiology student. In her lab work, she frequently has to count the number of bacteria colonies in her Petrie dishes by hand. I was sure taking a photo and uploading it to ChatGPT would be the way to go. But ChatGPT couldn’t do it.
Are you sure GPT-4o is not using Mixture of Expert? Are you sure the whole GPT-4 family of model is not MoE based? Where did you get the information is not? Using MoE doesn’t mean the model is not using Transformers architecture as MoE Expert are still using Transformers underneath. If your information are not correct will you agree to call it a human hallucination?
From the developer stand point, google Gemini has a governance problem trying to please many. Because most of the issues noted with Gemini are in the native context prompting. and chatgpt's governance is more lax and they easily agree on the best appealing presentation. If you noticed Gemini always avoids making a decision for you because that is how google's governance wants to present their AI. and I like the fact that you said "in the regular persons perspective". in development, they are not much different if you prompt them well.
My experience it refuses to answer or just repeats whatever I say as fact. Although all of them are pretty bad as consumer facing products currently. Certainly not worth the energy involved in terms of sustainability. Emissions are going to go up 35 40% for this?
Gemini is not a standalone conversation A.I like chatgpt who is meant to generate human like , creative responses with minimal outputs . Gemini is an A.I created to further enhance the user experience of Google products (is evident due to their integration into Gmail, UA-cam, Google searches) , it's like how apple created products which just seamlessly integrate with other apple products.
@@carlmathews5778 which is obvious 😅 , I meant like the base aim of chatgpt is being standalone (In a sense) , it's obvious that they are gonna customize it for buisness .
@@sangeetdass6844 You said what you said. ChatGPT that you keep referencing is for consumer facing purposes. The concept of ChatGPT looks way different in the business space
For the floor test when adjusted for outside the US: There's a tall building with a magic lift in it. When stopping on an even floor number, this lift connects to floor 1 instead. Starting on the ground floor, I take the magic lift 3 floors up. Exiting the lift, I then use the stairs to go 3 floors up again. Which do I end up on? When I tried this, ChatGPT 4o failed and Gemini (free version) nailed it.
I have made a custom gpt through absolute endless hours, that is quite good at both text based analytical reasoning questions and nuanced images too, so it can get physics problems correct and also an example i’ve been using is upload an image of a cloud that looks like multiple things but it can be interpreted, the one i have made recognised it was a rabbit every time now on 1st shot so it knows when something is unusual about an image even if you dont say anything is, it can also do iq test image reasoning pattern questions Its called Smarter Vision Multimodal image/text analysis Quite insane changing even a few things in the instructions can make the vision so much worse, just gotta follow the instructions given to get the right seed its 1 in 2 chance or so i have absolutely no idea why it needs that.
GPT 4o started on floor 0 for me and ended up on floor 6. Wildly different answers for the same AI. I even took it on a factual tour to see if it could figure out it's mistake and it never did until I told it. I've had these kind of issues with both Chat and Gemini with all of their versions. In my opinion, this makes generative AI, as a source of information, worthless. You might as well just ask your bullsh!t artist friend, Billy, and then wait and see if he continues his hot streak as Schrodinger's Scholar.
This video should have just been called "I work for OPENAI and I use chatgpt you should too, the paid one of course, leave a comment if you are using the paid version so you can ask questions about 2023 or about the russo ukrainian war
AI is still at his infancy! This kind of comparison is senseless! Most youtubers are biased! They all criticize "Google"! Chatgpt has some weak spots too! Nobody talk about!
👉 Click to start UA-cam Channel with my team salebot.site/kozwin_yt_6?KxySfdKtm-k
Gemini Advanced is the best in the business at trying to gaslight you when it doesn't know the answer.
Absolutely right...especially that "I am just a large language model thing" It's quite frustrating.
That's not really gaslighting. gaslighting is manipulation. What it's doing is akin to bluffing but it actually doesn't know it's wrong.
😂 so true
lol people use gaslighting so freely these days. It’s not gaslighting you, it’s “hallucinating” the answer aka making it up and being confident about it. It’s a very common issues with AI chat bots. Even ChatGPT does it even when I shared the pdf documentation with it.
My experience in using both is that ChatGPT information is less words packed better information and more factual information. While Gemini talks bla bla nonsense and too much.
ChatGPT voice dictation is just insane, absolutely amazing. I have never seen anything like it, and Google voice dictation is just normal stuff, which means bad.
yes yes,
I agree with you. Chatgpt4o is the winner. In few years, chatgpt will be on top list in terms of number of users.
@@mudangkano5267 $, gpt-4o is a little pricy
try asking it to generate an mp3 for download... sounds like text2voice from 2010
@@antb533 I didn't get you!
Gemini has a lot of problems and sucks when you compare to ChatGPT. It often gets the prompt wrong or refuses to do it saying it's just a large language model. While other language models handle the same tasks without any problems. I paid for both services so I used their latest models. Gemini lecture you on almost everything. Whether you are researching information or it thinks you wrote something wrong. It provides the least accurate information in my opinion. Why pay $20 a month for Gemini when other LLM are much better. They should have it priced at $10 a month because it's not worth the asking price of the subscription. Thanks for the video, I enjoyed it
Gemini is a novice and the primary focus is for the human centric approach. Which means the more is yet come and will have far lesser hallucinations.
There is more advanced update pending for public usage.
Right 😂 It's not worth paying for
they are giving away gemini pro bundled with 2tb of storage
@@mrak8948 As they should because it's a hot mess. I no longer see Google in the light i once did.
Hey, the purpose of gemini and chatgpt is different. Chatgpt is mostly based on its innate data that it has collected , with some browsing capabilities . While gemini is more like an extension to a Google search , in order to get good answers to you need ask questions in such a way that it utilizes it's google searching capabilities . Example , for the phone test you can instead write : I've given you specs for 3 phones , identify each of these phones (search the web , to find exact models of phones) . Compare these phones based on camera, screen , chipset and charging . Tell the strengths and weakness of each phone . (My personal opinion, i use the free web version of gemini and i have also use Chatgpt 4).
The funny thing is, you'd think the one thing Google would get right about Gemini would be it's help in shopping but it sucks at that too.
Cogent review. Each LLM has its own area of expertise. I source each based on the strengths that I have found. For my needs overall, GPT is the winner.
But why do you need to describe what image you gives, i mean if you find any image on the net and you don't have any clue about the content, and you want to ask that what is on the picture, why do you need to give the answer too, to get the answer.. if you already know the answer you won't ask it...
@@TamasKiss-yk4st Well, that's a fair point but, if you find a image which you don't know anything about , then, you could state the prompt in a different manner like, " Search the web , to find the source of this image , give me the link of the source " something like this. 😀
4o is the king especially when you work in the same chat for hours revising code. Its track of the context is magical
Gemini is like a fan of something tried to recreate that something as accurate as possible so much that they basically create a doppelganger of it that looks slightly off from the original
Google taking Ls 😅
I was considering switching to Gemini until I saw this video. Chat GPT 4.0 has really accelerated my coding. I do not want to lose that momentum
I fed them both the same prompt to see what the monthly principal payment would be to pay my mortgage off in five years. Gemini initially gave me a website with a mortgage calculator. ChatGPT noticed I forgot to provide the current interest rate with my prompt and upon providing that information gave me a detailed breakdown. I adjusted my prompt for Gemini and it gave me just the answer in what the amount would be. I preferred ChatGPT’s detailed response. I primarily use Gemini and defer to OpenAI and Anthropic when Gemini acts up.
Who in their right mind would even consider paying for 1 million tokens!? That's useless marketing bs, when as little as 4K tokens is enough for 99.9% of all use cases.
The context window of 1 million is probably what people would use gemini for.
If you don't know what that is then let me share you about it, "The Gemini context window refers to the amount of information an AI model can process at once. A larger context window allows for better understanding and reasoning across various forms of information like text, images, videos, audio, and code.
"
And you pay for each token. Google's pricing is 7 $ / 1 Million Tokens for each input with more than 128K tokens. So 10 interactions would cost you 70 $. 💸💸💸
@@Alex-nk8bw Then have gemini advanced, they have gemini 1.5 pro on there now. And you're not required to use 1 million tokens all at one place. 1 Million is just for convenience.
@@Alex-nk8bw What you mean? Google One has 19 $ subscription that contain "Gemini Advanced + 2 TB storage and more...
Gemini failed when I asked to hash the number 1 with sha 1. It provided a random hash. When I asked what he hashed, he couldn't provide an answer 😂😂.
Gemini cuts my code in half most of the time, GPT has the option to keep writing.
Did he compare gemini advanced and gpt 4o or gemini 1.5 pro and gpt 4o? or maybe gemini 1.5 pro and gemini advanced are same?
Every time you put your hands on the table, there's a terrible sound from the microphone
Can easily say that chatgpt is crazy good at coding while I had so hard time with gemini.
Excellent video comparison! Thank you!
Very helpful, thanks!
As far as AI humor goes, I've seen worse results. I actually found the dark mode joke quite funny. 🙂
gemini was worst but getting better now ...
Yep, Gemini is improving every day, not like ChatGBT it's improving every 2 months
If we lay 5 shirts out in the sun and it takes 4 hours to dry, how long would 20 shirts take to dry? Explain your reasoning step by step.
As a college student that works with data, ChatGPT 4o Plus is miles better.
My daughter is a microbiology student. In her lab work, she frequently has to count the number of bacteria colonies in her Petrie dishes by hand. I was sure taking a photo and uploading it to ChatGPT would be the way to go. But ChatGPT couldn’t do it.
Are you sure GPT-4o is not using Mixture of Expert? Are you sure the whole GPT-4 family of model is not MoE based? Where did you get the information is not?
Using MoE doesn’t mean the model is not using Transformers architecture as MoE Expert are still using Transformers underneath.
If your information are not correct will you agree to call it a human hallucination?
Could you upload pdfs on Gemini?
Yes
From the developer stand point, google Gemini has a governance problem trying to please many. Because most of the issues noted with Gemini are in the native context prompting. and chatgpt's governance is more lax and they easily agree on the best appealing presentation. If you noticed Gemini always avoids making a decision for you because that is how google's governance wants to present their AI. and I like the fact that you said "in the regular persons perspective". in development, they are not much different if you prompt them well.
My experience it refuses to answer or just repeats whatever I say as fact. Although all of them are pretty bad as consumer facing products currently. Certainly not worth the energy involved in terms of sustainability. Emissions are going to go up 35 40% for this?
Gemini is still the baby infront of chat gpt
Haha true 😂
baby always grow up to be better then the parent
@@hfdgsjhdagfhWho lie to you?😂
Gemini is not a standalone conversation A.I like chatgpt who is meant to generate human like , creative responses with minimal outputs . Gemini is an A.I created to further enhance the user experience of Google products (is evident due to their integration into Gmail, UA-cam, Google searches) , it's like how apple created products which just seamlessly integrate with other apple products.
Chat GPT is also not a standalone. That is just their public prompt screen. There are super advanced used case for Chat Gpt
@@carlmathews5778 which is obvious 😅 , I meant like the base aim of chatgpt is being standalone (In a sense) , it's obvious that they are gonna customize it for buisness .
@@sangeetdass6844 You said what you said. ChatGPT that you keep referencing is for consumer facing purposes. The concept of ChatGPT looks way different in the business space
For the floor test when adjusted for outside the US:
There's a tall building with a magic lift in it. When stopping on an even floor number, this lift connects to floor 1 instead.
Starting on the ground floor, I take the magic lift 3 floors up. Exiting the lift, I then use the stairs to go 3 floors up again.
Which do I end up on?
When I tried this, ChatGPT 4o failed and Gemini (free version) nailed it.
What's the answer
Ground floor@@mustang5431
Bto! Now we have Gemini 1,5 Flash & it's alot smarter. Google soon will be on ChatGPT level...just wait a minute.
That's weird my Gemini got the dot count correct
I have made a custom gpt through absolute endless hours, that is quite good at both text based analytical reasoning questions and nuanced images too, so it can get physics problems correct and also an example i’ve been using is upload an image of a cloud that looks like multiple things but it can be interpreted, the one i have made recognised it was a rabbit every time now on 1st shot so it knows when something is unusual about an image even if you dont say anything is, it can also do iq test image reasoning pattern questions
Its called Smarter Vision Multimodal image/text analysis
Quite insane changing even a few things in the instructions can make the vision so much worse, just gotta follow the instructions given to get the right seed its 1 in 2 chance or so i have absolutely no idea why it needs that.
GPT 4o started on floor 0 for me and ended up on floor 6. Wildly different answers for the same AI. I even took it on a factual tour to see if it could figure out it's mistake and it never did until I told it. I've had these kind of issues with both Chat and Gemini with all of their versions. In my opinion, this makes generative AI, as a source of information, worthless. You might as well just ask your bullsh!t artist friend, Billy, and then wait and see if he continues his hot streak as Schrodinger's Scholar.
Should've used an *android* for gemini
IQ test on AI please
He woke up, but she never did 🥺👉🏻👈🏻
i'm gonna win this...
This video should have just been called "I work for OPENAI and I use chatgpt you should too, the paid one of course, leave a comment if you are using the paid version so you can ask questions about 2023 or about the russo ukrainian war
Yandex Alice is 100 times better than CHATGPT and Gemini
BUT IT'S FEETHERS?!
Gemini is whack
blah blah blah
Gemini is trash compared to chatgpt
AI is still at his infancy! This kind of comparison is senseless! Most youtubers are biased! They all criticize "Google"! Chatgpt has some weak spots too! Nobody talk about!
Like what?
Way less than Google’s models I can tell you that.
some examples. Yes both does. But some examples for the normal person.
You gonna cry bro?
@@jeffman9260 Don't explain news of 2024, no real-time answers, don't give great answers of logic