Resources:
Join the discussion: Ask Jordan questions on OpenAI
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Try Our Free AI Prompting Course: Register for our free Prime, Prompt, and Polish AI Course!
OpenAI's New O1 And ChatGPT Pro
Is your business ready to embrace the promising horizon of AI-dominated workplaces? Our world is witnessing an era where over 70% of information workers employ AI to amplify productivity. The landscape of work is evolving at an unprecedented pace, with 'shadow AI' emerging from the integration of tools outside of organizational provision. Yet, as we delve deeper, there is an intriguing contrast between the employees eager to adopt AI and leaders prudent regarding its integration and risks.
It's an exciting time in the AI space, with OpenAI announcing subscription plans for their ChatGPT model. The subscription plans entail a $20/month plan for ChatGPT Plus and a whopping $200/month O1 Pro plan. The notable improvement in the O1 Pro model's performance sets it apart, positioning it as a viable tool for professionals in software development, research, and analytics.
However, the introduction of the expensive O1 Pro plan has sparked concerns about the creation of an "elite club." This new development seems to indicate a shift from the earlier focus on accessibility in generative AI.
Performance Overview and Compatibility Challenges
The O1 Pro model outperforms previous versions in benchmarks like math and coding, showing immense improvements over the O1 Preview. However, a pertinent issue lies with the model's integration into various software and apps, as OpenAI plans to release an API for the O1 Pro model which is not yet available. Despite this, the O1 Pro model exhibits increased reliability as it now requires the model to answer queries correctly 4 out of 4 attempts.
Feature Additions and Disruption Potential
New features, such as an unlimited voice mode, previously available only in the $20/month plan, are now included in the $200/month O1 Pro tier. Moreover, users can now upload images to the new AI model to perform high-level data analysis. For example, the model processed 14,000 cells of YouTube metrics data rapidly, offering actionable insights for content creation.
These advancements could disrupt medium-sized consulting businesses, offering a cheaper and more efficient alternative. The O1 Pro model shines a light on the cost-effectiveness of AI when compared to hiring expensive consultants.
OpenAI's Strategic Positioning
Despite fierce competition from key players such as Google, Anthropic, Mistral, and Meta, OpenAI, with its high-quality models and fair pricing, has managed to maintain its edge. Its initial strategy to capture market share before delivering premium products seems to be paying off.
Future Prospects
With predictions of a new model release, GPT 4.5, between the upcoming US election and Q1 2025, and a "12 days of ship miss" event by OpenAI, the AI ecosystem could witness significant model and feature releases. It's an exciting period to observe the evolution of AI and its impact on various professions and industries.
While the new plans' pricing represents a distinct shift in OpenAI's strategy, it's important to note that this move also corresponds with their bid to secure financial stability in a fiercely competitive AI landscape. Despite the change, the existing $20/month plan will likely maintain its value rather than being reduced.
Topics Covered in This Episode
1. Details about the O1 Pro Model
2. O1 Model Capabilities and Applications
3. OpenAI's Pricing Strategy and Market Position
4. Product Releases and Future Predictions
Podcast Transcript
Jordan Wilson [00:00:17]:
Open AI has just released a new model. That's right. The new open AI 0101 pro, but that's not it. There's more. We also have ChatGPT pro, and it comes at a price, both on the wallet at $200 a month, but, also, I think this is really going to impact businesses. And is this separation needed? Is there now, an elite club? Right? I think that one of the exciting things over the past 2 years since ChatGPT was first released was that it democratized access to large language models for everyone. But $20 a month for the world's most capable model, I think, is reasonable for most people, small business owners, and everyone. But $200 a month for ChatGPT Pro and for the new OpenAI 1 and 1 Pro models.
Jordan Wilson [00:01:28]:
Is this a revolution, or is it a rip off? Well, we're gonna at least start to dig into that. Alright. What's going on, y'all? My name is Jordan Wilson, and welcome to Everyday AI. This is your daily livestream podcast and free daily newsletter helping everyday people like you and me learn and leverage generative AI to grow our companies and our careers. If you're tuning in live, you might be like, wait. What's going on? Yeah. Double dose today. Yeah.
Jordan Wilson [00:01:56]:
We already, had a podcast and livestream today. Our newsletter already shipped out the door. But speaking of ship, it is OpenAI's Ship Miss, the 12, 12 days of OpenAI. And it started today. And, you know, we knew that they were probably gonna release, something big on day 1, and they did not disappoint. OpenAI. So, yes, this is our second podcast of the day. So, thank you everyone who's who's tuning in for the evening edition.
Jordan Wilson [00:02:29]:
Absolutely. Peter and Tara and Nathan and Douglas and, doctor Harvey, thanks for tuning in. So, let's jump straight into it, y'all. Let's talk about what's new, what this 12 days of open AI even is, and we'll talk a little bit about this new, o one. It is a new model. So we had o one preview and o one mini. Now we have the o one full model, as well as an o one pro model, as well as the new ChatGPT Pro, which is $200 a month, which go back a year ago. I said that we were gonna see a version of this that was gonna be multiple $100 a month and that I would pay for it, and I'm going to.
Jordan Wilson [00:03:13]:
Right? But we're gonna take a look at all of this a little bit here live. So, let's get to it, and, hey, podcast audience, appreciate, especially if you listen twice, today. So thanks thanks for tuning in. But I'm gonna try to do my best to describe anything that we're gonna do. We're gonna jump in and test out the new full, o one model, as well. So, let's talk a little bit about what was announced today. So CEO Sam Altman and some engineers at OpenAI did a livestream. So that's essentially what's gonna be happening, over the next, 11 now weekdays.
Jordan Wilson [00:03:46]:
So, at 10 AM, Pacific time, it's, you know, sure for my Chicago people, noon. So, you know, we'll be, probably sending out our newsletter a little late every day, probably right after noon. So if you wanna see what comes out, it should be, speaking of shipping, it should be shipping out then. So, CEO Sam Altman and crew today announced live, what was new, with OpenAI in their first, 12 days of Christmas release. So here's what it is. It is ChatGPT Pro. $200 a month. Unlimited use.
Jordan Wilson [00:04:18]:
That is the big thing. Right? So no more, no more messaging caps or anything like that. And and and let's talk about the models. Let's talk a little bit about the models. So OpenAI claims that the new o one is the smartest model in the world. We'll wait to see all the 3rd party benchmarks, but what OpenAI has released, it is bonkers good. So we'll wait for, all the 3rd party benchmarks just to see how big of a gap, is it, or will it be between this new o one model, and everyone else in the world. So, we've seen a lot of, you know, of the Chinese startups in the last week, you you know, with Quan, QWQ, Alibaba, all the big, kind of large language model and AI companies in China are also releasing reasoning models as well.
Jordan Wilson [00:05:05]:
So that's another important thing to know. We this is not GPT 5. I've been on the record saying we're gonna get a GPT 4.5 before we get a GPT 5. That is not it. So the o one models are a different class. They are not a GPT class. They are a reasoning model. What that means, if you haven't heard of it, maybe you're brand new, but, there was essentially some reports more than a year ago about something called QStar.
Jordan Wilson [00:05:31]:
Right? And then that QStar project that this was a codenamed. Then it was called strawberry. And we ultimately learned, it was, the o one model and the o one series. So when OpenAI did release this, it was o one preview, which was kind of like a light version, and then o one mini, which was a smaller, faster, cheaper version. So, I should say this. This is all live now. Right? So you can go use the complete o one model. You can use the new o one pro model, which we will be talking about, as well as the new ChatGPT pro plan, that $200 a month plan, which essentially lifts all the limits off of everything.
Jordan Wilson [00:06:12]:
So it lifts the limits off of the 4 o model. It lifts the limits off of, obviously, the o one models where normally you would have a certain number of messages, a week, or on the, o one preview. The o one mini, had a higher limit per day. So a lot going on here. But let's just talk about the model. So OpenAI says that we are seeing a 50%, performance increase. Alright. So let's go over.
Jordan Wilson [00:06:39]:
So, now just reading straight off of, OpenAI's website. And, you know, we're gonna give it a little a little, poke here at the end, but we're gonna be covering, this new model in the coming weeks, I'm sure as we get more time to go super in-depth. So here's what we're seeing. OpenAI says today we're adding ChatGPT Pro, a $200 monthly plan that enables scaled access to the best of OpenAI's models and tools. This plan includes unlimited access to our smartest model, OpenAI o one, as well as o one mini, GPT 4 o, and advanced voice mode. It also includes o one pro mode. Okay? So there's not a lot of details on this new o one pro. So there's technically now 3 models of o one.
Jordan Wilson [00:07:22]:
There's o one mini. There's o one, which used to be called o one preview. So o one preview is gone. So it's o one mini, o one, and then o one pro. But you have to have that $200 a month plan, to use the o one pro. Although this morning, it was popping up for me before the announcement. Now it's gone. Alright.
Jordan Wilson [00:07:40]:
So it, what OpenAI says, it says it, so it also includes o one pro mode, a version of o one that uses more compute to think harder and provide even better answers to the hardest problems. So, yeah, not a ton of, information yet. Aside from benchmarks, I'm sure that we'll be, seeing and hearing more from OpenAI, on what actually differentiates this new o one and the o one pro. But for now, they're essentially just saying, yo. You get more electricity, more or less. Right? You get more brains, you get more electricity, and it's better. Right? And, hey, look at the benchmarks. It's better.
Jordan Wilson [00:08:16]:
And then you have to think, is it actually a different model, or is there some throttling going on, right, in the normal kind of, like, quote, unquote, o one? So, you know, not sure. I'm sure we'll get a a a lot more information, but, essentially, they're saying we devote more compute, to to o one pro, to think harder and work better, essentially. Right? So let's look at the difference here. So, if you're listening to the podcast, it's it's a graph. Right? We're looking at graphs here. And we're looking at o one preview, right, which is now, not the default o one model, but it was released in September. It was the most powerful model in the world until today. So now that is o one.
Jordan Wilson [00:08:56]:
So if you are on the normal ChatGPT paid plan, that $20 a month, ChatGPT plus or if you're on ChatGPT teams, you have access to the o one model. And they're saying the o one pro, so that's the one you have to pay $200 a month for, it performs much, much better, in challenging, machine learning benchmarks across math, science, and coding. Right? So, if you are an analyst, if you're a researcher, if you're a software developer, if you are just a power user, you might wanna look at that $200 a month plan. Right? Sounds it sounds crazy to say that out loud. Right? Because we've been absolutely spoiled. Humans in the world, we've been spoiled, for the last, you know, year and a half since there was a paid plan of ChatGPT, and then Google Bard, which became Google Gemini, and then we had Claude Anthropic, Microsoft CoPilot, all these other models. We've been spoiled. Right? I still don't think people number 1, people don't care to learn about large language models, how they work, all that.
Jordan Wilson [00:10:02]:
Right? But to have the amount of of power into, now as how much a single person can accomplish for $20 a month, it's it's nuttier than a squirrel on keto. Right? So now you have to think, is this a good thing? Right? This new tier, it's almost like the, there's an elite club now. Whereas before, I think generative AI, large language models, it it it became this place for everyone. Right? And you could almost feel this this energy and, you know, almost inclusiveness of, you know, kind of leveling the playing field. And that's something, even for me personally, I loved about generative AI and large language models early on. I said anyone can do this. Literally, anyone can scrap up. You know? Almost anyone can scrap up, you know, $20 a month and it pays for itself immediately.
Jordan Wilson [00:10:55]:
Now we have tiers. Right? We have the elite club. We have not everyone. Not everyone can afford, you know, $200, a month for this new ChatGPT Pro. And I I haven't yet seen because, again, y'all, this is just hours in, hours, you know, new or hours old. Right? I I I haven't seen a lot of details yet on, Teams accounts, enterprise accounts, what the pricing, if it's even available, what it will be. As we figure that out, we'll let you know. But, anyways, benchmarks, I mean, o one Pro is much, much higher.
Jordan Wilson [00:11:28]:
Right? So in math. Right? So we go from o one preview, which 12 hours ago was the world's most powerful model. You got a 50 on competition math, which is the, aim e, aim e 20 24 benchmark. So we got a 50. Right? So now the new o one model, 78, and o one pro, 86. Yeah. So if you are in any any any role related to business intelligence, anything with with stats, research, analytics. That what I just said right there.
Jordan Wilson [00:12:04]:
And then, I mean, we're gonna get into coding and and, science research. Right? It is such a large gap between the o one preview model that we that we all had to the o one model now in the o one pro. Also, I I I should mention, OpenAI did say the API for this should be released soon. So, at least as of right now, this isn't on the API. And so what that means, if you're a newbie, don't worry. Everyday AI, it's for everyday people. What that means, right, there's probably, and you don't even know, probably so many of the different softwares, apps, enterprise tools that you use, websites. Right? A lot of them are powered by this technology under the hood.
Jordan Wilson [00:12:47]:
Right? So so many, thousands of enterprise, companies, software, apps, etcetera, use OpenAI's technology. They use their API. So right now, if you really want a taste of this and if you wanna, you you know, figure out for yourself if this is a revolution or rip off, you're gonna have to log into ChatGPT. This isn't available for free users as of now. I'm sure they're gonna give everyone a little taste. Right? Give everyone a gateway little gateway drug here of, o one, and o one pro maybe. But, I mean, the competition, coding. So code forces, another metric here.
Jordan Wilson [00:13:28]:
O one preview, 62. 0189. Alright? I mean, we are talking about 30% plus gains overnight. Right? So a lot of these instances where smart business leaders would look and they'd be like, you know, still not there. Well, now you have to think about it. But you also have to think, is it worth that $200 a month price tag? Alright. Another thing and, again, this is all ChatGPT. We gotta do our own kind of a very unofficial internal testing.
Jordan Wilson [00:13:58]:
So another thing that they're talking about is improved reliability. So they're saying that they used, in their evaluation, which there was a whole model card. Right? So we'll share that in the newsletter, you know, today or if you're listening live, tomorrow. So we will share this, but they're also reportedly reproving the, improving the reliability. So what they said here on their page is to highlight the main strengths of o one pro mode. They use a stricter evaluation setting. A model is only considered to solve a question if it gets the answer right in 4 out of 4 attempts. 4 out of 4 reliability, not just 1.
Jordan Wilson [00:14:38]:
So in kind of even benchmarking their own internal models, you see this huge, huge improvement. So, right, like we talked about that competition math, doubled the reliability, from 01 preview to 01 pro. Right? That's the that's the only thing people don't really understand about large language models. They're generative. Right? It's rolling the dice. Especially if you're not good at, you know, prompting or if you don't know how models work, you're even rolling the dice even more so. Right? If you know the proper prompt engineering, you're, you have a little bit more control of those, Dai. But I mean, the reliability scores here, are off the charts.
Jordan Wilson [00:15:15]:
Alright. So then let's quickly let's quickly look at the pricing. Alright. And then we're gonna run, we're gonna run a little test here. Don't worry. So $200 a month. Is that worth it? Is it worth it for you? Is it worth it for me? Is it worth it for my company? Is it worth it for your department? Is this where we should be looking now? Right? Is this what we're gonna see in the future? Y'all, I'm literally on record. I said last year that eventually there would be more premium versions of large language models that would cost 1,000 of dollars a month.
Jordan Wilson [00:16:03]:
That's a good thing about having a daily podcast. You know? Got receipts receipts for days days. Right? And I still think that's coming. Right? So, you know, a lot of people are are talking about this on online, on the Twitter machine, on Reddit. Right? And they're saying, oh, OpenAI is just doing this because they're reportedly losing $5,000,000,000 a month, and they need to get profitable. Well, yes. Right? And and anyone that follows this space has known all along, that especially, OpenAI was burning through cash. Right? If you go try to use Claude Anthropic at the same rate, you can't.
Jordan Wilson [00:16:40]:
It is so limited. Right? So OpenAI has been very generous, and I think that's one of the reasons why they are literally crushing everyone else. Right? Because for $20 a month, you get access to the world's most powerful models with the best features. And it's even though there's limits, it's very hard to hit those limits. So now with this pro plan, it is unlimited access. Again, that's $200 per month for 1 user. So you do get everything else, but you get unlimited access to GPT 4 0 and o 1. Right? So I haven't personally, jumped on the $200 a month pro plan.
Jordan Wilson [00:17:19]:
I will, just been been a little bit busy today, but, you know, we will be upgrading, to that plan. Our team here at Everyday AI, we will. But one of the things that I think is great is unlimited voice mode. Alright? Because the advanced voice mode from OpenAI is so sneaky good. It's so sneaky good. Right? And if you haven't used it before, I encourage you to. You do have to have at least the $20 a month plus plan to use that. But that's something that not a lot of people talk about.
Jordan Wilson [00:17:50]:
I'm like, yo. Like, the rates are different. I I I think it's been about I think you get, like, originally it was 24 or, 20 20 to 25 minutes, when it first came out. I think they, expanded it to 60 minutes, but that's a huge upgrade there as well. Alright. So let's let's jump in. We're gonna do a super, super unofficial unofficial test here. Alright? And we're gonna see.
Jordan Wilson [00:18:20]:
We're just gonna figure out if this is worth it or not. So I'm jumping into ChatGPT01. Again, not GPT01. I'm gonna run this little prompt, and we're gonna we're gonna look. So is is this any good? I don't know. Let's let's find out. Right? So I'm saying, please give me exact and specific examples of what OpenAI's o one model can do for someone's business. Be short and succinct yet very detailed.
Jordan Wilson [00:18:52]:
Do not waste words. Yada yada yada. And I'm saying, you know, give me 5 examples. And in doing so, tell me how a subject matter expert would tackle these same task, projects, or challenges, and then how the o one model could do it. Alright? And I'm saying, be short, be direct. Let's take a look. Right? So let's take a look because everyone's everyone's always talking about use cases. Alright? And so when we think about it, the the o one model, it is a reasoning model.
Jordan Wilson [00:19:21]:
Okay? So to simplify that, in theory, you're not getting something that wasn't possible before with ChatGPT's 4.0, it would just require to get the same outputs, some of the smartest and most talented humans in the world. Right. Because, you know, if you think of like kind of quote unquote prompt engineering, it's really steering the world's intelligence at the right place with the right skill set in the right way. Right? So, technically, nothing is is different in terms of what it can achieve. But by having a model that can use logic and reasoning, right, it's it's literally if you get the smartest human in the world, give them g v t four o, they can probably get similar results. It would just maybe take them many hours and probably only a handful of people in the world could even perform at the same level. So the o one model, it does that all for you. It uses this kind of chain of thought reasoning under the hood.
Jordan Wilson [00:20:22]:
So I like to tell people like this. If you're in the kind of the the default version of ChatGPT plus, gbt 4 o. Right? You it's it's kind of like you can get to the destination, but you have to help ChatGPT build the car. You have to tell it which road to take. You have to examine. Right? Oh, is this the right route? Right. So you still have to be very involved. It's like you're you're you're building the car.
Jordan Wilson [00:20:46]:
You're you're you're paving the roads. Right? But, again, it's capabilities 5 years ago we never thought we would have with that baseline model. But with the reasoning model and now with this new updated version of it, you tell it the destination and then this reasoning or, kind of, chain of thought processing does it all for you. Right? It does a lot of those, you know, best practice prompt engineering, for you under the hood. So if if if you haven't used, the o one model before, and this is a little bit different, user interface with this this newest version. So when I asked this. Right? So I said, hey. Tell me, you know, 5 specific things, that OpenAI would be great or or sorry.
Jordan Wilson [00:21:29]:
The o one model would be great at, and then tell me how a normal, very smart human would tackle the same things. So it said thought about o one model applications for 7 seconds. So I can click that. Okay. And then on my screen here, I can go see, kind of how the model handled this. So it said highlighting so the the bullets here says highlighting capabilities, clarifying business impact, highlighting precise tasks, crafting marketing campaigns, and transforming tasks. Right? So you can literally see how it's thinking, how it's piecing it together, and it's literally going through step by step. Right? Alright.
Jordan Wilson [00:22:08]:
So let's let's let's quickly see. What are the business values, and are these things that your company could use? Alright. So, let me let me see. Let me see something that hopefully, here we go. I'm not gonna read all 5, but I'm gonna read just, number 2 here. So it says producing accurate sales forecast reports from raw data. So it says a subject matter expert, has to manually extract historical sales figures from spreadsheets, identify seasonal patterns, apply statistical models, and write a thorough analysis report, a process that could take days. Right? A one model.
Jordan Wilson [00:22:47]:
Alright. It rapidly processes upload historical sales data, identifies trends, runs predictive modeling, and outputs a polished data backed forecast report with charts in minutes or maybe barely a minute. Alright. So I'm not gonna go through I'll I'll at least read all five examples, but not the details. So number 1, it said generating highly specific marketing copy for niche audiences. Number 2, producing accurate sales forecast reports from raw data. Number 3, conducting detailed code reviews for complex software. Number 4, summarizing and analyzing customer support tickets at scale.
Jordan Wilson [00:23:28]:
And then number 5, checking marketing materials against regulatory guidelines. Alright. So, I'm gonna do another, another quick little test here. And, hey, livestream audience. Let me know if you have any questions. I I've I've I've played around with the model, but let me know if you have any questions. And and and thanks thanks everyone for tuning in twice in one day. Love to see it.
Jordan Wilson [00:23:55]:
Alright. So I'm gonna do one other thing y'all because there's another feature that I haven't even talked about, in o one that was not previous before. So not only did we get the full o one model, if you pay for the $200, you get the o one pro. But there is one other nice little feature, that's available that was not previously available. So let's actually, take a look here. Alright. So you can upload photos. Alright.
Jordan Wilson [00:24:24]:
So you unfortunately can't upload PDFs. You can't upload spreadsheets, but you can upload PNGs and JPEGs. Right? So I can do probably something, you know, kind of fun and cutesy and, you know, we do some standard, test that we do here on the Everyday AI Show, pretty often. But I'm gonna do something else. Right? Because I want this. I want to be able to upload because I think one of the best things one of the best things that you can use a large language model for right now, is analyzing data. Right? Because the availability of data has made people that are not data analysts. Well, all of a sudden, you gotta make use of it.
Jordan Wilson [00:25:02]:
Right? Just think of, like, 15 years ago. Who put together sales decks and and presentations? Right? It was probably a graphic designer. Guess who puts them together now? Literally everyone. And I think being able to tell, stories with data is everyone's job now because data is available everywhere, and we can use it and leverage it and make sense of it in seconds. Alright. So I'm gonna try a little trick. Y'all haven't done this. So if anyone is, if you're looking at my screen on the live screen, it this is chaos.
Jordan Wilson [00:25:32]:
Right? So, we we we do this on our YouTube channel a little, a little bit, but I just export a ton of data here out of our YouTube channel. It's it's video titles, views, subscribers gained on each video, impressions, click through rates, shares. Right? All these just, thousands of data points. So right now I can copy and paste this in, which I'm actually gonna try. But here's a little tip that most people aren't gonna tell you. Alright? Zoom out a ton. Actually, I'm gonna I'm gonna put this to the test. Alright.
Jordan Wilson [00:26:03]:
So I just took a screenshot. Alright. And then I gotta find, where my ChatGPT went. Alright. So now I have a screenshot with a bunch of data. Right? You can barely read it. I zoomed all the way out like it would take a human, like, a frigging telescope and, a lot of time just to be able to see this data. I'm gonna see, okay, how are its computer vision capabilities.
Jordan Wilson [00:26:34]:
This if it does this, this is an extremely difficult task. Alright? So here's what I'm gonna say. Luckily, I typed this out so you didn't have to watch me, type this live. So I'm gonna go ahead and click enter, and I'm saying based on this breakdown, please give me 20 in-depth insights that can help me grow the YouTube channel. Be pins be pinpoint specific, drawing on between the line insights, and advanced tactics. For each insight, give me actual stats or examples from this spreadsheet. Also, give me step by step directions on how to implement all twenty of these insights. Make sure they are specific to the content I've uploaded.
Jordan Wilson [00:27:12]:
Do not suggest anything general. Only focus on the results, trends, and statistics, statistical insights you can glean from the data in this screenshot. Alright. So it fought for a whopping 18 seconds. So what I will, say is sometimes when you click it and you wanna see what's going on under the hood, and I encourage you to do that, it's gonna make you better at AI. It's gonna make you better at your job. So sometimes it says no details available. Sometimes it says, oh, there we go.
Jordan Wilson [00:27:37]:
It says, occasionally, the model generates the best response without anything to summarize. Right? So it fought for 18 seconds before replying, and then it says below are 20 pinpoint insights drawn directly from the data patterns visible on the spreadsheet. And here we go. Right? So, even though you can't upload spreadsheets, which I think people are going to love once you get that capability, because before today, when working with the o one model, you couldn't upload anything. So today, you can at least upload images. Right? So that leaky leaky pipe, you wanna, you know, prove prove you're a handyman or handywoman, go ahead, upload it now, to the o one model. Right? Any any photo image, if you're trying to, you know, read something in a different language, a complex problem, if there's, you know, something in a deck or a sales presentation that you don't understand and you can, you you know, snap a photo of it, a meme that's going over your head. Right? You can use the l one model now and you can upload a photo.
Jordan Wilson [00:28:34]:
But what she what people don't know as well, just take a photo of a spreadsheet. It actually does an extremely impressive job. And I'm, I'm double checking the data here because I do wanna make sure, this is correct, and it is correct. Right? At least on on on this one, I'm gonna go through and check it later. But, by by first, kind of count, it doesn't look like we have any, any hallucinations. So it's giving me some good stuff. Right? So, let's look at some of these. I'm not gonna read them all.
Jordan Wilson [00:29:06]:
Let's see. Okay. Okay. Here's here's a good thing. Right? So, on number 4, it said turn high engagement shorts, into let me make sure I'm sharing my right screen here. There we go. So number 4, it says turn high engagement shorts like metrics into a series. And then it says data point, perplexity AI overview, and what's the best PDF plugin.
Jordan Wilson [00:29:29]:
So those are 2 different YouTube video titles, I guess, in this, screenshot from the spreadsheet. And it says have viewed versus swiped away rates of 182% and 200% perspectively. So it actually went through and created its own metric by combining 2 different metrics. That metric does not exist. So it combined 2 different metrics, and then it said it implies a hook strong enough to capture viewers beyond a single impression, possibly shorts or very short form content. Then it says so it gave it gave me a data point. It said why it matters. So it says exceptional viewed versus swiped metrics indicate you're grabbing and holding attention quickly, something crucial in short form formats.
Jordan Wilson [00:30:11]:
That's so freaking impressive. Right? It literally I didn't give it any information. I didn't say, hey. You know, this is a YouTube channel. Here's a this is a screenshot of metrics. I nope. I I I uploaded a photo. I gave it a very specific task.
Jordan Wilson [00:30:24]:
It it's literally going through there in 18 seconds. And it not only quickly analyzed everything. Right? It converted a screenshot where the fonts, like, seemingly like 2, because I'm trying I'm doing a little work around here. It literally understood it that quickly and started to create its own metrics based on the data that it observed to then create a plan. Alright. So then it's it gave me the data point, why it matters, and then how to implement how to implement. So it says, 1, identify if these videos are short form content. If yes because I didn't get all of the stats, in that screenshot.
Jordan Wilson [00:31:03]:
And then it says, if yes, produce a mini series of similar short form explainers. 2, use identical opening hooks from these top performers, such as starting with a, pressing question or a bold claim. And then it says, 3, compare the viewed versus swipe metrics over 2 to 3 new uploads with the same hook style to confirm improvement. It's freaking good. That's good. Right? Like, I've been in different forms of marketing, advertising, communications for 20 years. That is good stuff. Right? I've hired expensive consultants who would have to take 18 days to come up with, that kind of thing.
Jordan Wilson [00:31:37]:
Right? Essentially creating your own, metric, or your own way to measure something and then giving me step by step ways to to play it out. Freaking amazing. Love to see it. So, alright. I see I see a couple questions. I'm gonna try to get them, here at the very end. So what do you guys think of this? You know? And if you're listening on the podcast, we do always, put our contact information. Let me know if you do wanna see more content around this new, o one model.
Jordan Wilson [00:32:08]:
Once we get our hands, hands a little dirty with the o one pro, I'm sure we'll do a show or 2 on this, but it's pretty impressive. So now what I'm gonna do so I'm back in this spreadsheet. Right? There's a ton of data. It I I mean, this is this is, I think it's like 13,000. It's like 13,000 data points. So I don't even know how much I can upload. I didn't see, context on this yet to see how much I can just copy and paste in there. Right? So, let's see.
Jordan Wilson [00:32:40]:
Let's see if we can break a one quick. Right? Alright. So let's go in, and I'm just gonna say, I don't know. Give me 10 trends from this data. I doubt I'm gonna be able to paste in, like 13. Let's see how many of this is exactly. 14,000 cells of information. Let's see.
Jordan Wilson [00:33:08]:
Okay. Looks like it at least took it. I didn't break it. Alright. So there we go. 14,000 cells of information. I just said, tell me 10 trends. Okay.
Jordan Wilson [00:33:20]:
And there we're done. It didn't even need to tell me any details of how it thought. So that's that's that's pretty impressive y'all. And I'm I'm just double checking some things. Right? So it says, okay. We compare them all. 18, a 111 shares on what, on on the best PDF plugin. So I'm just double checking this because, I don't wanna, you you know, be giving information that's not correct.
Jordan Wilson [00:33:45]:
So I'm checking over here on my other monitor, and surely enough that video has a 111 shares. So it's it's it's not like it just made up a bunch of information. It literally literally live here, gobbled up 14,000 cells of data and gave me 10 insightful trends in seconds In seconds. Yeah. Brian is saying, I bet the consultants you hired cost more than $200 a month, though. Yeah. That's a good point, Brian. And I I think that brings me to, you you know, as we kinda wrap up here, that brings me to the question now.
Jordan Wilson [00:34:27]:
Is this right for your business? Is this a revolution or a rip off? Right? There's no one answer. I don't think. I will say I'm gonna reserve my judgment till I go through the pro. But this new model, off the bat, very impressive. Very impressive. Right? And and there's there's a I think there's a lot to still explore. I'm impressed personally. Alright.
Jordan Wilson [00:34:54]:
Gareth asking, Jordan, do you think this new large pricing hike to is to help show that they can get close to being more profitable as a company or a sign of other financial issues? I can't I can't speak to that. I don't have any inside, intel on that. Right? I I do hear from people at OpenAI that say, you know, oh, they listen to the show and all that. So so thanks to them for their support. I don't have any intel on that. My my own, take is I mean, this has to be right. So OpenAI just, raised, I I I think $7,000,000,000 a couple of months ago. I mean, this has to be a part of their plan to profitability.
Jordan Wilson [00:35:29]:
Right? It makes sense that they were, you you know, losing 1,000,000,000 of dollars reportedly, every single year. I said this all along, right? I said eventually companies, especially startups, it's different for Google and Microsoft and Amazon, right? Amazon with their very impressive Nova model, which we haven't even talked about. I think because it's so hard to use. But I I I said very early on, I said, anthropic and OpenAI are gonna have to choose. You can either burn through money, right, and give people everything, Or you have to be extremely restrictive, right, and worry about profit right away, profit over users. It seems to me, OpenAI said, hey, let's just capture the market. They were first there. They were crushing everyone because there was no competition, right, in November of 2022.
Jordan Wilson [00:36:21]:
But, you know, as they started to see competition from from Google, from Anthropic, from, Mistral, from Meta. Right? They just kept just busting everyone over the head with the best model still for so cheap. So I'm not sure. But, you know, I I I do believe that a big part of their long term strategy was capture the market first, and then offer premium products later. Nathan saying, do you think the $20 will become the old free version with dial back capabilities compared to what is currently available on the $20? No. I don't think so. Because there's other things. Right? This whole 12 days of ship ship miss.
Jordan Wilson [00:36:57]:
Right? You know, open so we're gonna be getting a lot of releases. So, I'm not gonna play out my whole line of predictions, but I've been saying all along, sometime after the US election to the Q1 of 2025, we're gonna get a GBT 4.5. I personally think that's what we're gonna see over the next, you know, 11 days of this little series. I think we're probably gonna see Sora. Sora released, at at least to maybe it's just the ChatGPT, pro. I'm not sure. I I do also think that we will eventually see these tools start to be rolled out to other modes. Here's what I mean by that.
Jordan Wilson [00:37:34]:
I just showed you all o one. You can upload photos now. Right? That's an example of a quote, unquote tool, but it doesn't have, use of all the tools, advanced voice mode, web browsing, you you know, whether you wanna call it ChatGPT search or browse with Bing. It doesn't have all of those things. The same with advanced voice mode. So I don't necessarily think the $20 a month tier, is is is going to get stripped away from. I just think you're gonna see all these newer features. Right? Like, oh, it as soon as advanced voice mode, you can have the ability to upload files and it can browse the Internet, literally, game over for, like, 80% of medium sized consulting businesses if they don't pivot hard.
Jordan Wilson [00:38:14]:
Because then I can instantly upload all my information, get real time access to literally and if you know how to speak, to a model, that's that's bananas. That's bananas. Right? So, I don't necessarily think that the $20 a month version is is gonna get worse. Alright, y'all. That was a lot. But to recap, a lot of new stuff coming from OpenAI. So they just kicked off, their kind of 12 day event, with a new version of o one, and a $200 a month new tier coming in with the ChatGPT Pro, which gives unlimited usage as well as access to that o one pro model. So I hope this was helpful.
Jordan Wilson [00:39:04]:
Thanks to our livestream audience who joined us twice in a day. Y'all are amazing. Appreciate the support. I'd also appreciate it if you tell someone about this. If this was helpful, y'all, please click that little repost button. Right? It's it's the holiday season. I don't ask you for a $1,000,000,000,000 or, $7,000,000,000 of fundraising. I just ask you to share this with your network.
Jordan Wilson [00:39:26]:
We put in a lot of time, to make sure you stay up to date with the realest news, the realest thing in artificial intelligence. Please repost, share this with your friends. And if you haven't already, please go to your everydayai.com. Sign up for the free daily newsletter every single day, Monday through Friday. We not just do this podcast and this livestream, but we also recap, the, the show for the day in the newsletter. So thank you for tuning in. Hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all.