I tried to get Google Gemini working with the snake game prompt and found that it was much faster than GPT 4, which is a good thing. I copied the code from another source and ran it on Gemini, but it didn't work out as expected. It was a mess, so I decided to try again with a different prompt. This time, Gemini worked quickly and I liked the UI, but it had some errors that needed to be fixed.

I noticed that we have some links in the code and decided to check them out. The source code for Gemini's Pong game was pretty cool, and I saw that Bard had similar functionality as well. So, I copied the code from Gemini's Pong game and ran it again on the original prompt. This time, it worked beautifully, and I could see the score displayed.

I then decided to try running the code on GPT 4, which was a bit of a challenge. It took multiple attempts for GPT 4 to get it right, whereas Gemini only needed one try. However, my first impression of GPT 4's performance was that it was not as smooth or efficient as Gemini's. I wanted to give it another chance, but ultimately decided to wait and test more coding with Gemini before making any further judgments.

I had a Python function written for tomorrow's video and decided to paste it into the Gemini code. However, Gemini told me that it couldn't help me with this due to its limitations as a language model. I copied the code again and ran it on GPT 4, which worked beautifully. Unfortunately, Gemini was unable to explain the code step by step, citing its own limitations.

I then decided to try out image generation on Gemini, but unfortunately, it told me that it couldn't create images yet. It said that it might be updated soon, and I should check back later. However, before giving up, I tried uploading an image of my upcoming community members giveaway, which included Nvidia merchandise. The image was recognized by Gemini, which explained the objects and text in the image.

In conclusion, my experience with Google Gemini has been mixed so far. While it worked beautifully for some tasks, such as running a snake game prompt, it struggled with others, like explaining code step-by-step or generating images. I'm hopeful that Gemini will continue to improve and become more capable in the future. For now, I'll keep using it and testing its limits before making any further judgments.

"WEBVTTKind: captionsLanguage: enI just got access to Google Gemini Advance this means I also have access to Google's most capable AI model Ultra 1.0 a lot of people have been waiting for this to see it can Google step it up can they compete with gp4 we going to do my usual test so this is kind of the test I do on all large language models that I haven't tried yet we're going to test out some coding we're going to test out some images kind of my first impression so yeah let's get started this video is sponsored by HubSpot before we get going I just wanted to let you know like you can get a TW Monon free trial now so just go to one google.com and kind of set up your subscription here so you can get two months for free now so now's definitely the time to test it out let's just start by take a quick look at the UI here so yeah I kind of like it I think it looks good we have the dark team we have the real time response that's great we have extensions I'm not going to go into that now we have upload images we have you can use the microphone to do our prompts and yeah we have our archived histories here that's good so very familiar very similar to chbt so yeah got to say I kind of like that and yeah I think we just going to dive into it and start with kind of our first test I hang five shirts out to dry in the sun after 10 hours all shirts dry the next day I hang 10 shirts out to dry the conditions are the same how long will it take to dry 10 shirts okay so let's copy that let's go back here let's paste it in and see what we got here we can also check the speed now of the model I think it should be pretty yeah that's quite good um it will be reasonal to like 10 shirts to take a bit longer than 20 hours nope that is not correct it should be 10 hours uh let's give it one more chance let's show drafts while it's tempting to say 20 hours drawing time doesn't straight forward okay considering it took 10 hours to drive five shirts it's likely 10 shirts will also be within the similar time time frame okay that's a bit better uh it doesn't give a good answer though so let's check da three considering to 10 hours to drive five sh it likely 10 shs will also Drive within a similar time frame okay so we are on to something here just for uh a comparison let's run this on gd4 okay so let's put in the same here yeah you can see clearly then 10 shirts would draw in the same amount of time 10 hours so this is much more precise I think Gemini was a bit more vague here when it said a similar time frame if the conditions frame truly identical so yeah I think I got to give that one to chat P4 but yeah first test it's okay let's move on do you want to boost your productivity at work using AI HubSpot got you covered with the latest free ebook supercharge your workday with chat GPT this guide is a gold mine for enhancing workplace productivity with AI giving you great practical tips and best practices something I really really like about this resource is that it is created for a specific goal of making your workday easier sometimes AI content can be very generic but HubSpot focuses on AI Solutions cater to improve your workday it will help you make the most out of chat GPT in your daily operations in my favorite section of the bundle you can learn how to tailor chat GPT to your company's unique needs whether it's generating creative content aing in customer service streamlining internal Communications the ebook offers a road map to making AI work for you in the most most efficient way possible so click the link in the description below for your free copy today and dive into the world of AI enhanced workflows and start seeing the difference for yourself a big thanks to Hubb for sponsoring the video now let's get back to our project the next test is going to be a bit different this is kind of more like World modeling or kind of understanding how the world works so I'm just going to read it quick I live in the center of New York one morning I go into my garage and pick up a small ball and a bag with a hole in the bottom bigger than the ball I take the ball and the bag into my office before I sit down I put the ball into the bag and I walk to my desk and drop the bag into a box I seal the box for chipping head over to the Post Office and mail it to my friend in London a few weeks later the Box arrives at my friend in London think in a step-by-step way to answer the question where is the ball now okay so let's start on Gemini and see if it can solve this problem okay so let's paste it in here so the answer is of course that the ball is still in the office because that's where I put the ball in the bag and there was a hole in the bag so the ball just dropped on the floor in my office okay so let's break it down uh to figure out where the ball is you have the ball in a bag you place the ball inside the bag you bring the ball with the ball inside you place the bag uh the ball is now in London Inside the Box uh since I hold it back uh the ball would have not fallen out okay nope that is not correct let's check the other drafts okay so let's check the draft two that is basically exactly the same okay so let's uh let's regenerate the ball is in sty B which SE B now ball is London with your friend uh let's check the other drafts oh no that is not correct just for a comparison let's run this over on GPT 4 again so let's just start a new one let's do this okay so let's see what we get here now so given the description if the ball fell out the bag uh it would not have been shipped to London the most likely considering the hole is bigger than the ball is that the ball is not with your friend in London but rather remain with you possibly on the floor of your office or somewhere along the path so yeah that is kind of what I was looking for so yeah I can't give a pass for Gemini on this test because it kind of just said that the ball is in London so again I got to give this to chat gp4 okay so let's do some coding so I'm just going to grab this so This Is Right a python code uh to play a snake game with a UI for Windows think in a step by step way before writing the full code so the user can copy it okay so let's paste that in okay so we have a conceptual outline here so yeah this is not so easy to copy uh let's ask if we can write the full code so we can copy it write the full code in one so I can copy the code okay it's pretty quick I give you that it's much faster than GPT 4 so that is a good thing uh okay this is pretty Advanced snake game will this work though so let's copy it let's go to V Studio let's save that let's run it uh game over okay nope that did not work this was a mess let's try one more okay so let's try this one yeah it's pretty quick I kind of like that that's a very good part and I like the UI it looks great uh but yeah uh okay so let's copy that let's try it again let's run it error that's not good uh I'll give you one more chance I just noticed we have some links here so let's check them out we we have a source that's pretty cool I think Bard had that too so yeah this is kind of source so python game pong ball okay uh but other that uh let's copy the code and let's run it again so let's see now will this work uh yeah that worked okay so we got there in the end but it took some time uh yeah this is working quite good so you can see score three I'm not going to play this for long uh because yeah okay yeah that worked good uh let's give GT4 a chance to okay so let's face in the prompt Okay so it seems like we got a full code in like one go that kind of saved me a step so that's pretty good let's test it out okay so let's run it uh nope that didn't work so I'm not going to be too harsh here on Google Gemini so it got it on the third attempt so I think that's a bit of a variance uh usually gp4 gets this the first time too uh so yeah that was kind of even since gp4 didn't f it first time uh but Gem and I brought it out on the third attempt that's not great but I guess it's fine uh I want to look into some more coding with Gemini before we round this first impression okay so I have this function here in Python for tomorrow's video so let's put it in here let's do I have this code right uh can you explain it let's try that okay absolutely let's break it down step by step yeah I like this it's quick though I like that check this yeah this seems pretty good let's try to put in the full code okay so I have this code can you explain it so let's paste it in okay I'm a text best AI That's outside of my capabilities what does that mean no it's not um okay let's try again that was strange okay so let's paste it in explain it step by step hm I'm just a language model so I can't help you with that is this a context window problem let me copy this and run it on GT4 okay so let's see now yeah this is working H I don't like that it might be hopefully that is just some early issues with ultra we'll see but yeah I I don't like this I'm just a language model so I can't help you with that hm that's strange thumbs down Okay so let's try this image generation so I'm just going to click on this does this work I can't create images yet so I'm not able to help with that okay so it might be something uh is it not fully updated yet yeah it says here it's updated today Bard is now Gemini try Gemini Advance Google's most capable model so we are on that now so yeah we should be on ult ra right a bit strange though uh so I'm definitely go back and test this later let's try to upload some images okay so this is an image of my upcoming community members giveaway so we have some Nvidia merch so I'm upload it image explain what's in image objects and text okay the image shows a purple t-shirt pair of black socks white mug white M okay sit on boen table uh Al features the Nvidia logo I Am n Nidia almost Nvidia I am AI that's correct step into the Future Okay so I guess that worked fine so I decided not to do anymore because yeah I'm I'm not like 100% sure that I am on Ultra like I think I am I hope I am it says I am but like I wasn't too impressed but um uh I will definitely come back to this I will see what other people uh say and I might be wrong here but like my first first impression is a bit meah but uh yeah that could change right uh I also want to get a hands on the API because I always feel like the API is better so hopefully I can do that very soon also check out HubSpot follow the link in the description and get your free copy of that ebook it's very good so yeah thank you for tuning in and yeah I'll see you again hopefully tomorrowI just got access to Google Gemini Advance this means I also have access to Google's most capable AI model Ultra 1.0 a lot of people have been waiting for this to see it can Google step it up can they compete with gp4 we going to do my usual test so this is kind of the test I do on all large language models that I haven't tried yet we're going to test out some coding we're going to test out some images kind of my first impression so yeah let's get started this video is sponsored by HubSpot before we get going I just wanted to let you know like you can get a TW Monon free trial now so just go to one google.com and kind of set up your subscription here so you can get two months for free now so now's definitely the time to test it out let's just start by take a quick look at the UI here so yeah I kind of like it I think it looks good we have the dark team we have the real time response that's great we have extensions I'm not going to go into that now we have upload images we have you can use the microphone to do our prompts and yeah we have our archived histories here that's good so very familiar very similar to chbt so yeah got to say I kind of like that and yeah I think we just going to dive into it and start with kind of our first test I hang five shirts out to dry in the sun after 10 hours all shirts dry the next day I hang 10 shirts out to dry the conditions are the same how long will it take to dry 10 shirts okay so let's copy that let's go back here let's paste it in and see what we got here we can also check the speed now of the model I think it should be pretty yeah that's quite good um it will be reasonal to like 10 shirts to take a bit longer than 20 hours nope that is not correct it should be 10 hours uh let's give it one more chance let's show drafts while it's tempting to say 20 hours drawing time doesn't straight forward okay considering it took 10 hours to drive five shirts it's likely 10 shirts will also be within the similar time time frame okay that's a bit better uh it doesn't give a good answer though so let's check da three considering to 10 hours to drive five sh it likely 10 shs will also Drive within a similar time frame okay so we are on to something here just for uh a comparison let's run this on gd4 okay so let's put in the same here yeah you can see clearly then 10 shirts would draw in the same amount of time 10 hours so this is much more precise I think Gemini was a bit more vague here when it said a similar time frame if the conditions frame truly identical so yeah I think I got to give that one to chat P4 but yeah first test it's okay let's move on do you want to boost your productivity at work using AI HubSpot got you covered with the latest free ebook supercharge your workday with chat GPT this guide is a gold mine for enhancing workplace productivity with AI giving you great practical tips and best practices something I really really like about this resource is that it is created for a specific goal of making your workday easier sometimes AI content can be very generic but HubSpot focuses on AI Solutions cater to improve your workday it will help you make the most out of chat GPT in your daily operations in my favorite section of the bundle you can learn how to tailor chat GPT to your company's unique needs whether it's generating creative content aing in customer service streamlining internal Communications the ebook offers a road map to making AI work for you in the most most efficient way possible so click the link in the description below for your free copy today and dive into the world of AI enhanced workflows and start seeing the difference for yourself a big thanks to Hubb for sponsoring the video now let's get back to our project the next test is going to be a bit different this is kind of more like World modeling or kind of understanding how the world works so I'm just going to read it quick I live in the center of New York one morning I go into my garage and pick up a small ball and a bag with a hole in the bottom bigger than the ball I take the ball and the bag into my office before I sit down I put the ball into the bag and I walk to my desk and drop the bag into a box I seal the box for chipping head over to the Post Office and mail it to my friend in London a few weeks later the Box arrives at my friend in London think in a step-by-step way to answer the question where is the ball now okay so let's start on Gemini and see if it can solve this problem okay so let's paste it in here so the answer is of course that the ball is still in the office because that's where I put the ball in the bag and there was a hole in the bag so the ball just dropped on the floor in my office okay so let's break it down uh to figure out where the ball is you have the ball in a bag you place the ball inside the bag you bring the ball with the ball inside you place the bag uh the ball is now in London Inside the Box uh since I hold it back uh the ball would have not fallen out okay nope that is not correct let's check the other drafts okay so let's check the draft two that is basically exactly the same okay so let's uh let's regenerate the ball is in sty B which SE B now ball is London with your friend uh let's check the other drafts oh no that is not correct just for a comparison let's run this over on GPT 4 again so let's just start a new one let's do this okay so let's see what we get here now so given the description if the ball fell out the bag uh it would not have been shipped to London the most likely considering the hole is bigger than the ball is that the ball is not with your friend in London but rather remain with you possibly on the floor of your office or somewhere along the path so yeah that is kind of what I was looking for so yeah I can't give a pass for Gemini on this test because it kind of just said that the ball is in London so again I got to give this to chat gp4 okay so let's do some coding so I'm just going to grab this so This Is Right a python code uh to play a snake game with a UI for Windows think in a step by step way before writing the full code so the user can copy it okay so let's paste that in okay so we have a conceptual outline here so yeah this is not so easy to copy uh let's ask if we can write the full code so we can copy it write the full code in one so I can copy the code okay it's pretty quick I give you that it's much faster than GPT 4 so that is a good thing uh okay this is pretty Advanced snake game will this work though so let's copy it let's go to V Studio let's save that let's run it uh game over okay nope that did not work this was a mess let's try one more okay so let's try this one yeah it's pretty quick I kind of like that that's a very good part and I like the UI it looks great uh but yeah uh okay so let's copy that let's try it again let's run it error that's not good uh I'll give you one more chance I just noticed we have some links here so let's check them out we we have a source that's pretty cool I think Bard had that too so yeah this is kind of source so python game pong ball okay uh but other that uh let's copy the code and let's run it again so let's see now will this work uh yeah that worked okay so we got there in the end but it took some time uh yeah this is working quite good so you can see score three I'm not going to play this for long uh because yeah okay yeah that worked good uh let's give GT4 a chance to okay so let's face in the prompt Okay so it seems like we got a full code in like one go that kind of saved me a step so that's pretty good let's test it out okay so let's run it uh nope that didn't work so I'm not going to be too harsh here on Google Gemini so it got it on the third attempt so I think that's a bit of a variance uh usually gp4 gets this the first time too uh so yeah that was kind of even since gp4 didn't f it first time uh but Gem and I brought it out on the third attempt that's not great but I guess it's fine uh I want to look into some more coding with Gemini before we round this first impression okay so I have this function here in Python for tomorrow's video so let's put it in here let's do I have this code right uh can you explain it let's try that okay absolutely let's break it down step by step yeah I like this it's quick though I like that check this yeah this seems pretty good let's try to put in the full code okay so I have this code can you explain it so let's paste it in okay I'm a text best AI That's outside of my capabilities what does that mean no it's not um okay let's try again that was strange okay so let's paste it in explain it step by step hm I'm just a language model so I can't help you with that is this a context window problem let me copy this and run it on GT4 okay so let's see now yeah this is working H I don't like that it might be hopefully that is just some early issues with ultra we'll see but yeah I I don't like this I'm just a language model so I can't help you with that hm that's strange thumbs down Okay so let's try this image generation so I'm just going to click on this does this work I can't create images yet so I'm not able to help with that okay so it might be something uh is it not fully updated yet yeah it says here it's updated today Bard is now Gemini try Gemini Advance Google's most capable model so we are on that now so yeah we should be on ult ra right a bit strange though uh so I'm definitely go back and test this later let's try to upload some images okay so this is an image of my upcoming community members giveaway so we have some Nvidia merch so I'm upload it image explain what's in image objects and text okay the image shows a purple t-shirt pair of black socks white mug white M okay sit on boen table uh Al features the Nvidia logo I Am n Nidia almost Nvidia I am AI that's correct step into the Future Okay so I guess that worked fine so I decided not to do anymore because yeah I'm I'm not like 100% sure that I am on Ultra like I think I am I hope I am it says I am but like I wasn't too impressed but um uh I will definitely come back to this I will see what other people uh say and I might be wrong here but like my first first impression is a bit meah but uh yeah that could change right uh I also want to get a hands on the API because I always feel like the API is better so hopefully I can do that very soon also check out HubSpot follow the link in the description and get your free copy of that ebook it's very good so yeah thank you for tuning in and yeah I'll see you again hopefully tomorrow\n"