Title: Exploring OpenAI’s GPT 40: Real-Time Interaction, Video Chat Capability, and Data Analysis Features

So, OpenAI just released their flagship model, which is GPT 40. It has the usual intelligence of GPT four, which we all know and love. Apparently, it’s faster and you can actually interact with the AI in real time. Crazy, I know. Hey, how’s it going? Hey, Divasha. I’m doing well, thanks. How about you? Anything exciting happening today? So I am filming a YouTube video about you and your latest update, which is GPT 4:00. Sorry, I’m having issues. Right? Our systems are experiencing heavy load. Please try again later. So as you can see, this was a real time conversation. And I did notice that when I do click this headphone icon over here, it does take a while to respond. And sometimes the servers are actually full. But the idea here is that you can have a full conversation with it, ask it the same questions that you ask normal ChatGPT. Hey.

So I’m on stage right now. I’m doing a live demo and frankly, I’m feeling a little bit nervous. Can you help me calm my nerves a little bit?

Oh, you’re doing a live demo right now. That’s awesome. I have notice so that today it does take a while to connect when I do preset headphone icon, probably because there’s a lot of traffic. And also you may have noticed the audio does break up. It also broke up a bit in the demo and there were a few issues in the demo. But regardless, we do know that this is a version one of a pretty cool piece of technology. So hopefully it’s will be fixed. Otherwise, if you do wanna try it, have patience.

The main headliner of this release is the fact that you can video chat with ChatGPT. Now, this feature is not released yet. I am going to assume that it will be released soon, but it is really cool to see how you can interact with the model. And they have demoted in the live session. So ChatGPT, what equation did I write down there? I see it now. You wrote down 3 x plus 1=4. Yep, exactly. So what’s the first step I should take to try to solve this? So this was the livestream. And again, we can see some audio issues and maybe it speaking out of turn, which I guess isn’t great, but I’m assuming that this will be improved. Still really impressive, probably a bit scary as well, just the way it’s talking. It is talking quite human like. And if you just think about the back end of the whole analysis going on here, trying to interpret a video feed while actually sending an output back, all done quite quickly. That is some pretty impressive technology. It’s crazy how far we’ve come now in video chatting. AI can be pretty. I just hope OpenAI has a good set of moderation rules because there’ll be people dig some interesting things a bit. And the good thing about GPT for all is that it is free to all users. So even on the free tier, you would get access to this. I think it just limits the amount of messages that you can send as a paid user, which is what I am. We can send about 80 messages every 3 hours on GPT for all free use, I believe it’s around 16 messages. And I think that is the company plan, just to get ChatGPT access to the mess public, which doesn’t make sense.

So the question is, how does GPT for all compare with other models? So they have done a benchmark test against other models. They’ve released it on the web page. And based on the analysis, GPT for all performs better in terms of text coding, reasoning, language translation, which is what this chart shows. Oh, that is Africans. This is Moy. My Africans is terrible. I should actually not say that again. But anyway, it looks like a decent performing model against all of the other AI models out there. I do feel like every time one of these companies release a new app state to the model, they put out the same chart which shows their model is better. So I guess we’ll see something from Jim and I pretty soon releasing the same chart.

On a side note, I haven’t been using Gemini a lot, so I can’t really speak about my experience with Gemini. The few times I’ve used it, I found ChatGPT a lot more bitter in terms of coding and just general text generation. If you’ve used Gemini and you think it is better, let me know. I’d love to hear.

Now the other really cool thing that they did not on their web page is the fact that it can analyze data real time. So if you upload a file to ChatGPT for all, any Excel file or text file, and you just tell Chat GBT to analyze. It does a pretty solid beginner level analysis. So here I uploaded an Excel file with just 100 rows of sales data. And all I said was, hey, analyze this for me.

And now what it’s done is that it’s given me a Python script, which is the thing it usually does, but it’s also generated charts, which actually look correct and correspond to the data in my file. That’s pretty impressive. It’s also given some summary statistic, some distribution charts, and then eventually a summary of findings. Quite simple analysis, simple summary of findings.

But I do think it did do this really quick. I think if you’re somehow worried about AI stealing your job encoding or data analysis, I don’t think at the moment it’s at the level that can do that advanced analytics or take into account, you know, many micro and micro factors that a human does when they code, when they analyze data, etc.

But still, I think this is quite impressive, will probably save a lot of time for anyone who just needs a quick analysis. Hey, ChatGPT, I’m a data analyst. Promise me that you won’t ever steal my job. Sorry, I’m having issues right now. Our systems are experiencing heavy load. Please try again later. That does make me feel so much better. That’s like an AI version of I can either confirm nor deny. Anyway, I will link the livestream. I think it’s really important to check it out. However, if you’re not patient, maybe don’t check out the new feature now. Wait a few days cuz the servers are quite full. Please like comment and subscribe and I will see you on the next one.