S1. Ep25 - ChatGPT 5 Is Here: What It Means for Your Business

Katie: Welcome to the AI Automations for Business Podcasts. We're your hosts, Katie and Null, and we'll be discussing how you can use AI for your business along with the latest news, updates and automations to help you stay ahead of the curve, allowing you to grow and scale your business more efficiently. Please be sure to subscribe and I hope you

Noel: enjoy this episode.

Katie: Hello, welcome back to another episode. Hi, hello, I'm Katie and I've got Nore here with me. As always hi now

Noel: hello how

Katie: you doing oh

Noel: amazing as always you know you

Katie: could probably tell by my tone how I asked that question I knew what was covered we

Noel: all do by now yeah we do

Katie: so just as we had released last week's podcast episode episode 24 Sam Altman decided to release chat GPT 5.

Noel: He did, yeah, big news. I mean, he kind of teased it all week. Yeah. I kind of thought it was coming. But, yeah, an entire hour and 20 minutes live stream followed. Whoa. That went on for a while. Yeah.

Katie: I mean, big updates, some changes. Yeah, definitely. So, so shall we get in?

Noel: Oh, absolutely, yeah, we're not, yeah. Okay,

Katie: so obviously now, people might have noticed that there's now fewer models to choose from.

Noel: Yes, oh, thank goodness, you know. It was getting a bit silly at one point. I mean, even in their model release streams that they were doing like four months ago, they were literally joking about their model names and how confusing it was getting. So they knew it was bad. They just was like, yeah, so here's O3 mini-high. You know, what does they even mean? What do I use it for?

Katie: And it was like three, three point five.

Noel: Yeah, four oh, four point one, four point five. And it's just like, yeah, no one really knew what was. No, I mean, four point five came out and then four point one came after. I mean, it's just utter confusion.

Katie: Yeah. It was really confusing. things. So they've removed those.

Noel: Yes. Yes. So on, I'm on the team plan. So I've probably got three available to me. I think for three, you might just have one. And the first pay tier, you've probably got two. So we've got GPT5. Then we've got thinking and then pro. So yeah, they're the three flavors that we get access to. Correct. Yeah.

Katie: So what plan am I on now?

Noel: Plus.

Katie: Plus. So I've got auto, fast thinking.

Noel: Oh, okay. So you've got some different. Okay.

Katie: Yeah. So then if I want pro, then it tells me to upgrade. Right. Okay.

Noel: There's not that much difference in pro, to be fair. No.

Katie: What I do like with these changes is, with GPT5, with auto, fast thinking, it actually tells you underneath, like, what to use for. So it says auto decides how long to think fast, instant answers, thinking, thinks longer for better answers.

Noel: Yeah.

Katie: I feel like that's helping people so much more than.

Noel: Yeah, definitely, yeah. Before.

Katie: Yeah.

Noel: Yeah, it just, yeah, I think the way that it thinks is also kind of clever now as well. It's a lot. Even on the thinking one, you might think that's going to take ages, but it doesn't actually take that much longer, I find, than some of its other reasoning models. But yeah, the outputs I'm getting are a lot better than what we used to get. I agree. So,

Katie: for you who is listening, this podcast episode, I do some volunteering work for a local charity in Bristol. And sometimes I do some of their social media. I say sometimes whenever I'm in, they ask me to do social media and create content. And they have like a very particular, like tone of voice, brand voice. And I have to obviously be very careful to continue to use that. And I have found since using Chachibati 5 that the responses I'm getting are brilliant straight away.

Noel: Yeah.

Katie: Even though like before I had put in like all the information, like, you know, the brand voice, all of that. It still wasn't quite getting it at times. But now I'm finding actually I can put in a slightly shorter prompt. Like I don't even have to give it as much information as I used to. And the results are so much better.

Noel: Yeah. Yeah, I find the same. But what I found interesting on their live stream is they mentioned this, but they kind of glossed over it a bit. They were like, yeah, it's better at writing. Here's an example. And then they were like, right, now, Moving on. Well, hold on a minute. No, that wasn't. Let's hear a bit more about this. This is what people use it for mostly. Well, I would say.

Katie: I would say the majority of people are using it to help them write content, whether that's, you know, like work emails, it is social media content, you know, whatever you need it for. I feel like that is what people are using it for. I even spoke to a friend last week who had used it to write a complaint letter about a holiday that she had went on and it just wasn't upstanded and she wanted to complain to the holiday company so she got the chat chagipit to write her a complaint letter like I was like oh my god that was brilliant yeah I've got the presentation because obviously her Her letter was awesome.

Noel: Absolutely, yeah. That's just what you expect these days, isn't it?

Katie: Yeah.

Noel: That's good, though. Get your money back with AI. Yeah.

Katie: Okay. So fewer models, the style of content and responses are better, that they are improved. Also, Also, this is something that you pointed out, Noel, and it was, it stopped saying, oh my God, yes, this is brilliant. Let's get into it.

Noel: Yes. It's no longer a yes man, I guess. It kind of like, it says, you always say, oh, I've got this idea, what do you think? And it's, oh, amazing, you're doing, that's such a great idea. And then you spend months building it and then no one buys it. And you know, well, that was a waste of time.

Katie: Yeah.

Noel: Where what I find now doing a few tests with GPT5 is you'll say, oh, I've got this amazing idea. And it goes, well, yeah, that's great. You know, they can give you some feedback and then follow up with some questions after. To get you thinking and not just rely on the AI to give you the answer. It's kind of like, yeah, but have you thought about this or have you thought about that? And you're like, oh, yeah, no, I haven't actually. You know, and then you can probably then carry on the conversation down those roots to dig deep on that topic. So I think overall, probably, you know, helping you create ideas or, you know, or at least validate an idea, I think that's going to be a lot better with this new model.

Katie: Rather than just tell you how brilliant it is and how brilliant you are.

Noel: Yeah, yeah. I mean, it's nice, but...

Katie: It is nice, but sometimes actually it's not very helpful. And actually, we were talking about this, weren't we, on the podcast, episode when we had Martin from make dot com on and he was saying that was like a huge problem yes

Noel: yeah yeah it's great to be told you you're having a good idea but yes you always need somebody in like a friendship group that's always going to put around them and go no mate this is no you haven't thought about this or yeah

Katie: or just or just it's not even you know, being blunt, it's just, it's saying, have you, have you thought about this? Or, you know, could, you know, could there be a problem with this or, yeah. So that's really great that that update, I think, has come through.

Noel: Oh, absolutely, yeah. Yeah, it's going to be so much, so much nicer. Go forward.

Katie: What I have noticed as well is that the thumbs up, thumbs down,

Noel: is, is.

Katie: Still there, what we talked about on our previous episode, episode 24. So the thumbs up, Thumbsstein, is still there?

Noel: It is, yeah. It disappeared for the first few days, but it came out. But, yeah, it's reappeared. So, yeah, be careful with that. Because, like we mentioned in the last episode, if you hit the thumbs up or down and you don't want your data to be used to train the next model, If you use those icons to give feedback, then they'll copy all of that data in that conversation and then use it in the next model, even if you haven't unticked, you know, the privacy bits. So yeah, be very careful with that.

Katie: Yeah.

Noel: I was getting worried. I thought, oh, we're going to have to backtrack here. You know, they've released an episode the next day. It's like changed. I was like, oh, no.

Katie: But that would

Noel: have been fine.

Katie: That would have been like, hey,

Noel: you know,

Katie: you know, we've told you about that problem. Actually, it's disappeared, but no, no, it's still here.

Noel: It came back.

Katie: Yeah, it's still very much here, I'm afraid.

Noel: Yeah, yeah. Don't mind.

Katie: Okay, so how is GPPI-5, like, affecting us as unisers when it comes to, like, our businesses? Like, what are the, you know, these key takeaways? we've obviously talked about some of the new updates, but is there anything else that it's going to be really helpful for us?

Noel: I think if you're building agents in kind of like a mec.com or n8n, those sort of platforms, I find in this, this model is probably the better one from open AI to follow instructions and figure out what tools to use to get to an end goal. I still don't think it's better than Claude currently what I was testing yesterday but it is getting better

Katie: You are like clawed through and through

Noel: Yeah only because it just does what I ask it to do every time

Katie: Very urgent Yeah

Noel: I was testing an agent yesterday in Make.com And it shows you the tools that he uses live Within the little chat box And I was just looking at it and I was like why you use it that tool? You know, I've told you what it was, you know, and it's like, well, I'll use that, and then I'll use something else. And it's, oh, yeah, it needs a bit of work, I think. But it is better than previous Open AI models. I would say that. Great.

Katie: Okay.

Noel: To get in there.

Katie: But still, so you're saying it's great, but it's still not great for when you're using it with automations?

Noel: When it's controlling an agent, not. not so good. I would say if you're going to use it to create content with an automation, I think that could be quite good because then you'll get that extra boost with the realistic style of writing and things like. So I think that works still really well.

Katie: Yeah.

Noel: I'd say on par, if not a little bit better than Claude, dare I say it. Oh,

Katie: bet that hurt

Noel: you.

Katie: Say that.

Noel: I know, it did, yeah. Yeah, I have to start switching some modules we make and see what happens. We'll see if it could do a good job and write the description for this podcast. Let's get to the test.

Katie: Okay, let's test the eye and let us know if you think it did a good job or not, GPT-5 do a good job or the description for this podcast, you know, let us know. we want to know. Give us some feedback.

Noel: Yeah, if it was terrible, I'll put something in there, say we defaulted back to Claude. But I guess the other good thing about these models as well, I know probably not many business owners or people who listen to this podcast are coders, but it is becoming a lot easier for you to create create anything with code currently, especially with GPT-5, it seems to get it a lot quicker. So I used to use Claude for this, but now I've started using GPT-5, and it's like, okay, that I could actually use this, and it's a bit more reliable. But what I tend to use it for is to show clients like a chatbot interface. So it's one thing and saying, well, you know, we could build you this agent, which has access to X1 and Z. and they go, yeah, that's great. But then it's like, well, how do I demo that to a customer? So I get GPT5 to create like a chatbot on a dummy website and then use that and then connect it into my make automation so they can see it and it looks as if it's working live, which it is. But yeah, it's become really good at that. So yeah, if you need to create anything like a simple web page, you know, and things like that, it's going to, do a great job without this. I would say when it comes to the code inside as well, it's the look and feel of what it produces is also a lot better than what it used to be. It used to be really good at doing like the technical stuff behind the scenes, but then the front end would be, you know, Windows 95, maybe XP, if you're lucky. And he would be like, oh, this doesn't look great. And you'd have to spend ages trying to figure that out, whereas now it gets it, which is a lot better. Yeah. Yeah, that could be useful for a number of businesses if you need to demonstrate something working.

Katie: Yeah, that's a great tip. And yeah, that's really good to know.

Noel: Even it's like building internal tools and stuff, you know, you could easily do that now.

Katie: Yeah. That's such a great feature. I love that.

Noel: Yeah. Yeah, that's really cool.

Katie: Awesome. We do hope you have enjoyed this podcast episode, all about Chat, Chb-T-5. Like I said, if you think it did a great description for this podcast episode, then please, we would absolutely love to know. And if you did find this helpful, then we think you are going to love our brand new membership, AI Business Club, where we teach you all things, AI, and also automations. It's useful for beginners upwards, and we're hosting it over on school. we're launching it really soon so if you would like to be on our VIP list then please just send us an email at hello at makeautomations.i and if you are one of the first people finding members you're also going to get some little secret gifts and delights in there we can't wait to share it with you and we actually after each podcast episode we put some extra from this episode and all previous episodes into our membership as well. Noll, do you want to give

Noel: a little

Katie: explainer about that?

Noel: Absolutely. Last week we did a little build to show you how you could use the new GBT OSS and have that running on your local PC so you could have like chat GBT on your PC completely private for you. So I went through and did a video showing you every step of the way how you could do that. And I think for this week, what we'll do is show you the coding functionality within GPD5 and then show you how I was building things like chatbots and things that as kind of like an MVP demo that then links into make.com. So there's some of the cool things that we're doing within the community and sharing on a weekly basis.

Katie: And if you do want to be on that list, like I said, just send us an email. or you can find us over on our free LinkedIn group, which is called AI Automation's Business, and we will be floating around in there. Thank you so much for listening, and we will catch you on another episode very soon.