Loading summary
Transcript1 lines
- [00:00]
A
Welcome to the AI Chat podcast. I'm your host, Jaden Schaefer. Today on the podcast I'm going to be breaking down the Latest news from OpenAI, which is their search feature. They've released some really interesting new competition to Google and also Perplexity, which is a built in search function right inside of Google. It's kind of what they've been calling dubbing search GPT for a while. It's finally out. I'm going to demo it, talk about it and say some of the big implications that I think this has in the field, which I think is going to be pretty exciting. Before I get into that, if you've ever wanted to make money with AI tools or you want to grow and scale your current business startup, whatever you're currently working on your projects with AI tools, I would love to have you as a member of the AI Hustle school community. This is a community that I started and every single week I release exclusive content that I don't post anywhere else. I share all of the tools that I'm using to grow and scale my own software and personal companies and businesses. And we have over 200 members that are from, you know, that have started a hundred million dollar companies to brand new startups. You get a wide range of opinions and feedback and ideas on different side hustles to start different ways to make money. So anyways, it's an amazing time, it's an amazing community and I would love to have you as a member of it. So this sounds interesting to you. It's 19amonth. We're going to raise the price eventually, but if you get in at 19amonth, you will never have the price raised on you. You're locked in for forever. So you can check it out, the link is in the description. Now let's get into what OpenAI is doing with their new search feature. So the first thing I want to do is just show you a little demo of or and talk about a demo or when I was testing this out essentially. So the thing that's kind of interesting, if you say, you know, like what's the weather like in Madrid, it's going to give you a breakdown. Do you remember back in the past, oh my gosh, it seems so long ago when they're like, I'm sorry but my cutoff window is, you know, 2003 in September, whatever. And it was just kind of this annoying thing and before that it was 2001, this cutoff window. Okay, so now we're getting live stuff and I was Trying to do like the most live changing thing so you, you know, we could really tell. This is super accurate. So tells me what the weather is like in Madrid. Madrid as of today, as of the exact time 5:45pm Friday, when I did the search. So to me, this is very, very interesting. It's. They kind of have their own UI for like, weather stuff that I think they've built in. That's cool. I asked what the score for the latest hockey game in Vancouver was, which is Vancouver Canucks. It breaks down the scores of the hockey games that have recently happened as of, you know, yesterday's game when The Canucks lost 0 to 6 to the new Jersey Devils. Canucks are my home team, so it's a little painful, but, you know, we'll survive. And then I was like, okay, well, so we've done weather, we've done sports, let's do some politics. I said, who's leading in the polls in the US Election? Now, this is the response to this, I think is very interesting and very useful for whatever you're working on. So it said, you know, as of November 1st, the whatever, it's exceptionally close. There's like a lead in this poll for Harris and versus Trump. However, this poll indicates that Trump has a narrower margin. Okay, whatever. Who cares about the politics of it? The interesting part is after each of the statements, they have a link to the exact same study or poll or news article that this was pulled off of. So this is similar to some stuff that we've seen happen of perplexity, right? They source everything. OpenAI is now doing this and it's really, really interesting. And then right below that, they have actual news articles with more information on those topics. So a lot of things are great about this One is I love seeing the sources. I love seeing news articles about whatever topic you're talking about. But there's another use case that I want to show you. I'll do it live. I said, write me an outline on a training for executives on how to have better communication skills. And instead of clicking the search button on ChatGPT, I'm going to click the search the web button. So we kind of have, we have two buttons now and I believe, like, depending on what you ask, right? Like if you ask it for the weather or for sports, it's going to search the web and get you those things. But sometimes when you're like, hey, write me, you know, a description for my podcast, X, Y and Z, it knows it doesn't need to search the web, so it's just going to give it to you based off of its training data. But if you want it to search the web to give you more relevant information about that, and I'll give you a good example of that in a second, you can click this kind of search the web bar and everything that you ask it will be search the web. So an example of this is I always ask it to write, you know, like the descriptions on these podcasts. If you look at them, it's like in this podcast we talk about X, Y and Z. That is just chatgpt kind of writing a description just in case someone wanted a two sentence summary of what we're talking about in the podcast. Now the problem that I previously is I, I would essentially give it the title of the podcast and tell it to write a two sentence description about what's in the podcast and it's just basically guessing based off of the title what's in the thing. But every single podcast episode I'm talking about now is news, right? So there's all of these articles being written about this topic in the news. And so if I did search the web and had it search the web, when it writes my description, it can not only just get any information from the, I'm like just based off of this title, write me the description beyond just the title, it can now actually search the Internet and get relevant little bits of information that maybe weren't in the title and add that. And so now I know it's going to be more reliable, it's going to be more accurate. Anyways, this is fantastic. So with this search the web example that I did for this executive summary, it went and it searched the web and wrote kind of a description on how to write or how to, you know, have a training on, for executives on how to have better communication skills. But it didn't seem to find that there was in this. Like it didn't seem to think it really needed Google or the Internet too much for this. But what it did do was come up with a source for the information it gave me. So it didn't have sources cited throughout the whole response, kind of like it did with the news bit. But it did come up with a button at the bottom that says sources. And when I click on this list of sources, it shows me like 10 different links and search results that actually used to pull together to put together this response. So to me this is really interesting. You can see exactly where your data is coming from. I personally, if I'm being 100% honest, I like it better when I'M not. I like it better when it's citing the sources, but I get that sometimes it can be annoying when you're copying and pasting things and like the links are like embedded in there. So I could see how this could be a little bit annoying and why some people might not want to search the Internet. Personally. Yeah, personally I think it's great. But the big difference I want to say here is like, if I asked it the same question twice of kind of this like executive training thing that I wanted to do and I did it once where I asked it to do a search to come up with a response, and once where I didn't and where I didn't. Both of them worked, right? So both of them gave me a response. The one where I didn't, though, you can tell it's just chat GPTs, like whatever, it's trained in there and you're not going to get anything new or anything that's cutting edge. So if a new study came out yesterday or last week, it's most likely not going to be embedded. It's not going to be talked about here. So if you're talking about things that are newsworthy, cutting edge, new research, new information, you're going to want to use this search button every single time. Because I got a completely different response. Um, and I was able to see the sources of where all of these responses, you know, Harvard Business School, Online, mit, Sloan Executive Education and so at Coursera, Harvard. It's so fascinating sometimes. And another thing I would say is like, even if you don't care so much about the search results or citing the sources or anything, I think it's really useful to know where it's pulling your data from. It seems like a black box, but all of a sudden when you, when you have this search the web feature enabled, you're going to get the sources to everything it gives you, the black box disappears. And I would highly recommend this. If you're writing any sort of professional content or if you're trying to write a book or a post and you want to fact check and like test to make sure everything's legit, you're going to want to have this enabled because then you can see the sources on every response and make sure no hallucinations are happening. So overall, I think this is a fantastic feature. I'm super excited about it. I'll keep you up to date on anything new that OpenAI rolls out in regards to this because I think the battle right now is now between perplexity, which has been doing a lot of this stuff for quite some time and Google and a lot of people are saying, oh no, is. Is, you know, chat, GPT and OpenAI going to start squashing Perplexity and what they're doing? I don't think it is because at this point Google has such massive market share with a billion users that I think perplexity and OpenAI are actually going to just be fighting over Google's market share. Probably not so much from each other. Too much. That being said, that the AI community is like pretty tight and maybe some people that were using propelloxy before will be like, oh, I don't need this anymore, I'll go over to OpenAI. Perplexity has a ton of cool spaces, a kind of a ton of cool features that I think make it very unique that people would continue to use this. But and to be honest, I personally like how it cites its sources better. I like how it always cites its sources. You don't have to have a button enabled. Like there's a whole bunch of things I like that Perplexity is doing. I'm excited to see OpenAI jumping into the race, especially for people that might be listening to this that don't have a premium Perplexity account. But that being said, exciting times, both of these are going to be battling against Google. Who's going to be the main person losing market share to this? Google's in a bit of a tricky place. That's a podcast for another day I'll update you on, but thanks so much for tuning in. If you're watching on YouTube, drop a like and comment. I really appreciate it. Helps the algorithm recommend us. And if you're listening to this on Spotify or Apple, leave a review. It really would mean the world to me if you would take two seconds out of your day to leave a review. Helps motivate me to keep making these podcasts. Hope that you all have a fantastic rest of your day and I will catch you next time.