WEBVTT
00:00:01.280 --> 00:00:02.479
Yeah, yeah, yeah.
00:00:02.720 --> 00:00:11.919
OpenAI's new GPT-5.6 soul is amazing and the best and most practical AI model in the world right now.
00:00:12.240 --> 00:00:18.160
But they actually just pulled off one of the smartest naming tricks I've seen in a while.
00:00:18.399 --> 00:00:33.039
See, they took the popular agentic coding platform codex, softened the edges, got rid of that scary word code, and just repackaged it as ChatGPT work, the super app.
00:00:33.679 --> 00:00:46.960
But everyone is rightfully paying attention to the new model itself, GPT 5.6 Soul, OpenAI's flagship release built for longer, cheaper, and more reliable execution of all types of work outputs.
00:00:47.280 --> 00:00:58.640
This sounds like one launch of two related products, but it's really a model story, a platform story, and a competitive story, all wrapped up in one.
00:00:58.719 --> 00:01:02.159
And there's some details that I think people are overlooking.
00:01:02.560 --> 00:01:07.599
So that's exactly what we're going to be tackling today on everyday AI.
00:01:07.760 --> 00:01:09.840
So here is the big picture.
00:01:10.000 --> 00:01:14.480
OpenAI just released their new 5.6 models.
00:01:14.560 --> 00:01:16.319
Yeah, GPT 5.6 models.
00:01:16.400 --> 00:01:22.079
There's multiple, and a product they're calling Chat GPT Work or the Super App.
00:01:22.719 --> 00:01:32.959
So the models themselves, well, by benchmarks and my limited use so far, look and feel like real upgrades built for long-running, hands-off professional work.
00:01:33.040 --> 00:01:39.280
So Chat GPT work, because you're going to be hearing me say work and the words chat GPT work a lot.
00:01:39.439 --> 00:01:48.640
Uh, it's well, it's just kind of repackaging of some of the tools you didn't know you had inside of Chat GPT.
00:01:48.879 --> 00:02:08.400
And in case you did miss it, it's been a very busy like 36 hours for OpenAI because not only did they come out with probably now the best model in the world, the best harness and making it more accessible for more people, but they also brought what I think could be more impactful in the long run.
00:02:08.560 --> 00:02:20.719
Uh, yesterday, no, two days ago, my gosh, time's flying by the new GPT live model, which brings that uh duplex two-way uh voice interaction.
00:02:20.960 --> 00:02:24.159
But the big thing is GPT 5.6.
00:02:24.560 --> 00:02:33.599
So on today's show, you're gonna learn why Chat GPT work on the web is mostly just pretty packaging and where the real upgrade actually lives.
00:02:33.759 --> 00:02:44.879
I'm gonna tell you five overlooked features that people aren't really focusing on, and they probably should, including the full merge super app and the quiet death of Atlas.
00:02:44.960 --> 00:02:47.280
But don't worry, it's a good thing.
00:02:47.439 --> 00:02:49.199
And I'm gonna give you my hot take.
00:02:49.439 --> 00:02:59.919
OpenAI's positioning play and why I think Anthropic should be very nervous of a model that's not GPT-5-6 soul.
00:03:00.080 --> 00:03:04.240
Yeah, a different open AI model is gonna cause anthropic headaches.
00:03:04.400 --> 00:03:05.280
Let's get into it.
00:03:05.439 --> 00:03:06.719
Welcome to everyday AI.
00:03:06.800 --> 00:03:10.080
If you're new here, my name is Jordan Wilson, and well, do this every day.
00:03:10.240 --> 00:03:19.439
Uh, it's for you, it's your daily live stream, podcast, and free daily newsletter, unedited, unscripted, helping business leaders like you and me keep up with the AI updates.
00:03:19.680 --> 00:03:22.960
I go through sleep deprivation, so you don't have to.
00:03:23.120 --> 00:03:27.280
You get to benefit from uh just the good stuff to grow your company and your career.
00:03:27.439 --> 00:03:31.840
So it starts here, but make sure you go to your everydayai.com, sign up for the free daily newsletter.
00:03:31.919 --> 00:03:36.879
We're gonna be recapping all the my random ramblings into the too long didn't listen version.
00:03:36.960 --> 00:03:42.639
So make sure you go do that and check out everything else that's happening in the world of AI in our newsletter.
00:03:42.879 --> 00:03:45.759
All right, so let's get to it.
00:03:45.919 --> 00:03:49.360
We're talking about chat GPT work, the new GPT-5.
00:03:50.000 --> 00:03:50.319
Soul.
00:03:50.479 --> 00:03:54.879
I'm gonna tell you what's new, five overlooked features, and one hot take.
00:03:55.039 --> 00:03:58.000
So GPT 56 soul.
00:03:58.159 --> 00:04:05.599
If you missed the boat or maybe you've been on a long vacation and haven't listened to the podcast in two weeks, OpenAI already announced this.
00:04:05.680 --> 00:04:07.680
This wasn't new, we knew it was coming.
00:04:07.840 --> 00:04:19.680
They actually announced it about two weeks ago, uh, but they had to work with the federal government now that we have kind of the permission slip uh you know, way of doing AI here in the US because of anthropic.
00:04:19.759 --> 00:04:23.199
Uh, but here it is, brand new model.
00:04:23.360 --> 00:04:33.920
Uh, but the difference is well, it is actually maybe a little more confusing in the long run because before we just had things like GPT 5.5, right?
00:04:34.000 --> 00:04:36.959
And then you had different levels of reasoning and all that.
00:04:37.120 --> 00:04:46.720
So now we still have the different levels of reasoning, we still have the number, the five six, but now we also have soul, which is the big model.
00:04:46.959 --> 00:04:50.959
Uh, we have Terra, which is the medium model, and then we have Luna, right?
00:04:51.040 --> 00:04:54.560
So Sun, Earth, Moon, um, which is great.
00:04:54.720 --> 00:04:56.879
And I think that that's going to carry on.
00:04:57.040 --> 00:05:00.319
Uh, so let's talk a little bit about the new model.
00:05:00.399 --> 00:05:02.720
And this is from OpenAI's release.
00:05:02.879 --> 00:05:14.480
So they said we're launching the GPT-5.6 family of models for general availability following our limited preview, our new flagship Sol alongside Terra, a balanced model for everyday work, and Luna, our most cost-efficient model.
00:05:14.720 --> 00:05:30.319
GPT-5.6 Sol sets a new standard for both intelligence and efficiency, achieving state-of-the-art results across coding, knowledge work, cybersecurity, and science while outperforming previous and competing frontier models with fewer tokens and at a lower estimated cost.
00:05:30.480 --> 00:05:38.000
The result is stronger performance per dollar, more successful work for the same spend, or comparable results at a lower total cost.
00:05:38.240 --> 00:05:41.600
We also introduce a new way to accelerate the most demanding work.
00:05:41.839 --> 00:05:49.519
Ultra is our highest capability setting, coordinating multiple agents across parallel work streams to finish complex tasks faster.
00:05:49.759 --> 00:05:53.920
Stronger computer use and design judgment make GPT-5.
00:05:56.639 --> 00:06:00.800
Helping it inspect, refine, and deliver ready-to-use results.
00:06:01.120 --> 00:06:01.519
All right.
00:06:01.600 --> 00:06:04.800
So we're going to talk a little bit more uh about GPT-5.
00:06:05.600 --> 00:06:11.519
I'm going to do a little bit of comparison, not too much on this show, uh, but obviously Fable 5, right?
00:06:11.600 --> 00:06:15.759
We have to talk about those two models together when we go over the benchmarks here in a couple of minutes.
00:06:15.920 --> 00:06:19.600
But I want to quickly tell you a little bit about the other big release.
00:06:19.759 --> 00:06:25.839
So we knew uh GPT 5.6 Soul was coming out yesterday on Thursday.
00:06:26.079 --> 00:06:31.920
What we didn't necessarily know is well, we were essentially getting the super app, right?
00:06:32.000 --> 00:06:37.360
Uh uh internally, they've kind of saying it's it's the merge, but this is the super app.
00:06:37.519 --> 00:06:46.639
So this is kind of the combination of ChatGPT, uh codex, which has been on fire the last uh five months, and the Atlas browser.
00:06:46.720 --> 00:06:49.199
So the Atlas browser is no more now.
00:06:49.279 --> 00:07:18.160
Codex is ChatGPT and Chat GPT is codex, it's actually both an amazing play, just call simply calling it chat GPT work, but also kind of confusing with how it rolled out because if you go download codecs or if you go update codecs, you know, now it says, you know, chat GPT, but you can change it, you know, you can say, hey, I want it to be the codecs icon and still call it codecs, but it's the same thing, uh, right.
00:07:18.560 --> 00:07:26.879
But uh kind of how OpenAI uh says uh put their post out about ChatGPT work.
00:07:27.279 --> 00:07:41.360
Uh they said powered by GPT5.6, ChatGPT Work brings together context from your team's tools to turn scattered notes, drafts, and ideas into finished work and keeps projects moving while you stay in control.
00:07:41.600 --> 00:07:51.040
They say ChatGPT work gathers context, plans the approach, and takes action across your tools, files, and desktop apps to create published spreadsheets, docs, and slides.
00:07:51.199 --> 00:08:00.959
Uh ChatGPT can turn context from your tools and files into published documents, presentations, and analysis to better uh that better follow your templates and preferred formats.
00:08:01.120 --> 00:08:02.480
And then they say sites.
00:08:02.639 --> 00:08:05.920
Yeah, updates on sites, and I'm gonna share about that more in a minute.
00:08:06.000 --> 00:08:20.319
Uh, sites let you turn ideas, plans, and data into interactive websites and web apps that are easy to share with your team, create dashboards, project trackers, launch calendars, prototypes, reports, and more, and keep them up to date as information changes.
00:08:20.480 --> 00:08:33.519
Uh, so then talking about the unified plugins, they say with more than 1400 plugins available, ChatGPT can pull context from the tools and workflows you already use to help move projects forward.
00:08:33.600 --> 00:08:41.120
Uh, they say create one-time recurring tasks, monitor updates, and check progress from your phone while you're away from your desk.
00:08:41.200 --> 00:08:45.440
And I'm going to be talking about those ChatGPT remote updates because I think they're really good.
00:08:45.600 --> 00:08:50.639
Uh, then the new uh side-by-side uh in ChatGPT, which I like.
00:08:50.799 --> 00:09:02.480
Uh, it's the new built-in browser experience in the ChatGPT desktop app, makes it easier to work across your tools, files, and accounts with support for multiple tabs and richer agentic workflows.
00:09:02.559 --> 00:09:10.480
And I'm gonna share at the end some of my uh tips on the best way to do that because that's for me the biggest unlock.
00:09:10.639 --> 00:09:12.799
A very small piece of this is the biggest unlock.
00:09:12.879 --> 00:09:15.679
Uh, and then review the approach before work begins.
00:09:15.759 --> 00:09:21.200
In plan mode, ChatGPT gathers context, asks questions, and creates a step-by-step plan.
00:09:21.279 --> 00:09:26.159
You can suggest changes or approve the play, uh, the plan to start the work.
00:09:26.720 --> 00:09:27.200
All right.
00:09:27.440 --> 00:09:30.559
So let me translate all of that marketing speak.
00:09:30.639 --> 00:09:32.159
Uh, it's just codex.
00:09:33.279 --> 00:09:34.000
That's it.
00:09:34.240 --> 00:09:35.600
There's really nothing new.
00:09:35.759 --> 00:09:44.720
Um, so I don't know if if most people remember this, but uh OpenAI added a feature about two weeks ago, or sorry, about two months ago.
00:09:44.799 --> 00:09:50.399
Uh, essentially, it allowed you to choose how you wanted codecs to display.
00:09:50.559 --> 00:09:56.879
So you could either have a more technical display or something they just called for everyday work.
00:09:56.960 --> 00:10:01.440
Uh, so you could just choose how much technical detail codex showed when it was working.
00:10:01.600 --> 00:10:08.879
So, did you want to see all the tokens spinning and all the Pythons calling and all the you know tools spinning off the shelf, right?
00:10:09.039 --> 00:10:14.399
If so, you would choose the coding approach, which is technically now codecs.
00:10:14.480 --> 00:10:22.879
Uh, if you didn't, if you just wanted a more minimalistic and you didn't really care about what was going on under the hood, well, that's now just called Chat GPT work.
00:10:23.039 --> 00:10:28.559
I spent so much time and I'm like, this can't be true, both on the web, because they did the same thing on the web.
00:10:28.720 --> 00:10:36.639
On the web, now you have chat GPT, uh, and there's a toggle for chat and a toggle for work.
00:10:37.120 --> 00:10:43.919
And you know, on the web at least, chat GPT work looks a little different.
00:10:44.080 --> 00:10:49.840
There's technically no real new functionality aside from being powered by the new model.
00:10:50.000 --> 00:10:54.000
The plugins are unified, but there's no new anything, right?
00:10:54.080 --> 00:10:57.279
So you look at it and you're like, oh my gosh, this is gonna be amazing.
00:10:57.440 --> 00:10:58.480
What's new here?
00:10:58.720 --> 00:11:02.240
All this is, which it's smart, don't get me wrong, right?
00:11:02.320 --> 00:11:12.720
I just was like, okay, I just you know read the blog post six times, watched every single video that open AI put out, uh, you know, played with every feature, and I'm like, oh, it's literally the same.
00:11:12.879 --> 00:11:14.399
But I don't think that's a bad thing.
00:11:14.559 --> 00:11:16.639
I think it's smart what OpenAI did here.
00:11:16.720 --> 00:11:26.000
Because if I'm being honest, even power users of Chad GPT, I could go in, I feel show them things, and then people would be like, wait, what?
00:11:26.080 --> 00:11:27.519
Like, where'd you find that setting?
00:11:27.600 --> 00:11:27.840
Right.
00:11:27.919 --> 00:11:34.159
There's literally be some things that were some of the best parts of Chad GPT that you couldn't even find, right?
00:11:34.240 --> 00:11:40.000
You had to go into like a setting like four or five rows in, and it's like, oh my gosh, you can do that in Chat GPT.
00:11:40.080 --> 00:11:41.279
It's like, yeah, you can.
00:11:41.759 --> 00:11:54.159
Um, you know, like even like scheduled tasks, they used to uh they took that away for a couple of months, but you could still access it via a URL or by clicking like four or five links deep inside of a setting panel.
00:11:54.320 --> 00:11:58.799
So you're gonna probably be hearing a lot of chatter about the new ChatGPT work.
00:11:58.960 --> 00:11:59.759
I do think it's smart.
00:11:59.840 --> 00:12:10.080
So on the web, it's just some more uh, you know, it does look like more codex on the web, but you don't really have the codex features per se.
00:12:10.559 --> 00:12:16.799
Um but on the desktop, it's just codecs.
00:12:17.120 --> 00:12:32.559
Chat GPT work is codecs, and you can choose to call it chat GPT or call it codecs, but you will have a uh kind of a switcher at the top, and you can switch between uh ChatGPT work and ChatGPT codex, but nothing really changes.
00:12:32.879 --> 00:12:35.919
But like I said, I think it's smart.
00:12:36.000 --> 00:12:47.840
And here's the play it still baffled me how many smart people I've talked to over the last five or six months about codex, right?
00:12:48.000 --> 00:12:56.399
Because on one hand, yes, it is an extremely powerful agentic platform that you can automate your life, you can code, you can build things, uh, right.
00:12:56.480 --> 00:12:58.960
And that's intimidating for people, right?
00:12:59.120 --> 00:13:00.879
But and I've been saying this on the show.
00:13:01.200 --> 00:13:10.559
You know, I've been, you know, captain codex since February, and it absolutely crushes every competitor, and it's not even close.
00:13:10.720 --> 00:13:14.320
It's way better than Claude Desktop, Claude Code, Claude Cowork.
00:13:14.559 --> 00:13:15.600
It's better than Cursor.
00:13:15.679 --> 00:13:17.759
I think Cursor is probably in second place.
00:13:17.840 --> 00:13:20.080
Uh, it's better than anti-gravity Gemini.
00:13:20.320 --> 00:13:24.000
It is the best desktop AI app, and it is not even close.
00:13:24.159 --> 00:13:27.039
But the problem is, is the word codex, right?
00:13:27.200 --> 00:13:29.039
People said, Oh, I'm not a coder, right?
00:13:29.120 --> 00:13:34.559
Yeah, I've had so many recent conversations, people who are even somewhat technical, but they're not coders.
00:13:34.639 --> 00:13:36.080
And they're like, Oh, I can't use codecs.
00:13:36.159 --> 00:13:37.679
And I'm like, you absolutely have to.
00:13:37.759 --> 00:13:40.480
So, from that aspect, it's a very smart play.
00:13:40.559 --> 00:13:47.360
They literally just slapped the word work on, didn't really change anything, just shows fewer details.
00:13:47.600 --> 00:13:48.159
That's it.
00:13:48.320 --> 00:13:48.720
All right.
00:13:48.799 --> 00:13:52.399
So don't get confused when you're like, what's this new Chad GPT work?
00:13:52.559 --> 00:13:53.759
It's really nothing new.
00:13:53.919 --> 00:13:54.240
All right.
00:13:54.320 --> 00:14:08.000
So yeah, here on my screen, I have the Chat GPT work on the web, just looks kind of like Codex Lite, but it doesn't really do anything new or different that you couldn't previously do on the web inside of Chad GPT.
00:14:08.159 --> 00:14:12.879
So yeah, the web version just bundles agent work, deep research projects, connected apps.
00:14:13.039 --> 00:14:18.480
So it just brought more some of the more uh technically advanced features, the ones that I used all the time.
00:14:18.559 --> 00:14:22.480
And it kind of gave it its own uh dedicated panel in the work.
00:14:22.639 --> 00:14:25.440
You can still do all those things in the normal chat panel.
00:14:25.600 --> 00:14:37.679
It just, you know, makes the chat panel maybe a little bit easier um and less intimidating with you know fewer interfaces, and then the work panel just makes it easier and then it suggests different things.
00:14:38.320 --> 00:14:38.960
All right.
00:14:39.919 --> 00:14:41.600
And then, well, there's one new thing.
00:14:41.679 --> 00:14:47.600
There's a nice little toggle, right, inside the chat chat GPT uh desktop app.
00:14:47.919 --> 00:14:50.320
Uh yeah, there's nothing really new there.
00:14:50.399 --> 00:14:52.240
Uh, but Atlas is built in.
00:14:52.320 --> 00:14:53.440
Sorry, I take that back.
00:14:53.600 --> 00:14:57.440
Atlas is now built into the browser, and it is really good.
00:14:57.600 --> 00:15:03.360
It is improved, and I'm gonna give you here in a couple of minutes uh when I go over my five overlooked features.
00:15:03.519 --> 00:15:08.639
One of them is going to hopefully, I think, change the way a lot of people work.
00:15:09.120 --> 00:15:21.840
All right, so more on GPT-5.6, you know, got to get into the benchmarks, the comparisons, all that, but I need a quick water break, and we have to quickly hear uh word from our partners.
00:15:22.320 --> 00:15:27.279
Here's what most AI tools still can't do work outside their old little box.
00:15:27.440 --> 00:15:29.759
The all new Slack bot just changed that.
00:15:29.919 --> 00:15:36.799
It's your AI teammate inside Slack, and now it can read, write, and act across the other apps your team already uses.
00:15:36.960 --> 00:15:42.240
No more tool switching, no more re-explaining yourself every time you open a new tab.
00:15:42.480 --> 00:15:48.879
One ops team at engine says the summary feature alone saves them 15 to 20 minutes of use.
00:15:49.039 --> 00:15:52.320
See what Slackbot can do at Slack.com.
00:15:53.759 --> 00:15:56.960
All right, so a little bit more on the new models.
00:15:57.120 --> 00:16:00.159
So GPT-5.6, here's the basics.
00:16:00.320 --> 00:16:03.120
There's three models Sol, Terra, Luna.
00:16:03.279 --> 00:16:09.840
Soul is the flagship, it is aimed at the heavy users, coding, knowledge work, cybersecurity, and science.
00:16:10.000 --> 00:16:22.799
If you are someone that's using AI for three, five, six, twelve hours a day, you're probably gonna want to use Seoul because you will be able to take advantage of those full capabilities.
00:16:22.879 --> 00:16:25.279
We always talk about the capability gap, uh, right?
00:16:25.360 --> 00:16:29.120
If if you're doing lighter work, maybe you can use Terra, right?
00:16:29.200 --> 00:16:32.240
Uh, that is the more everyday work at a lower cost.
00:16:32.320 --> 00:16:35.440
And then Luna is the fastest, cheapest tier.
00:16:35.519 --> 00:16:38.399
So, yeah, that's our our new three naming tier.
00:16:38.559 --> 00:16:44.399
And then on top of that, you have um Soul Ultra as well.
00:16:45.279 --> 00:16:47.759
So let's look at that.
00:16:48.000 --> 00:16:52.639
So with Soul, so we're gonna get into the uh the benchmarks here.
00:16:52.799 --> 00:16:55.840
Uh, it did extremely well.
00:16:56.080 --> 00:17:05.920
So, even on a newer benchmark, which I really like, uh, if you've heard of humanities last exam, it's essentially all these questions that are not in training data.
00:17:06.079 --> 00:17:16.319
So, there's a similar one called Agent's Last Exam, and it's been one of the more popular in trending benchmarks of 2026, and it just crushed the competition.
00:17:16.480 --> 00:17:21.839
So, you know, and even if you're sitting there and you're like, I don't need agents, I'm not a coder.
00:17:21.920 --> 00:17:23.599
Yes, you do, and yes, you will.
00:17:23.759 --> 00:17:25.039
You just don't know it yet.
00:17:25.200 --> 00:17:27.759
Uh, and and and I'm being dead serious, right?
00:17:28.079 --> 00:17:36.880
The future of work, if you are a knowledge worker sitting in front of a computer, you are an agent orchestrator.
00:17:37.039 --> 00:17:42.559
It's it's how I've been doing my work since February, since Codex came out.
00:17:42.720 --> 00:17:50.079
Um, you know, constantly providing a lot of feedback uh on the front end, you know, keeping an eye on your agents, seeing what they return you.
00:17:50.240 --> 00:18:03.519
That's your work, you know, you polish it a little bit, but you're able to get four, five, six, ten times uh done that of the type of work that you would normally do, and uh, you know, unlocking a whole new skill set that you never thought you could do.
00:18:03.680 --> 00:18:07.839
Um, so but the company's core pitch here is just performance per dollar.
00:18:08.000 --> 00:18:09.839
And we're gonna get into that in some of the charts.
00:18:10.000 --> 00:18:15.759
And then there is technically a tier-ish on top of Seoul, and that's Seoul Ultra.
00:18:15.839 --> 00:18:25.200
And that's just when they run uh essentially four uh kind of sub-agents in parallel, all of the highest setting, and you kind of get the best of the best.
00:18:25.680 --> 00:18:33.119
So Ultra does kind of coordinate those four agents in parallel by default to finish those demanding tasks faster.
00:18:33.279 --> 00:18:40.559
Uh, we've gone over pricing before, but Sol is five dollars input and thirty dollars per output for million tokens.
00:18:40.640 --> 00:18:43.039
So that's just if you're paying on the API side, right?
00:18:43.119 --> 00:18:50.400
So if you're in uh if you have a Chat GPT subscription, FYI, that now means you have a Chat GPT work or a codex subscription.
00:18:50.559 --> 00:18:52.640
It's the same thing, it shares those limits.
00:18:52.799 --> 00:18:58.799
So obviously you're not, you know, having to worry about that, but the pricing is extremely important.
00:18:59.200 --> 00:19:05.680
Uh, and then obviously the cheaper per task uh can still drain uh plans quickly.
00:19:05.920 --> 00:19:13.759
So, you know, if you are using uh the ultra salt ultra, that's that's gonna eat eat things away.
00:19:14.640 --> 00:19:18.559
All right, let's talk a little bit about some benchmarks.
00:19:18.640 --> 00:19:21.519
We're not gonna go too deep into it, but it's important to talk about.
00:19:21.680 --> 00:19:29.200
So I think probably the best um single source for benchmarks uh is artificial analysis.
00:19:29.279 --> 00:19:35.359
So it is uh conglomerate, it just brings in, you know, different benchmarks all at once.
00:19:35.519 --> 00:19:37.839
So it averages all these different benchmarks.
00:19:37.920 --> 00:19:40.400
Uh so it's one of the best sources to look at.
00:19:40.559 --> 00:19:48.000
So uh GPT 5.6 Seoul Max, uh, pretty much on par uh with Claude Fable 5.
00:19:48.079 --> 00:19:52.559
It's a one-point difference, uh, which is pretty minuscule, right?
00:19:52.720 --> 00:19:57.920
I'd say you can start power users can start noticing at about a three-point difference.
00:19:58.079 --> 00:20:03.680
Um, average users maybe like a six-point difference, and they're not gonna tell the difference.
00:20:03.839 --> 00:20:14.079
Uh, but you know, if if you are a heavy user, you're not gonna be able to tell the difference between a 59 and a 60 on artificial analysis intelligence index.
00:20:14.160 --> 00:20:18.000
That means they are essentially one A and one AAB.
00:20:18.079 --> 00:20:21.039
It's not one A and one B, it's one A and one AA B.
00:20:21.119 --> 00:20:25.039
But that's just on that, the artificial analysis intelligence benchmark.
00:20:25.200 --> 00:20:26.640
Here's where it gets interesting.
00:20:26.799 --> 00:20:29.839
On uh, they have essentially they have three main benchmarks.
00:20:29.920 --> 00:20:35.119
I usually just talk about the uh intelligence index because that brings a little bit of everything.
00:20:35.359 --> 00:20:51.359
But in this case, when we're talking about work, knowledge work becoming more and more agentic, I think that's actually when the other two artificial uh analysis um benchmarks become more important, and that is their artificial analysis coding index, uh, which GPT-5.
00:20:55.039 --> 00:21:01.359
So it beats it on coding and on the agentic index, uh 5.5 also beats it.
00:21:01.519 --> 00:21:10.079
So essentially, I would argue by most standards, that would mean that uh GPT-5.6 soul is better.
00:21:10.240 --> 00:21:14.240
I mean, we'll see when all the benchmarks come out in the LM arena.
00:21:14.319 --> 00:21:15.759
So, what do humans prefer?
00:21:15.839 --> 00:21:16.480
Uh, right.
00:21:16.720 --> 00:21:20.880
And you know, OpenAI has said they're working on front-end design.
00:21:21.039 --> 00:21:24.160
I know so much, which I've never understood, right?
00:21:24.480 --> 00:21:32.000
Because so much of like vibes that are out there for models, it's always like, okay, I created this video game and like look at how good it looks.
00:21:32.079 --> 00:21:39.039
It's like, okay, well, maybe like what 0.001% of the world uses models for that, right?
00:21:39.200 --> 00:21:40.720
I use models for knowledge work.
00:21:40.799 --> 00:21:51.519
I use models for personalizing research, carrying my context over from different apps, uh, you know, building, um, you know, building decks, building little sites that I can share, right?
00:21:51.680 --> 00:21:55.359
Uh, you know, helping me grow my business, right?
00:21:55.440 --> 00:21:59.759
Uh I'm I'm not creating 3G 3JS worlds, uh, you know, I.
00:22:00.079 --> 00:22:04.160
I do a little bit of design things maybe for some of the apps that I do build.
00:22:04.319 --> 00:22:08.400
But you know, I think it's important to look at these three benchmarks in tandem.
00:22:08.640 --> 00:22:15.680
But the one that is the most telling, and this is we're gonna get to this in a little bit more, is the price.
00:22:15.839 --> 00:22:21.200
So for artificial analysis, this is essentially the cost per intelligence, right?
00:22:21.440 --> 00:22:24.799
Because these models can all get the tasks done.
00:22:25.119 --> 00:22:27.839
But well, how much does it cost?
00:22:28.079 --> 00:22:33.200
And right now, GPT-5.6, I will argue, is a slightly better model.
00:22:33.279 --> 00:22:38.559
If you look at all these, uh, you know, the three different artificial analysis indexes that are the most important.
00:22:38.720 --> 00:22:51.519
Uh, you know, they're one point behind on one winning the other two, but they're doing it at yeah, a dollar per task, a dollar and four cents versus uh Fable 5,$2.75.
00:22:52.240 --> 00:22:52.880
Yeah.
00:22:53.359 --> 00:22:54.079
Entropic.
00:22:54.319 --> 00:22:54.559
All right.
00:22:54.720 --> 00:22:59.599
So uh we went over Chad GPT work, we went over codex.
00:22:59.920 --> 00:23:04.720
Now I'm gonna give you the five overlooked features and my one hot take.
00:23:04.799 --> 00:23:07.359
All right, overlook feature number one.
00:23:07.519 --> 00:23:12.160
Even though OpenAI didn't say it out loud, this is the super app, right?
00:23:12.240 --> 00:23:26.400
Uh, I think there are a lot of super app um, you know, rah-rah back in like March, April, May, um, and at the time OpenAI said, well, codex is the super app, uh, right.
00:23:26.559 --> 00:23:30.400
They said, you know, they said we're building it in disguise, but this is it.
00:23:30.720 --> 00:23:38.000
But is bringing together chat work, codex, browser, files, and actions.
00:23:38.079 --> 00:23:42.640
And if you know absolutely nothing about like, well, codex, right?
00:23:42.720 --> 00:23:43.920
Or why does it matter?
00:23:44.000 --> 00:23:51.519
You know, why would why do I need a desktop, you know, something running on my desktop versus something, you know, that I can just do in the cloud, right?
00:23:51.680 --> 00:23:56.319
If ChatGPT has this new ChatGPT work on the web, why would I ever use Codex?
00:23:56.799 --> 00:24:01.039
Well, the main reason is it can access your local system.
00:24:01.200 --> 00:24:06.720
So it can use your entire computer, it can use your browser, uh, right?
00:24:06.880 --> 00:24:11.279
It can access, read and write all of your local files.
00:24:11.440 --> 00:24:13.759
So those are you know three big advantages.
00:24:13.920 --> 00:24:17.599
And, you know, there's a little bit more that you can do with scheduling and things like that.
00:24:17.680 --> 00:24:24.240
If you have a machine like I do that just stays on, my poor Mac Studio, I don't think, has taken a nap since I got it.
00:24:24.319 --> 00:24:24.720
Uh, right.
00:24:24.799 --> 00:24:29.440
So I have a lot of things that are scheduled, you know, loops, goals, heartbeats, all these things.
00:24:29.599 --> 00:24:38.480
There's a little bit more uh technical um capabilities that you can do with the desktop app, but it is now all one.
00:24:38.799 --> 00:24:44.799
There's some syncing issues, FYI, and I'll probably reach out to OpenAI about it.
00:24:44.880 --> 00:24:51.680
I I don't know if they're fully um aware because, as an example, some things change names, right?
00:24:51.839 --> 00:24:55.759
So if you were using codecs before, there was a thing called chats.
00:24:55.920 --> 00:24:57.599
Now there's no codex chats.
00:24:57.759 --> 00:25:04.640
Codex chats are now tasks, so they could be called codex tasks or they could be called chat GPT work tasks, anyways.
00:25:04.799 --> 00:25:20.640
But you can there's a new chats panel, and that brings in your chats from ChatGPT, but right now at least only chats from the chat tab inside chat GPT are syncing and not the chats from the work tab inside chat GPT.
00:25:20.720 --> 00:25:22.240
Yeah, confusing, right?
00:25:22.400 --> 00:25:29.519
So there are some things that you know, the the the super app and bringing all this together in one cohesive ball.
00:25:29.680 --> 00:25:33.200
I think that there's a couple of things that the team is probably gonna tighten up.
00:25:33.359 --> 00:25:37.680
Um, but I mean, first iteration, it is really, really good.
00:25:37.839 --> 00:25:46.400
All right, the other thing, small footnote, sites, all right, chat GPT sites rolling out to all paid plans.
00:25:46.559 --> 00:25:46.880
All right.
00:25:46.960 --> 00:25:55.440
I think there's some might be some uh geography restrictions at first, but uh rolling out to all paid plans, and this is important.
00:25:55.599 --> 00:26:02.000
Also, there's public publishing, live URLs, and it's on Chad GPT on the web, not just codecs.
00:26:02.079 --> 00:26:05.839
So we are starting to see, you know, a little bit more feature parity.
00:26:05.920 --> 00:26:12.799
Uh, you know, certain like a good example is uh Chad GPT sites used to only be on codecs.
00:26:12.880 --> 00:26:17.599
So, well, now that's also uh on Chad GPT on the web as well.
00:26:17.759 --> 00:26:25.839
So here's why this is important because before sites was number one only available on Codecs, so now it's available on the web as well.
00:26:26.000 --> 00:26:30.240
Uh, but it's also um available for all paid plans.
00:26:30.400 --> 00:26:32.559
Before it was just for team plans.
00:26:32.720 --> 00:26:44.480
If if if I'm being honest, like it was one of the reasons that I still was holding on uh to my team plan, uh, even though I'm mainly the only person that uses it, was for ChatGPT sites.
00:26:44.559 --> 00:26:49.839
I really like it, but on my$200 a month pro plan, it's like I wanted to be able to use sites now.
00:26:50.000 --> 00:26:50.319
I can't.
00:26:50.400 --> 00:26:52.960
So if you have a paid plan, ChatGPT sites is great.
00:26:53.039 --> 00:26:55.759
It's like a a lovable light, right?
00:26:55.920 --> 00:26:59.839
So all you know, uh I I did a full episode on it a couple of weeks ago.
00:26:59.920 --> 00:27:00.799
Think of it like this.
00:27:00.960 --> 00:27:04.640
Um, you know, think of how many versions of decks that you have, right?
00:27:04.720 --> 00:27:09.359
There's 10 different versions on 10 different people's machines, v1, v3, v4, final.
00:27:09.599 --> 00:27:11.440
Oh, did we have edits on this one, right?
00:27:11.599 --> 00:27:18.559
You probably waste uh hundreds of hours a year just dealing with uh who has that file, right?
00:27:18.720 --> 00:27:23.200
ChatGPT sites can just get rid of that because it's just one source of truth.
00:27:23.279 --> 00:27:26.400
And anytime one person updates it, it's updated everywhere.
00:27:26.480 --> 00:27:38.240
So you can build full-on you know, databases, dashboards, uh, you know, little apps that your team can use or just decks and websites to keep track of anything type of work.
00:27:38.400 --> 00:27:42.559
Uh, so it's nice that it's rolling out to all plans and available on the web.
00:27:42.720 --> 00:27:46.799
Next, Chad GPT Mobile got a huge update.
00:27:46.960 --> 00:27:51.359
Uh, it now you can manage plugins from Chad GPT app.
00:27:51.599 --> 00:27:54.799
You can see and manage all of your schedule tasks.
00:27:54.960 --> 00:27:59.440
So uh that's great, but also uh they changed the name.
00:27:59.599 --> 00:28:08.640
So you used to open up the Chad GPT app and it would say codex, and then you could essentially control your uh any physical machine that was running codecs.
00:28:08.799 --> 00:28:12.079
So now it's just called remote, which I kind of like, right?
00:28:12.160 --> 00:28:16.079
So they're not slapping the name codecs everywhere and confusing people.
00:28:16.160 --> 00:28:19.519
And if you see remote, you're like, cool, even if you're a non-technical person.
00:28:19.680 --> 00:28:28.400
And now all of a sudden you're like, wait, I can use the chat GPT app, which you know, uh OpenAI has like a billion weekly active users, and everyone knows the chat GPT app.
00:28:28.480 --> 00:28:36.319
And then they see this thing called remote, and they're like, Oh, I can, you know, control my desktop if I have the ChatGPT work app.
00:28:36.480 --> 00:28:37.279
Wow, great.
00:28:37.440 --> 00:28:55.359
So again, not really new functionality there, it's just cleaned up and it does work a little bit better, but actually, no, new, there is new functionality because you have more control over plugins and scheduled tasks on the actual codep, uh, sorry, uh, on the actual remote section of the ChatGPT app.
00:28:55.839 --> 00:28:59.200
All right, number four, uh, yeah, Atlas is technically gone.
00:28:59.359 --> 00:29:05.759
So OpenAI's browser experiment, it is no more, but it's for the best because now this is gonna sound small.
00:29:05.920 --> 00:29:16.720
All right, but two big updates to uh the browser, uh formerly known as Atlas, that is now just baked into the super app, right?
00:29:16.960 --> 00:29:31.359
So Codex had always had a browser, but it was a singular browser, and it didn't have all the normal browser capabilities, like you couldn't open multiple tabs and it didn't keep your saved passwords.
00:29:31.599 --> 00:29:38.079
I actually spent like 1.8 billion tokens building what I called, I forgot what I called it.
00:29:38.160 --> 00:29:55.920
I think I just called it like Codex Vault, but I essentially built an app that would save all of my passwords so I could run, you know, certain things agentically that required you to log in because you know the codex browser, you know, couldn't uh log in to things without saved passwords.
00:29:56.079 --> 00:30:05.920
Um, so that is available now, and you can import from any of your Chrome profiles because the browser is based on Chromium like Atlas was.
00:30:06.079 --> 00:30:10.720
So uh it supports cookies to keep you logged in and it'll support passwords.
00:30:10.960 --> 00:30:26.079
So I was playing around with it a little bit um because it was a little finicky to actually get computer used to click login, because some websites, depending on how they're set up, you know, um the actual browser control can log you in.
00:30:26.160 --> 00:30:31.359
So it'll pull up your saved, you know, uh websites, uh, your saved credentials, right?
00:30:31.440 --> 00:30:38.400
Just like in Chrome, if you you know sign into something and you've saved your password in Chrome, it'll be there and you just have to click log in.
00:30:38.559 --> 00:30:41.200
Um, so I had to toy around with it a little bit.
00:30:41.359 --> 00:30:49.279
I might just uh if I can refine it, I just might save it as a skill and you know put it in our newsletter or on our website because it is a little finicky now.
00:30:49.359 --> 00:30:50.640
I'm sure they'll get it fixed.
00:30:50.799 --> 00:30:52.799
But this is the big unlock here, y'all.
00:30:52.880 --> 00:30:53.759
And here's why.
00:30:53.920 --> 00:30:59.519
Yes, you know, there's there's connectors and plugins and mcps and all these things, right?
00:30:59.680 --> 00:31:09.920
But I think for knowledge work, so much of what we do is going to all of these different websites that don't have MCP and they don't have direct connectors, right?
00:31:10.000 --> 00:31:16.319
So, in my example, you know, the our podcast is hosted on BuzzSprout, which is an older host.
00:31:16.480 --> 00:31:18.319
They don't have any AI anything.
00:31:18.480 --> 00:31:24.720
I spend so much time, you know, going from this program, uh, you know, downloading this, uploading this file here.
00:31:24.799 --> 00:31:30.400
I clean the audio, then I bring it in, you know, get the timestamps, and it's like five different programs.
00:31:30.640 --> 00:31:51.359
Well, now I can just fully automate that completely because codex can open multiple tabs uh at once and it can save the passwords because you know it does save them, it had previously saved them for a certain amount of time, but every once in a while it just kicks you out like any normal uh you know SaaS program would.
00:31:51.440 --> 00:31:59.759
So now with the cookies and the saved passwords, huge in terms of the new knowledge work that you can do on a schedule.
00:31:59.920 --> 00:32:01.680
Think of all those monotonous tasks.
00:32:01.759 --> 00:32:04.480
You have to go log in, click this thing 12 times.
00:32:04.640 --> 00:32:06.079
You're like, why am I doing this?
00:32:06.240 --> 00:32:06.480
Right?
00:32:06.640 --> 00:32:15.920
Why, you know, even if you're a high-paid executive, you're like, I I, you know, I just every single day I have 45 minutes of things that you know a monkey could do, right?
00:32:16.079 --> 00:32:18.880
I just go here, click this, wait for it to load, three minutes.
00:32:19.039 --> 00:32:20.160
Oops, got distracted.
00:32:20.319 --> 00:32:22.480
Click this, export, download, right?
00:32:22.640 --> 00:32:24.000
You don't have to do any of that anymore.
00:32:24.079 --> 00:32:24.640
It's amazing.
00:32:24.720 --> 00:32:28.720
All right, and then last but not least, computer use is much faster.
00:32:28.880 --> 00:32:33.039
It is faster, it is better, and it has this real cool new thing.
00:32:33.119 --> 00:32:34.799
It's this picture in picture.
00:32:34.960 --> 00:32:52.640
Uh, so when computer use is happening, um you can see, you know, if you've ever seen picture in picture on your TV, uh, it's just right there in Codex, or if you decide to call your codex chat GPT work, uh, so you can see what's happening um right there in codex.
00:32:52.720 --> 00:33:00.640
Oh, another thing that I didn't mention, which I should have mentioned on number four with Atlas, um, related to this and related to this super app.
00:33:00.720 --> 00:33:04.000
And I don't even think OpenAI announced this.
00:33:04.160 --> 00:33:17.759
Um, and I didn't see anyone else talk about it, but in the Chrome extension, because you can also from Codex not just control your computer, not just control the built-in browser, but you can also control a Chrome browser, which is a little different.
00:33:17.920 --> 00:33:18.640
Uh, all right.
00:33:18.720 --> 00:33:31.680
Uh, but they actually have a Chrome extension, um, and there's an update in the Chrome extension where you can have a side panel in Chrome and you can see what is being worked on on your desktop, which is really cool.
00:33:31.839 --> 00:33:40.240
So, you know, if I have a a window open, which is actually unlocks a lot because normally on my setup, right, I have two screens.
00:33:40.400 --> 00:33:46.720
One screen is just full, you know, full codex, uh, you know, and then I'll have, you know, probably a full clawed desktop.
00:33:46.880 --> 00:33:50.160
Um, and then I have to swap between them because I'm like, oh, okay.
00:33:50.319 --> 00:33:56.319
And I know, you know, codex has the little pets and they say, all right, your thing's ready, but I'm still like having to always keep an eye on that.
00:33:56.480 --> 00:34:04.160
So having a sidebar in Chrome and I can see what Codex is working on and track it live.
00:34:04.319 --> 00:34:06.079
That's been like a small little thing.
00:34:06.160 --> 00:34:08.639
And I'm like, wait, why didn't they talk about this?
00:34:08.800 --> 00:34:14.239
This is huge, especially if you're like me and you're giving you know agents long-running tasks.
00:34:14.400 --> 00:34:19.199
You know, at any point, I might have, you know, five to 20 different threads running.
00:34:19.280 --> 00:34:22.239
Um, and it's like, I don't want to always have to check in on them.
00:34:22.320 --> 00:34:26.000
And sometimes the little codex pet for me is a little distracting.
00:34:26.239 --> 00:34:31.039
And I'm always in a Chrome browser, so I'm like, oh, okay, I can just have that going and work with it right there.
00:34:31.199 --> 00:34:32.400
Actually, amazing.
00:34:32.559 --> 00:34:32.960
All right.
00:34:33.199 --> 00:34:37.920
So that is the five um overlooked features.
00:34:38.000 --> 00:34:40.960
And let me end with this hot take.
00:34:42.800 --> 00:34:48.559
Uh for those of you that stuck around, Anthropic's in trouble.
00:34:49.840 --> 00:34:51.679
They are in trouble.
00:34:51.920 --> 00:34:57.199
If they do not immediately pivot, it's not gonna look good.
00:34:57.519 --> 00:34:57.920
All right.
00:34:58.079 --> 00:35:04.800
And here's the thing Yeah, Anthropic is scared, and they should be, right?
00:35:04.880 --> 00:35:07.840
And this is great because it's going to drive competition.
00:35:08.079 --> 00:35:12.719
Anthropic did something that I don't think they've ever done right after this announcement.
00:35:12.800 --> 00:35:18.639
And this and this is how I know Anthropic is in huge, huge trouble.
00:35:19.119 --> 00:35:27.760
Um, they reset weekly limits, which I don't think they've ever done, except one time when they made a big mistake.
00:35:27.840 --> 00:35:29.360
There's this big miscommunication.
00:35:29.519 --> 00:35:33.440
I think they might have reset five-hour limits, not weekly limits.
00:35:33.599 --> 00:35:37.039
That's something if you use codecs, you like we've been spoiled, right?
00:35:37.119 --> 00:35:41.280
Tebow over there uh at OpenAI always resetting your weekly limits.
00:35:41.599 --> 00:35:46.800
Infrappit did this, they never do it, and they did it right when OpenAI announced something.
00:35:46.880 --> 00:35:53.920
And that to me is like they know they're scared, and they know I'm I'm I'm not gonna make projections, man.
00:35:54.079 --> 00:35:55.039
I'll make a projection.
00:35:55.119 --> 00:36:02.320
I I'd say it's very, very safe to say that yesterday they lost probably eight figures in revenue, I would guess.
00:36:02.480 --> 00:36:11.360
Um, and I will continue to say that until they make a change, they're gonna be losing eight figures of revenue every single day, maybe more.
00:36:11.519 --> 00:36:15.199
It's they're gonna be losing a lot of money if they do not pivot immediately.
00:36:15.360 --> 00:36:18.079
So I even heard some people say, and I'm not gonna say this, right?
00:36:18.320 --> 00:36:20.559
People are like, oh, Fable Five is dead.
00:36:20.719 --> 00:36:26.239
It's not dead, uh, it's still, you know, a top-tier model, and there's still gonna be some people out there paying for it.
00:36:26.559 --> 00:36:31.360
But I would absolutely not pay a dime for Fable V.
00:36:31.599 --> 00:36:38.639
And Anthropic, they said, oh, we're gonna, you know, extend Fable V in uh subscriptions through July 12th.
00:36:38.800 --> 00:36:43.519
Uh well, if they don't extend it past that, they're gonna lose subscribers.
00:36:43.599 --> 00:36:47.199
But the majority uh of Anthropic's business comes from two main things.
00:36:47.360 --> 00:36:50.480
It comes from coding via the API, um, right?
00:36:50.639 --> 00:36:54.880
And up until 5.6 soul, um, they had a sizable lead, right?
00:36:55.039 --> 00:36:58.159
Fable five um had a sizable lead over five.
00:36:59.199 --> 00:37:02.000
From GPT and you know, Google, we'll see what happens.
00:37:02.159 --> 00:37:09.039
Uh, but here's the thing I don't think anthropic is worried about GPT 5.6 soul.
00:37:09.519 --> 00:37:15.360
They are worried about GPT 5.6 Terra.
00:37:15.760 --> 00:37:16.000
Okay.
00:37:16.400 --> 00:37:17.199
Here's why.
00:37:17.440 --> 00:37:21.119
So Anthropic makes money from coding.
00:37:21.599 --> 00:37:23.119
Yes, they have co-work.
00:37:23.280 --> 00:37:24.639
Yes, that's popular.
00:37:25.199 --> 00:37:29.519
They make their money from coding, but they sell tokens at an extreme premium.
00:37:29.840 --> 00:37:30.480
All right.
00:37:32.000 --> 00:37:38.480
Looking at the artificial analysis coding index, and that takes into account a lot of the popular coding benchmarks.
00:37:38.639 --> 00:37:40.719
Um, you know, soul, GPT-5.
00:37:43.119 --> 00:37:44.800
But guess who's in second place?
00:37:45.199 --> 00:37:48.960
GPT-5.6 Terra, the medium model.
00:37:49.360 --> 00:37:52.239
It is better than the big bad fable.
00:37:52.559 --> 00:38:01.039
The one that we had uh, you know, have a government anthropic fight, the one that, you know, anthropic had been, you know, saying this thing's too dangerous.
00:38:01.119 --> 00:38:14.239
They've been they were hyping it up for for three, three months, and now we have this uh you know permission slip AI uh in the US because of anthropic overhyping uh Fable and Mythos.
00:38:14.559 --> 00:38:14.880
Okay.
00:38:15.519 --> 00:38:20.559
The medium model is eating Fable Five's lunch.
00:38:20.719 --> 00:38:21.760
And why are they eating it?
00:38:21.840 --> 00:38:33.280
Not just because it's increment incrementally better at coding, it is more than five times cheaper, according to artificial analysis.
00:38:34.639 --> 00:38:40.960
Five times cheaper to do the same tasks versus Fable Five.
00:38:41.440 --> 00:38:41.760
Yeah.
00:38:41.920 --> 00:38:50.000
So what I said on the on the Twitter machine, I said anthropic has like a week or two until enterprises start catching up.
00:38:50.079 --> 00:38:57.119
Uh, because that's how long it takes for even the decision makers that are spending multiple millions of dollars sometimes a month on these products.
00:38:57.199 --> 00:39:07.840
And yes, there's vendor lock-in, but any company that's up for, you know, if they signed a one-year uh deal with anthropic, there's always companies rolling.
00:39:08.320 --> 00:39:09.679
They're gonna churn.
00:39:09.760 --> 00:39:16.719
There is zero, zero reason for any company right now to be paying for anthropic.
00:39:17.039 --> 00:39:17.519
Zero.
00:39:18.239 --> 00:39:20.719
Because of what OpenAI just did.
00:39:20.880 --> 00:39:24.239
They undercut on price, they overdelivered on performance.
00:39:24.480 --> 00:39:30.960
And if Fable, if anthropic doesn't change Fable Five pricing, it is DOA.
00:39:31.199 --> 00:39:31.920
It's done.
00:39:32.079 --> 00:39:40.400
So they've got to either quickly release Fable 5.1 um so they can, you know, knock down the pricing on Fable 5.
00:39:40.559 --> 00:39:46.159
Or I do think that they will be bleeding five figures potentially daily, right?
00:39:46.239 --> 00:39:51.039
So not just from people going over on contract renewals, but just on the API side.
00:39:51.119 --> 00:40:15.679
Um, it's it's gonna be interesting to look at these third-party sites um like open router that track uh you know third-party token usage, because I I would assume you're gonna see the open AI share go up exponentially, and you're gonna see people that were using at least for Fable um go down, Opus go down, because why would you use Opus now when Terra is cheaper?
00:40:15.920 --> 00:40:21.199
So a big portion of Anthropic's revenue just got cut off at the knees.
00:40:21.280 --> 00:40:24.159
And if nothing else, this is just fun for us, right?
00:40:24.239 --> 00:40:30.480
We get to watch, but that means we're gonna get number one, a better model from Anthropic sooner rather than later.
00:40:30.639 --> 00:40:38.320
They have to, uh, and they're gonna have to cut their prices and or uh extend and keep Fable Five in subscriptions.
00:40:38.400 --> 00:40:40.960
There's really no other option, right?
00:40:41.039 --> 00:40:45.440
Uh, because they can't just keep bleeding money because they're gonna be bleeding money.
00:40:45.519 --> 00:40:45.840
All right.
00:40:46.000 --> 00:40:46.880
So that's it.
00:40:47.039 --> 00:40:54.159
Now you know Chat GPT work, you know GPT 56 Soul, what's new, the five overlooked features, and my one hot take.
00:40:54.239 --> 00:40:55.280
I hope this was helpful.
00:40:55.440 --> 00:40:59.599
If so, please go to your everydayai.com, sign up for the free daily newsletter.
00:40:59.679 --> 00:41:00.239
Uh, thanks.
00:41:00.320 --> 00:41:05.039
We'll see you uh, well, not tomorrow, but Monday and every day for more everyday AI.
00:41:05.119 --> 00:41:06.159
Thanks, y'all.
00:00:01.280 --> 00:00:02.479
Yeah, yeah, yeah.
00:00:02.720 --> 00:00:11.919
OpenAI's new GPT-5.6 soul is amazing and the best and most practical AI model in the world right now.
00:00:12.240 --> 00:00:18.160
But they actually just pulled off one of the smartest naming tricks I've seen in a while.
00:00:18.399 --> 00:00:33.039
See, they took the popular agentic coding platform codex, softened the edges, got rid of that scary word code, and just repackaged it as ChatGPT work, the super app.
00:00:33.679 --> 00:00:46.960
But everyone is rightfully paying attention to the new model itself, GPT 5.6 Soul, OpenAI's flagship release built for longer, cheaper, and more reliable execution of all types of work outputs.
00:00:47.280 --> 00:00:58.640
This sounds like one launch of two related products, but it's really a model story, a platform story, and a competitive story, all wrapped up in one.
00:00:58.719 --> 00:01:02.159
And there's some details that I think people are overlooking.
00:01:02.560 --> 00:01:07.599
So that's exactly what we're going to be tackling today on everyday AI.
00:01:07.760 --> 00:01:09.840
So here is the big picture.
00:01:10.000 --> 00:01:14.480
OpenAI just released their new 5.6 models.
00:01:14.560 --> 00:01:16.319
Yeah, GPT 5.6 models.
00:01:16.400 --> 00:01:22.079
There's multiple, and a product they're calling Chat GPT Work or the Super App.
00:01:22.719 --> 00:01:32.959
So the models themselves, well, by benchmarks and my limited use so far, look and feel like real upgrades built for long-running, hands-off professional work.
00:01:33.040 --> 00:01:39.280
So Chat GPT work, because you're going to be hearing me say work and the words chat GPT work a lot.
00:01:39.439 --> 00:01:48.640
Uh, it's well, it's just kind of repackaging of some of the tools you didn't know you had inside of Chat GPT.
00:01:48.879 --> 00:02:08.400
And in case you did miss it, it's been a very busy like 36 hours for OpenAI because not only did they come out with probably now the best model in the world, the best harness and making it more accessible for more people, but they also brought what I think could be more impactful in the long run.
00:02:08.560 --> 00:02:20.719
Uh, yesterday, no, two days ago, my gosh, time's flying by the new GPT live model, which brings that uh duplex two-way uh voice interaction.
00:02:20.960 --> 00:02:24.159
But the big thing is GPT 5.6.
00:02:24.560 --> 00:02:33.599
So on today's show, you're gonna learn why Chat GPT work on the web is mostly just pretty packaging and where the real upgrade actually lives.
00:02:33.759 --> 00:02:44.879
I'm gonna tell you five overlooked features that people aren't really focusing on, and they probably should, including the full merge super app and the quiet death of Atlas.
00:02:44.960 --> 00:02:47.280
But don't worry, it's a good thing.
00:02:47.439 --> 00:02:49.199
And I'm gonna give you my hot take.
00:02:49.439 --> 00:02:59.919
OpenAI's positioning play and why I think Anthropic should be very nervous of a model that's not GPT-5-6 soul.
00:03:00.080 --> 00:03:04.240
Yeah, a different open AI model is gonna cause anthropic headaches.
00:03:04.400 --> 00:03:05.280
Let's get into it.
00:03:05.439 --> 00:03:06.719
Welcome to everyday AI.
00:03:06.800 --> 00:03:10.080
If you're new here, my name is Jordan Wilson, and well, do this every day.
00:03:10.240 --> 00:03:19.439
Uh, it's for you, it's your daily live stream, podcast, and free daily newsletter, unedited, unscripted, helping business leaders like you and me keep up with the AI updates.
00:03:19.680 --> 00:03:22.960
I go through sleep deprivation, so you don't have to.
00:03:23.120 --> 00:03:27.280
You get to benefit from uh just the good stuff to grow your company and your career.
00:03:27.439 --> 00:03:31.840
So it starts here, but make sure you go to your everydayai.com, sign up for the free daily newsletter.
00:03:31.919 --> 00:03:36.879
We're gonna be recapping all the my random ramblings into the too long didn't listen version.
00:03:36.960 --> 00:03:42.639
So make sure you go do that and check out everything else that's happening in the world of AI in our newsletter.
00:03:42.879 --> 00:03:45.759
All right, so let's get to it.
00:03:45.919 --> 00:03:49.360
We're talking about chat GPT work, the new GPT-5.
00:03:50.000 --> 00:03:50.319
Soul.
00:03:50.479 --> 00:03:54.879
I'm gonna tell you what's new, five overlooked features, and one hot take.
00:03:55.039 --> 00:03:58.000
So GPT 56 soul.
00:03:58.159 --> 00:04:05.599
If you missed the boat or maybe you've been on a long vacation and haven't listened to the podcast in two weeks, OpenAI already announced this.
00:04:05.680 --> 00:04:07.680
This wasn't new, we knew it was coming.
00:04:07.840 --> 00:04:19.680
They actually announced it about two weeks ago, uh, but they had to work with the federal government now that we have kind of the permission slip uh you know, way of doing AI here in the US because of anthropic.
00:04:19.759 --> 00:04:23.199
Uh, but here it is, brand new model.
00:04:23.360 --> 00:04:33.920
Uh, but the difference is well, it is actually maybe a little more confusing in the long run because before we just had things like GPT 5.5, right?
00:04:34.000 --> 00:04:36.959
And then you had different levels of reasoning and all that.
00:04:37.120 --> 00:04:46.720
So now we still have the different levels of reasoning, we still have the number, the five six, but now we also have soul, which is the big model.
00:04:46.959 --> 00:04:50.959
Uh, we have Terra, which is the medium model, and then we have Luna, right?
00:04:51.040 --> 00:04:54.560
So Sun, Earth, Moon, um, which is great.
00:04:54.720 --> 00:04:56.879
And I think that that's going to carry on.
00:04:57.040 --> 00:05:00.319
Uh, so let's talk a little bit about the new model.
00:05:00.399 --> 00:05:02.720
And this is from OpenAI's release.
00:05:02.879 --> 00:05:14.480
So they said we're launching the GPT-5.6 family of models for general availability following our limited preview, our new flagship Sol alongside Terra, a balanced model for everyday work, and Luna, our most cost-efficient model.
00:05:14.720 --> 00:05:30.319
GPT-5.6 Sol sets a new standard for both intelligence and efficiency, achieving state-of-the-art results across coding, knowledge work, cybersecurity, and science while outperforming previous and competing frontier models with fewer tokens and at a lower estimated cost.
00:05:30.480 --> 00:05:38.000
The result is stronger performance per dollar, more successful work for the same spend, or comparable results at a lower total cost.
00:05:38.240 --> 00:05:41.600
We also introduce a new way to accelerate the most demanding work.
00:05:41.839 --> 00:05:49.519
Ultra is our highest capability setting, coordinating multiple agents across parallel work streams to finish complex tasks faster.
00:05:49.759 --> 00:05:53.920
Stronger computer use and design judgment make GPT-5.
00:05:56.639 --> 00:06:00.800
Helping it inspect, refine, and deliver ready-to-use results.
00:06:01.120 --> 00:06:01.519
All right.
00:06:01.600 --> 00:06:04.800
So we're going to talk a little bit more uh about GPT-5.
00:06:05.600 --> 00:06:11.519
I'm going to do a little bit of comparison, not too much on this show, uh, but obviously Fable 5, right?
00:06:11.600 --> 00:06:15.759
We have to talk about those two models together when we go over the benchmarks here in a couple of minutes.
00:06:15.920 --> 00:06:19.600
But I want to quickly tell you a little bit about the other big release.
00:06:19.759 --> 00:06:25.839
So we knew uh GPT 5.6 Soul was coming out yesterday on Thursday.
00:06:26.079 --> 00:06:31.920
What we didn't necessarily know is well, we were essentially getting the super app, right?
00:06:32.000 --> 00:06:37.360
Uh uh internally, they've kind of saying it's it's the merge, but this is the super app.
00:06:37.519 --> 00:06:46.639
So this is kind of the combination of ChatGPT, uh codex, which has been on fire the last uh five months, and the Atlas browser.
00:06:46.720 --> 00:06:49.199
So the Atlas browser is no more now.
00:06:49.279 --> 00:07:18.160
Codex is ChatGPT and Chat GPT is codex, it's actually both an amazing play, just call simply calling it chat GPT work, but also kind of confusing with how it rolled out because if you go download codecs or if you go update codecs, you know, now it says, you know, chat GPT, but you can change it, you know, you can say, hey, I want it to be the codecs icon and still call it codecs, but it's the same thing, uh, right.
00:07:18.560 --> 00:07:26.879
But uh kind of how OpenAI uh says uh put their post out about ChatGPT work.
00:07:27.279 --> 00:07:41.360
Uh they said powered by GPT5.6, ChatGPT Work brings together context from your team's tools to turn scattered notes, drafts, and ideas into finished work and keeps projects moving while you stay in control.
00:07:41.600 --> 00:07:51.040
They say ChatGPT work gathers context, plans the approach, and takes action across your tools, files, and desktop apps to create published spreadsheets, docs, and slides.
00:07:51.199 --> 00:08:00.959
Uh ChatGPT can turn context from your tools and files into published documents, presentations, and analysis to better uh that better follow your templates and preferred formats.
00:08:01.120 --> 00:08:02.480
And then they say sites.
00:08:02.639 --> 00:08:05.920
Yeah, updates on sites, and I'm gonna share about that more in a minute.
00:08:06.000 --> 00:08:20.319
Uh, sites let you turn ideas, plans, and data into interactive websites and web apps that are easy to share with your team, create dashboards, project trackers, launch calendars, prototypes, reports, and more, and keep them up to date as information changes.
00:08:20.480 --> 00:08:33.519
Uh, so then talking about the unified plugins, they say with more than 1400 plugins available, ChatGPT can pull context from the tools and workflows you already use to help move projects forward.
00:08:33.600 --> 00:08:41.120
Uh, they say create one-time recurring tasks, monitor updates, and check progress from your phone while you're away from your desk.
00:08:41.200 --> 00:08:45.440
And I'm going to be talking about those ChatGPT remote updates because I think they're really good.
00:08:45.600 --> 00:08:50.639
Uh, then the new uh side-by-side uh in ChatGPT, which I like.
00:08:50.799 --> 00:09:02.480
Uh, it's the new built-in browser experience in the ChatGPT desktop app, makes it easier to work across your tools, files, and accounts with support for multiple tabs and richer agentic workflows.
00:09:02.559 --> 00:09:10.480
And I'm gonna share at the end some of my uh tips on the best way to do that because that's for me the biggest unlock.
00:09:10.639 --> 00:09:12.799
A very small piece of this is the biggest unlock.
00:09:12.879 --> 00:09:15.679
Uh, and then review the approach before work begins.
00:09:15.759 --> 00:09:21.200
In plan mode, ChatGPT gathers context, asks questions, and creates a step-by-step plan.
00:09:21.279 --> 00:09:26.159
You can suggest changes or approve the play, uh, the plan to start the work.
00:09:26.720 --> 00:09:27.200
All right.
00:09:27.440 --> 00:09:30.559
So let me translate all of that marketing speak.
00:09:30.639 --> 00:09:32.159
Uh, it's just codex.
00:09:33.279 --> 00:09:34.000
That's it.
00:09:34.240 --> 00:09:35.600
There's really nothing new.
00:09:35.759 --> 00:09:44.720
Um, so I don't know if if most people remember this, but uh OpenAI added a feature about two weeks ago, or sorry, about two months ago.
00:09:44.799 --> 00:09:50.399
Uh, essentially, it allowed you to choose how you wanted codecs to display.
00:09:50.559 --> 00:09:56.879
So you could either have a more technical display or something they just called for everyday work.
00:09:56.960 --> 00:10:01.440
Uh, so you could just choose how much technical detail codex showed when it was working.
00:10:01.600 --> 00:10:08.879
So, did you want to see all the tokens spinning and all the Pythons calling and all the you know tools spinning off the shelf, right?
00:10:09.039 --> 00:10:14.399
If so, you would choose the coding approach, which is technically now codecs.
00:10:14.480 --> 00:10:22.879
Uh, if you didn't, if you just wanted a more minimalistic and you didn't really care about what was going on under the hood, well, that's now just called Chat GPT work.
00:10:23.039 --> 00:10:28.559
I spent so much time and I'm like, this can't be true, both on the web, because they did the same thing on the web.
00:10:28.720 --> 00:10:36.639
On the web, now you have chat GPT, uh, and there's a toggle for chat and a toggle for work.
00:10:37.120 --> 00:10:43.919
And you know, on the web at least, chat GPT work looks a little different.
00:10:44.080 --> 00:10:49.840
There's technically no real new functionality aside from being powered by the new model.
00:10:50.000 --> 00:10:54.000
The plugins are unified, but there's no new anything, right?
00:10:54.080 --> 00:10:57.279
So you look at it and you're like, oh my gosh, this is gonna be amazing.
00:10:57.440 --> 00:10:58.480
What's new here?
00:10:58.720 --> 00:11:02.240
All this is, which it's smart, don't get me wrong, right?
00:11:02.320 --> 00:11:12.720
I just was like, okay, I just you know read the blog post six times, watched every single video that open AI put out, uh, you know, played with every feature, and I'm like, oh, it's literally the same.
00:11:12.879 --> 00:11:14.399
But I don't think that's a bad thing.
00:11:14.559 --> 00:11:16.639
I think it's smart what OpenAI did here.
00:11:16.720 --> 00:11:26.000
Because if I'm being honest, even power users of Chad GPT, I could go in, I feel show them things, and then people would be like, wait, what?
00:11:26.080 --> 00:11:27.519
Like, where'd you find that setting?
00:11:27.600 --> 00:11:27.840
Right.
00:11:27.919 --> 00:11:34.159
There's literally be some things that were some of the best parts of Chad GPT that you couldn't even find, right?
00:11:34.240 --> 00:11:40.000
You had to go into like a setting like four or five rows in, and it's like, oh my gosh, you can do that in Chat GPT.
00:11:40.080 --> 00:11:41.279
It's like, yeah, you can.
00:11:41.759 --> 00:11:54.159
Um, you know, like even like scheduled tasks, they used to uh they took that away for a couple of months, but you could still access it via a URL or by clicking like four or five links deep inside of a setting panel.
00:11:54.320 --> 00:11:58.799
So you're gonna probably be hearing a lot of chatter about the new ChatGPT work.
00:11:58.960 --> 00:11:59.759
I do think it's smart.
00:11:59.840 --> 00:12:10.080
So on the web, it's just some more uh, you know, it does look like more codex on the web, but you don't really have the codex features per se.
00:12:10.559 --> 00:12:16.799
Um but on the desktop, it's just codecs.
00:12:17.120 --> 00:12:32.559
Chat GPT work is codecs, and you can choose to call it chat GPT or call it codecs, but you will have a uh kind of a switcher at the top, and you can switch between uh ChatGPT work and ChatGPT codex, but nothing really changes.
00:12:32.879 --> 00:12:35.919
But like I said, I think it's smart.
00:12:36.000 --> 00:12:47.840
And here's the play it still baffled me how many smart people I've talked to over the last five or six months about codex, right?
00:12:48.000 --> 00:12:56.399
Because on one hand, yes, it is an extremely powerful agentic platform that you can automate your life, you can code, you can build things, uh, right.
00:12:56.480 --> 00:12:58.960
And that's intimidating for people, right?
00:12:59.120 --> 00:13:00.879
But and I've been saying this on the show.
00:13:01.200 --> 00:13:10.559
You know, I've been, you know, captain codex since February, and it absolutely crushes every competitor, and it's not even close.
00:13:10.720 --> 00:13:14.320
It's way better than Claude Desktop, Claude Code, Claude Cowork.
00:13:14.559 --> 00:13:15.600
It's better than Cursor.
00:13:15.679 --> 00:13:17.759
I think Cursor is probably in second place.
00:13:17.840 --> 00:13:20.080
Uh, it's better than anti-gravity Gemini.
00:13:20.320 --> 00:13:24.000
It is the best desktop AI app, and it is not even close.
00:13:24.159 --> 00:13:27.039
But the problem is, is the word codex, right?
00:13:27.200 --> 00:13:29.039
People said, Oh, I'm not a coder, right?
00:13:29.120 --> 00:13:34.559
Yeah, I've had so many recent conversations, people who are even somewhat technical, but they're not coders.
00:13:34.639 --> 00:13:36.080
And they're like, Oh, I can't use codecs.
00:13:36.159 --> 00:13:37.679
And I'm like, you absolutely have to.
00:13:37.759 --> 00:13:40.480
So, from that aspect, it's a very smart play.
00:13:40.559 --> 00:13:47.360
They literally just slapped the word work on, didn't really change anything, just shows fewer details.
00:13:47.600 --> 00:13:48.159
That's it.
00:13:48.320 --> 00:13:48.720
All right.
00:13:48.799 --> 00:13:52.399
So don't get confused when you're like, what's this new Chad GPT work?
00:13:52.559 --> 00:13:53.759
It's really nothing new.
00:13:53.919 --> 00:13:54.240
All right.
00:13:54.320 --> 00:14:08.000
So yeah, here on my screen, I have the Chat GPT work on the web, just looks kind of like Codex Lite, but it doesn't really do anything new or different that you couldn't previously do on the web inside of Chad GPT.
00:14:08.159 --> 00:14:12.879
So yeah, the web version just bundles agent work, deep research projects, connected apps.
00:14:13.039 --> 00:14:18.480
So it just brought more some of the more uh technically advanced features, the ones that I used all the time.
00:14:18.559 --> 00:14:22.480
And it kind of gave it its own uh dedicated panel in the work.
00:14:22.639 --> 00:14:25.440
You can still do all those things in the normal chat panel.
00:14:25.600 --> 00:14:37.679
It just, you know, makes the chat panel maybe a little bit easier um and less intimidating with you know fewer interfaces, and then the work panel just makes it easier and then it suggests different things.
00:14:38.320 --> 00:14:38.960
All right.
00:14:39.919 --> 00:14:41.600
And then, well, there's one new thing.
00:14:41.679 --> 00:14:47.600
There's a nice little toggle, right, inside the chat chat GPT uh desktop app.
00:14:47.919 --> 00:14:50.320
Uh yeah, there's nothing really new there.
00:14:50.399 --> 00:14:52.240
Uh, but Atlas is built in.
00:14:52.320 --> 00:14:53.440
Sorry, I take that back.
00:14:53.600 --> 00:14:57.440
Atlas is now built into the browser, and it is really good.
00:14:57.600 --> 00:15:03.360
It is improved, and I'm gonna give you here in a couple of minutes uh when I go over my five overlooked features.
00:15:03.519 --> 00:15:08.639
One of them is going to hopefully, I think, change the way a lot of people work.
00:15:09.120 --> 00:15:21.840
All right, so more on GPT-5.6, you know, got to get into the benchmarks, the comparisons, all that, but I need a quick water break, and we have to quickly hear uh word from our partners.
00:15:22.320 --> 00:15:27.279
Here's what most AI tools still can't do work outside their old little box.
00:15:27.440 --> 00:15:29.759
The all new Slack bot just changed that.
00:15:29.919 --> 00:15:36.799
It's your AI teammate inside Slack, and now it can read, write, and act across the other apps your team already uses.
00:15:36.960 --> 00:15:42.240
No more tool switching, no more re-explaining yourself every time you open a new tab.
00:15:42.480 --> 00:15:48.879
One ops team at engine says the summary feature alone saves them 15 to 20 minutes of use.
00:15:49.039 --> 00:15:52.320
See what Slackbot can do at Slack.com.
00:15:53.759 --> 00:15:56.960
All right, so a little bit more on the new models.
00:15:57.120 --> 00:16:00.159
So GPT-5.6, here's the basics.
00:16:00.320 --> 00:16:03.120
There's three models Sol, Terra, Luna.
00:16:03.279 --> 00:16:09.840
Soul is the flagship, it is aimed at the heavy users, coding, knowledge work, cybersecurity, and science.
00:16:10.000 --> 00:16:22.799
If you are someone that's using AI for three, five, six, twelve hours a day, you're probably gonna want to use Seoul because you will be able to take advantage of those full capabilities.
00:16:22.879 --> 00:16:25.279
We always talk about the capability gap, uh, right?
00:16:25.360 --> 00:16:29.120
If if you're doing lighter work, maybe you can use Terra, right?
00:16:29.200 --> 00:16:32.240
Uh, that is the more everyday work at a lower cost.
00:16:32.320 --> 00:16:35.440
And then Luna is the fastest, cheapest tier.
00:16:35.519 --> 00:16:38.399
So, yeah, that's our our new three naming tier.
00:16:38.559 --> 00:16:44.399
And then on top of that, you have um Soul Ultra as well.
00:16:45.279 --> 00:16:47.759
So let's look at that.
00:16:48.000 --> 00:16:52.639
So with Soul, so we're gonna get into the uh the benchmarks here.
00:16:52.799 --> 00:16:55.840
Uh, it did extremely well.
00:16:56.080 --> 00:17:05.920
So, even on a newer benchmark, which I really like, uh, if you've heard of humanities last exam, it's essentially all these questions that are not in training data.
00:17:06.079 --> 00:17:16.319
So, there's a similar one called Agent's Last Exam, and it's been one of the more popular in trending benchmarks of 2026, and it just crushed the competition.
00:17:16.480 --> 00:17:21.839
So, you know, and even if you're sitting there and you're like, I don't need agents, I'm not a coder.
00:17:21.920 --> 00:17:23.599
Yes, you do, and yes, you will.
00:17:23.759 --> 00:17:25.039
You just don't know it yet.
00:17:25.200 --> 00:17:27.759
Uh, and and and I'm being dead serious, right?
00:17:28.079 --> 00:17:36.880
The future of work, if you are a knowledge worker sitting in front of a computer, you are an agent orchestrator.
00:17:37.039 --> 00:17:42.559
It's it's how I've been doing my work since February, since Codex came out.
00:17:42.720 --> 00:17:50.079
Um, you know, constantly providing a lot of feedback uh on the front end, you know, keeping an eye on your agents, seeing what they return you.
00:17:50.240 --> 00:18:03.519
That's your work, you know, you polish it a little bit, but you're able to get four, five, six, ten times uh done that of the type of work that you would normally do, and uh, you know, unlocking a whole new skill set that you never thought you could do.
00:18:03.680 --> 00:18:07.839
Um, so but the company's core pitch here is just performance per dollar.
00:18:08.000 --> 00:18:09.839
And we're gonna get into that in some of the charts.
00:18:10.000 --> 00:18:15.759
And then there is technically a tier-ish on top of Seoul, and that's Seoul Ultra.
00:18:15.839 --> 00:18:25.200
And that's just when they run uh essentially four uh kind of sub-agents in parallel, all of the highest setting, and you kind of get the best of the best.
00:18:25.680 --> 00:18:33.119
So Ultra does kind of coordinate those four agents in parallel by default to finish those demanding tasks faster.
00:18:33.279 --> 00:18:40.559
Uh, we've gone over pricing before, but Sol is five dollars input and thirty dollars per output for million tokens.
00:18:40.640 --> 00:18:43.039
So that's just if you're paying on the API side, right?
00:18:43.119 --> 00:18:50.400
So if you're in uh if you have a Chat GPT subscription, FYI, that now means you have a Chat GPT work or a codex subscription.
00:18:50.559 --> 00:18:52.640
It's the same thing, it shares those limits.
00:18:52.799 --> 00:18:58.799
So obviously you're not, you know, having to worry about that, but the pricing is extremely important.
00:18:59.200 --> 00:19:05.680
Uh, and then obviously the cheaper per task uh can still drain uh plans quickly.
00:19:05.920 --> 00:19:13.759
So, you know, if you are using uh the ultra salt ultra, that's that's gonna eat eat things away.
00:19:14.640 --> 00:19:18.559
All right, let's talk a little bit about some benchmarks.
00:19:18.640 --> 00:19:21.519
We're not gonna go too deep into it, but it's important to talk about.
00:19:21.680 --> 00:19:29.200
So I think probably the best um single source for benchmarks uh is artificial analysis.
00:19:29.279 --> 00:19:35.359
So it is uh conglomerate, it just brings in, you know, different benchmarks all at once.
00:19:35.519 --> 00:19:37.839
So it averages all these different benchmarks.
00:19:37.920 --> 00:19:40.400
Uh so it's one of the best sources to look at.
00:19:40.559 --> 00:19:48.000
So uh GPT 5.6 Seoul Max, uh, pretty much on par uh with Claude Fable 5.
00:19:48.079 --> 00:19:52.559
It's a one-point difference, uh, which is pretty minuscule, right?
00:19:52.720 --> 00:19:57.920
I'd say you can start power users can start noticing at about a three-point difference.
00:19:58.079 --> 00:20:03.680
Um, average users maybe like a six-point difference, and they're not gonna tell the difference.
00:20:03.839 --> 00:20:14.079
Uh, but you know, if if you are a heavy user, you're not gonna be able to tell the difference between a 59 and a 60 on artificial analysis intelligence index.
00:20:14.160 --> 00:20:18.000
That means they are essentially one A and one AAB.
00:20:18.079 --> 00:20:21.039
It's not one A and one B, it's one A and one AA B.
00:20:21.119 --> 00:20:25.039
But that's just on that, the artificial analysis intelligence benchmark.
00:20:25.200 --> 00:20:26.640
Here's where it gets interesting.
00:20:26.799 --> 00:20:29.839
On uh, they have essentially they have three main benchmarks.
00:20:29.920 --> 00:20:35.119
I usually just talk about the uh intelligence index because that brings a little bit of everything.
00:20:35.359 --> 00:20:51.359
But in this case, when we're talking about work, knowledge work becoming more and more agentic, I think that's actually when the other two artificial uh analysis um benchmarks become more important, and that is their artificial analysis coding index, uh, which GPT-5.
00:20:55.039 --> 00:21:01.359
So it beats it on coding and on the agentic index, uh 5.5 also beats it.
00:21:01.519 --> 00:21:10.079
So essentially, I would argue by most standards, that would mean that uh GPT-5.6 soul is better.
00:21:10.240 --> 00:21:14.240
I mean, we'll see when all the benchmarks come out in the LM arena.
00:21:14.319 --> 00:21:15.759
So, what do humans prefer?
00:21:15.839 --> 00:21:16.480
Uh, right.
00:21:16.720 --> 00:21:20.880
And you know, OpenAI has said they're working on front-end design.
00:21:21.039 --> 00:21:24.160
I know so much, which I've never understood, right?
00:21:24.480 --> 00:21:32.000
Because so much of like vibes that are out there for models, it's always like, okay, I created this video game and like look at how good it looks.
00:21:32.079 --> 00:21:39.039
It's like, okay, well, maybe like what 0.001% of the world uses models for that, right?
00:21:39.200 --> 00:21:40.720
I use models for knowledge work.
00:21:40.799 --> 00:21:51.519
I use models for personalizing research, carrying my context over from different apps, uh, you know, building, um, you know, building decks, building little sites that I can share, right?
00:21:51.680 --> 00:21:55.359
Uh, you know, helping me grow my business, right?
00:21:55.440 --> 00:21:59.759
Uh I'm I'm not creating 3G 3JS worlds, uh, you know, I.
00:22:00.079 --> 00:22:04.160
I do a little bit of design things maybe for some of the apps that I do build.
00:22:04.319 --> 00:22:08.400
But you know, I think it's important to look at these three benchmarks in tandem.
00:22:08.640 --> 00:22:15.680
But the one that is the most telling, and this is we're gonna get to this in a little bit more, is the price.
00:22:15.839 --> 00:22:21.200
So for artificial analysis, this is essentially the cost per intelligence, right?
00:22:21.440 --> 00:22:24.799
Because these models can all get the tasks done.
00:22:25.119 --> 00:22:27.839
But well, how much does it cost?
00:22:28.079 --> 00:22:33.200
And right now, GPT-5.6, I will argue, is a slightly better model.
00:22:33.279 --> 00:22:38.559
If you look at all these, uh, you know, the three different artificial analysis indexes that are the most important.
00:22:38.720 --> 00:22:51.519
Uh, you know, they're one point behind on one winning the other two, but they're doing it at yeah, a dollar per task, a dollar and four cents versus uh Fable 5,$2.75.
00:22:52.240 --> 00:22:52.880
Yeah.
00:22:53.359 --> 00:22:54.079
Entropic.
00:22:54.319 --> 00:22:54.559
All right.
00:22:54.720 --> 00:22:59.599
So uh we went over Chad GPT work, we went over codex.
00:22:59.920 --> 00:23:04.720
Now I'm gonna give you the five overlooked features and my one hot take.
00:23:04.799 --> 00:23:07.359
All right, overlook feature number one.
00:23:07.519 --> 00:23:12.160
Even though OpenAI didn't say it out loud, this is the super app, right?
00:23:12.240 --> 00:23:26.400
Uh, I think there are a lot of super app um, you know, rah-rah back in like March, April, May, um, and at the time OpenAI said, well, codex is the super app, uh, right.
00:23:26.559 --> 00:23:30.400
They said, you know, they said we're building it in disguise, but this is it.
00:23:30.720 --> 00:23:38.000
But is bringing together chat work, codex, browser, files, and actions.
00:23:38.079 --> 00:23:42.640
And if you know absolutely nothing about like, well, codex, right?
00:23:42.720 --> 00:23:43.920
Or why does it matter?
00:23:44.000 --> 00:23:51.519
You know, why would why do I need a desktop, you know, something running on my desktop versus something, you know, that I can just do in the cloud, right?
00:23:51.680 --> 00:23:56.319
If ChatGPT has this new ChatGPT work on the web, why would I ever use Codex?
00:23:56.799 --> 00:24:01.039
Well, the main reason is it can access your local system.
00:24:01.200 --> 00:24:06.720
So it can use your entire computer, it can use your browser, uh, right?
00:24:06.880 --> 00:24:11.279
It can access, read and write all of your local files.
00:24:11.440 --> 00:24:13.759
So those are you know three big advantages.
00:24:13.920 --> 00:24:17.599
And, you know, there's a little bit more that you can do with scheduling and things like that.
00:24:17.680 --> 00:24:24.240
If you have a machine like I do that just stays on, my poor Mac Studio, I don't think, has taken a nap since I got it.
00:24:24.319 --> 00:24:24.720
Uh, right.
00:24:24.799 --> 00:24:29.440
So I have a lot of things that are scheduled, you know, loops, goals, heartbeats, all these things.
00:24:29.599 --> 00:24:38.480
There's a little bit more uh technical um capabilities that you can do with the desktop app, but it is now all one.
00:24:38.799 --> 00:24:44.799
There's some syncing issues, FYI, and I'll probably reach out to OpenAI about it.
00:24:44.880 --> 00:24:51.680
I I don't know if they're fully um aware because, as an example, some things change names, right?
00:24:51.839 --> 00:24:55.759
So if you were using codecs before, there was a thing called chats.
00:24:55.920 --> 00:24:57.599
Now there's no codex chats.
00:24:57.759 --> 00:25:04.640
Codex chats are now tasks, so they could be called codex tasks or they could be called chat GPT work tasks, anyways.
00:25:04.799 --> 00:25:20.640
But you can there's a new chats panel, and that brings in your chats from ChatGPT, but right now at least only chats from the chat tab inside chat GPT are syncing and not the chats from the work tab inside chat GPT.
00:25:20.720 --> 00:25:22.240
Yeah, confusing, right?
00:25:22.400 --> 00:25:29.519
So there are some things that you know, the the the super app and bringing all this together in one cohesive ball.
00:25:29.680 --> 00:25:33.200
I think that there's a couple of things that the team is probably gonna tighten up.
00:25:33.359 --> 00:25:37.680
Um, but I mean, first iteration, it is really, really good.
00:25:37.839 --> 00:25:46.400
All right, the other thing, small footnote, sites, all right, chat GPT sites rolling out to all paid plans.
00:25:46.559 --> 00:25:46.880
All right.
00:25:46.960 --> 00:25:55.440
I think there's some might be some uh geography restrictions at first, but uh rolling out to all paid plans, and this is important.
00:25:55.599 --> 00:26:02.000
Also, there's public publishing, live URLs, and it's on Chad GPT on the web, not just codecs.
00:26:02.079 --> 00:26:05.839
So we are starting to see, you know, a little bit more feature parity.
00:26:05.920 --> 00:26:12.799
Uh, you know, certain like a good example is uh Chad GPT sites used to only be on codecs.
00:26:12.880 --> 00:26:17.599
So, well, now that's also uh on Chad GPT on the web as well.
00:26:17.759 --> 00:26:25.839
So here's why this is important because before sites was number one only available on Codecs, so now it's available on the web as well.
00:26:26.000 --> 00:26:30.240
Uh, but it's also um available for all paid plans.
00:26:30.400 --> 00:26:32.559
Before it was just for team plans.
00:26:32.720 --> 00:26:44.480
If if if I'm being honest, like it was one of the reasons that I still was holding on uh to my team plan, uh, even though I'm mainly the only person that uses it, was for ChatGPT sites.
00:26:44.559 --> 00:26:49.839
I really like it, but on my$200 a month pro plan, it's like I wanted to be able to use sites now.
00:26:50.000 --> 00:26:50.319
I can't.
00:26:50.400 --> 00:26:52.960
So if you have a paid plan, ChatGPT sites is great.
00:26:53.039 --> 00:26:55.759
It's like a a lovable light, right?
00:26:55.920 --> 00:26:59.839
So all you know, uh I I did a full episode on it a couple of weeks ago.
00:26:59.920 --> 00:27:00.799
Think of it like this.
00:27:00.960 --> 00:27:04.640
Um, you know, think of how many versions of decks that you have, right?
00:27:04.720 --> 00:27:09.359
There's 10 different versions on 10 different people's machines, v1, v3, v4, final.
00:27:09.599 --> 00:27:11.440
Oh, did we have edits on this one, right?
00:27:11.599 --> 00:27:18.559
You probably waste uh hundreds of hours a year just dealing with uh who has that file, right?
00:27:18.720 --> 00:27:23.200
ChatGPT sites can just get rid of that because it's just one source of truth.
00:27:23.279 --> 00:27:26.400
And anytime one person updates it, it's updated everywhere.
00:27:26.480 --> 00:27:38.240
So you can build full-on you know, databases, dashboards, uh, you know, little apps that your team can use or just decks and websites to keep track of anything type of work.
00:27:38.400 --> 00:27:42.559
Uh, so it's nice that it's rolling out to all plans and available on the web.
00:27:42.720 --> 00:27:46.799
Next, Chad GPT Mobile got a huge update.
00:27:46.960 --> 00:27:51.359
Uh, it now you can manage plugins from Chad GPT app.
00:27:51.599 --> 00:27:54.799
You can see and manage all of your schedule tasks.
00:27:54.960 --> 00:27:59.440
So uh that's great, but also uh they changed the name.
00:27:59.599 --> 00:28:08.640
So you used to open up the Chad GPT app and it would say codex, and then you could essentially control your uh any physical machine that was running codecs.
00:28:08.799 --> 00:28:12.079
So now it's just called remote, which I kind of like, right?
00:28:12.160 --> 00:28:16.079
So they're not slapping the name codecs everywhere and confusing people.
00:28:16.160 --> 00:28:19.519
And if you see remote, you're like, cool, even if you're a non-technical person.
00:28:19.680 --> 00:28:28.400
And now all of a sudden you're like, wait, I can use the chat GPT app, which you know, uh OpenAI has like a billion weekly active users, and everyone knows the chat GPT app.
00:28:28.480 --> 00:28:36.319
And then they see this thing called remote, and they're like, Oh, I can, you know, control my desktop if I have the ChatGPT work app.
00:28:36.480 --> 00:28:37.279
Wow, great.
00:28:37.440 --> 00:28:55.359
So again, not really new functionality there, it's just cleaned up and it does work a little bit better, but actually, no, new, there is new functionality because you have more control over plugins and scheduled tasks on the actual codep, uh, sorry, uh, on the actual remote section of the ChatGPT app.
00:28:55.839 --> 00:28:59.200
All right, number four, uh, yeah, Atlas is technically gone.
00:28:59.359 --> 00:29:05.759
So OpenAI's browser experiment, it is no more, but it's for the best because now this is gonna sound small.
00:29:05.920 --> 00:29:16.720
All right, but two big updates to uh the browser, uh formerly known as Atlas, that is now just baked into the super app, right?
00:29:16.960 --> 00:29:31.359
So Codex had always had a browser, but it was a singular browser, and it didn't have all the normal browser capabilities, like you couldn't open multiple tabs and it didn't keep your saved passwords.
00:29:31.599 --> 00:29:38.079
I actually spent like 1.8 billion tokens building what I called, I forgot what I called it.
00:29:38.160 --> 00:29:55.920
I think I just called it like Codex Vault, but I essentially built an app that would save all of my passwords so I could run, you know, certain things agentically that required you to log in because you know the codex browser, you know, couldn't uh log in to things without saved passwords.
00:29:56.079 --> 00:30:05.920
Um, so that is available now, and you can import from any of your Chrome profiles because the browser is based on Chromium like Atlas was.
00:30:06.079 --> 00:30:10.720
So uh it supports cookies to keep you logged in and it'll support passwords.
00:30:10.960 --> 00:30:26.079
So I was playing around with it a little bit um because it was a little finicky to actually get computer used to click login, because some websites, depending on how they're set up, you know, um the actual browser control can log you in.
00:30:26.160 --> 00:30:31.359
So it'll pull up your saved, you know, uh websites, uh, your saved credentials, right?
00:30:31.440 --> 00:30:38.400
Just like in Chrome, if you you know sign into something and you've saved your password in Chrome, it'll be there and you just have to click log in.
00:30:38.559 --> 00:30:41.200
Um, so I had to toy around with it a little bit.
00:30:41.359 --> 00:30:49.279
I might just uh if I can refine it, I just might save it as a skill and you know put it in our newsletter or on our website because it is a little finicky now.
00:30:49.359 --> 00:30:50.640
I'm sure they'll get it fixed.
00:30:50.799 --> 00:30:52.799
But this is the big unlock here, y'all.
00:30:52.880 --> 00:30:53.759
And here's why.
00:30:53.920 --> 00:30:59.519
Yes, you know, there's there's connectors and plugins and mcps and all these things, right?
00:30:59.680 --> 00:31:09.920
But I think for knowledge work, so much of what we do is going to all of these different websites that don't have MCP and they don't have direct connectors, right?
00:31:10.000 --> 00:31:16.319
So, in my example, you know, the our podcast is hosted on BuzzSprout, which is an older host.
00:31:16.480 --> 00:31:18.319
They don't have any AI anything.
00:31:18.480 --> 00:31:24.720
I spend so much time, you know, going from this program, uh, you know, downloading this, uploading this file here.
00:31:24.799 --> 00:31:30.400
I clean the audio, then I bring it in, you know, get the timestamps, and it's like five different programs.
00:31:30.640 --> 00:31:51.359
Well, now I can just fully automate that completely because codex can open multiple tabs uh at once and it can save the passwords because you know it does save them, it had previously saved them for a certain amount of time, but every once in a while it just kicks you out like any normal uh you know SaaS program would.
00:31:51.440 --> 00:31:59.759
So now with the cookies and the saved passwords, huge in terms of the new knowledge work that you can do on a schedule.
00:31:59.920 --> 00:32:01.680
Think of all those monotonous tasks.
00:32:01.759 --> 00:32:04.480
You have to go log in, click this thing 12 times.
00:32:04.640 --> 00:32:06.079
You're like, why am I doing this?
00:32:06.240 --> 00:32:06.480
Right?
00:32:06.640 --> 00:32:15.920
Why, you know, even if you're a high-paid executive, you're like, I I, you know, I just every single day I have 45 minutes of things that you know a monkey could do, right?
00:32:16.079 --> 00:32:18.880
I just go here, click this, wait for it to load, three minutes.
00:32:19.039 --> 00:32:20.160
Oops, got distracted.
00:32:20.319 --> 00:32:22.480
Click this, export, download, right?
00:32:22.640 --> 00:32:24.000
You don't have to do any of that anymore.
00:32:24.079 --> 00:32:24.640
It's amazing.
00:32:24.720 --> 00:32:28.720
All right, and then last but not least, computer use is much faster.
00:32:28.880 --> 00:32:33.039
It is faster, it is better, and it has this real cool new thing.
00:32:33.119 --> 00:32:34.799
It's this picture in picture.
00:32:34.960 --> 00:32:52.640
Uh, so when computer use is happening, um you can see, you know, if you've ever seen picture in picture on your TV, uh, it's just right there in Codex, or if you decide to call your codex chat GPT work, uh, so you can see what's happening um right there in codex.
00:32:52.720 --> 00:33:00.640
Oh, another thing that I didn't mention, which I should have mentioned on number four with Atlas, um, related to this and related to this super app.
00:33:00.720 --> 00:33:04.000
And I don't even think OpenAI announced this.
00:33:04.160 --> 00:33:17.759
Um, and I didn't see anyone else talk about it, but in the Chrome extension, because you can also from Codex not just control your computer, not just control the built-in browser, but you can also control a Chrome browser, which is a little different.
00:33:17.920 --> 00:33:18.640
Uh, all right.
00:33:18.720 --> 00:33:31.680
Uh, but they actually have a Chrome extension, um, and there's an update in the Chrome extension where you can have a side panel in Chrome and you can see what is being worked on on your desktop, which is really cool.
00:33:31.839 --> 00:33:40.240
So, you know, if I have a a window open, which is actually unlocks a lot because normally on my setup, right, I have two screens.
00:33:40.400 --> 00:33:46.720
One screen is just full, you know, full codex, uh, you know, and then I'll have, you know, probably a full clawed desktop.
00:33:46.880 --> 00:33:50.160
Um, and then I have to swap between them because I'm like, oh, okay.
00:33:50.319 --> 00:33:56.319
And I know, you know, codex has the little pets and they say, all right, your thing's ready, but I'm still like having to always keep an eye on that.
00:33:56.480 --> 00:34:04.160
So having a sidebar in Chrome and I can see what Codex is working on and track it live.
00:34:04.319 --> 00:34:06.079
That's been like a small little thing.
00:34:06.160 --> 00:34:08.639
And I'm like, wait, why didn't they talk about this?
00:34:08.800 --> 00:34:14.239
This is huge, especially if you're like me and you're giving you know agents long-running tasks.
00:34:14.400 --> 00:34:19.199
You know, at any point, I might have, you know, five to 20 different threads running.
00:34:19.280 --> 00:34:22.239
Um, and it's like, I don't want to always have to check in on them.
00:34:22.320 --> 00:34:26.000
And sometimes the little codex pet for me is a little distracting.
00:34:26.239 --> 00:34:31.039
And I'm always in a Chrome browser, so I'm like, oh, okay, I can just have that going and work with it right there.
00:34:31.199 --> 00:34:32.400
Actually, amazing.
00:34:32.559 --> 00:34:32.960
All right.
00:34:33.199 --> 00:34:37.920
So that is the five um overlooked features.
00:34:38.000 --> 00:34:40.960
And let me end with this hot take.
00:34:42.800 --> 00:34:48.559
Uh for those of you that stuck around, Anthropic's in trouble.
00:34:49.840 --> 00:34:51.679
They are in trouble.
00:34:51.920 --> 00:34:57.199
If they do not immediately pivot, it's not gonna look good.
00:34:57.519 --> 00:34:57.920
All right.
00:34:58.079 --> 00:35:04.800
And here's the thing Yeah, Anthropic is scared, and they should be, right?
00:35:04.880 --> 00:35:07.840
And this is great because it's going to drive competition.
00:35:08.079 --> 00:35:12.719
Anthropic did something that I don't think they've ever done right after this announcement.
00:35:12.800 --> 00:35:18.639
And this and this is how I know Anthropic is in huge, huge trouble.
00:35:19.119 --> 00:35:27.760
Um, they reset weekly limits, which I don't think they've ever done, except one time when they made a big mistake.
00:35:27.840 --> 00:35:29.360
There's this big miscommunication.
00:35:29.519 --> 00:35:33.440
I think they might have reset five-hour limits, not weekly limits.
00:35:33.599 --> 00:35:37.039
That's something if you use codecs, you like we've been spoiled, right?
00:35:37.119 --> 00:35:41.280
Tebow over there uh at OpenAI always resetting your weekly limits.
00:35:41.599 --> 00:35:46.800
Infrappit did this, they never do it, and they did it right when OpenAI announced something.
00:35:46.880 --> 00:35:53.920
And that to me is like they know they're scared, and they know I'm I'm I'm not gonna make projections, man.
00:35:54.079 --> 00:35:55.039
I'll make a projection.
00:35:55.119 --> 00:36:02.320
I I'd say it's very, very safe to say that yesterday they lost probably eight figures in revenue, I would guess.
00:36:02.480 --> 00:36:11.360
Um, and I will continue to say that until they make a change, they're gonna be losing eight figures of revenue every single day, maybe more.
00:36:11.519 --> 00:36:15.199
It's they're gonna be losing a lot of money if they do not pivot immediately.
00:36:15.360 --> 00:36:18.079
So I even heard some people say, and I'm not gonna say this, right?
00:36:18.320 --> 00:36:20.559
People are like, oh, Fable Five is dead.
00:36:20.719 --> 00:36:26.239
It's not dead, uh, it's still, you know, a top-tier model, and there's still gonna be some people out there paying for it.
00:36:26.559 --> 00:36:31.360
But I would absolutely not pay a dime for Fable V.
00:36:31.599 --> 00:36:38.639
And Anthropic, they said, oh, we're gonna, you know, extend Fable V in uh subscriptions through July 12th.
00:36:38.800 --> 00:36:43.519
Uh well, if they don't extend it past that, they're gonna lose subscribers.
00:36:43.599 --> 00:36:47.199
But the majority uh of Anthropic's business comes from two main things.
00:36:47.360 --> 00:36:50.480
It comes from coding via the API, um, right?
00:36:50.639 --> 00:36:54.880
And up until 5.6 soul, um, they had a sizable lead, right?
00:36:55.039 --> 00:36:58.159
Fable five um had a sizable lead over five.
00:36:59.199 --> 00:37:02.000
From GPT and you know, Google, we'll see what happens.
00:37:02.159 --> 00:37:09.039
Uh, but here's the thing I don't think anthropic is worried about GPT 5.6 soul.
00:37:09.519 --> 00:37:15.360
They are worried about GPT 5.6 Terra.
00:37:15.760 --> 00:37:16.000
Okay.
00:37:16.400 --> 00:37:17.199
Here's why.
00:37:17.440 --> 00:37:21.119
So Anthropic makes money from coding.
00:37:21.599 --> 00:37:23.119
Yes, they have co-work.
00:37:23.280 --> 00:37:24.639
Yes, that's popular.
00:37:25.199 --> 00:37:29.519
They make their money from coding, but they sell tokens at an extreme premium.
00:37:29.840 --> 00:37:30.480
All right.
00:37:32.000 --> 00:37:38.480
Looking at the artificial analysis coding index, and that takes into account a lot of the popular coding benchmarks.
00:37:38.639 --> 00:37:40.719
Um, you know, soul, GPT-5.
00:37:43.119 --> 00:37:44.800
But guess who's in second place?
00:37:45.199 --> 00:37:48.960
GPT-5.6 Terra, the medium model.
00:37:49.360 --> 00:37:52.239
It is better than the big bad fable.
00:37:52.559 --> 00:38:01.039
The one that we had uh, you know, have a government anthropic fight, the one that, you know, anthropic had been, you know, saying this thing's too dangerous.
00:38:01.119 --> 00:38:14.239
They've been they were hyping it up for for three, three months, and now we have this uh you know permission slip AI uh in the US because of anthropic overhyping uh Fable and Mythos.
00:38:14.559 --> 00:38:14.880
Okay.
00:38:15.519 --> 00:38:20.559
The medium model is eating Fable Five's lunch.
00:38:20.719 --> 00:38:21.760
And why are they eating it?
00:38:21.840 --> 00:38:33.280
Not just because it's increment incrementally better at coding, it is more than five times cheaper, according to artificial analysis.
00:38:34.639 --> 00:38:40.960
Five times cheaper to do the same tasks versus Fable Five.
00:38:41.440 --> 00:38:41.760
Yeah.
00:38:41.920 --> 00:38:50.000
So what I said on the on the Twitter machine, I said anthropic has like a week or two until enterprises start catching up.
00:38:50.079 --> 00:38:57.119
Uh, because that's how long it takes for even the decision makers that are spending multiple millions of dollars sometimes a month on these products.
00:38:57.199 --> 00:39:07.840
And yes, there's vendor lock-in, but any company that's up for, you know, if they signed a one-year uh deal with anthropic, there's always companies rolling.
00:39:08.320 --> 00:39:09.679
They're gonna churn.
00:39:09.760 --> 00:39:16.719
There is zero, zero reason for any company right now to be paying for anthropic.
00:39:17.039 --> 00:39:17.519
Zero.
00:39:18.239 --> 00:39:20.719
Because of what OpenAI just did.
00:39:20.880 --> 00:39:24.239
They undercut on price, they overdelivered on performance.
00:39:24.480 --> 00:39:30.960
And if Fable, if anthropic doesn't change Fable Five pricing, it is DOA.
00:39:31.199 --> 00:39:31.920
It's done.
00:39:32.079 --> 00:39:40.400
So they've got to either quickly release Fable 5.1 um so they can, you know, knock down the pricing on Fable 5.
00:39:40.559 --> 00:39:46.159
Or I do think that they will be bleeding five figures potentially daily, right?
00:39:46.239 --> 00:39:51.039
So not just from people going over on contract renewals, but just on the API side.
00:39:51.119 --> 00:40:15.679
Um, it's it's gonna be interesting to look at these third-party sites um like open router that track uh you know third-party token usage, because I I would assume you're gonna see the open AI share go up exponentially, and you're gonna see people that were using at least for Fable um go down, Opus go down, because why would you use Opus now when Terra is cheaper?
00:40:15.920 --> 00:40:21.199
So a big portion of Anthropic's revenue just got cut off at the knees.
00:40:21.280 --> 00:40:24.159
And if nothing else, this is just fun for us, right?
00:40:24.239 --> 00:40:30.480
We get to watch, but that means we're gonna get number one, a better model from Anthropic sooner rather than later.
00:40:30.639 --> 00:40:38.320
They have to, uh, and they're gonna have to cut their prices and or uh extend and keep Fable Five in subscriptions.
00:40:38.400 --> 00:40:40.960
There's really no other option, right?
00:40:41.039 --> 00:40:45.440
Uh, because they can't just keep bleeding money because they're gonna be bleeding money.
00:40:45.519 --> 00:40:45.840
All right.
00:40:46.000 --> 00:40:46.880
So that's it.
00:40:47.039 --> 00:40:54.159
Now you know Chat GPT work, you know GPT 56 Soul, what's new, the five overlooked features, and my one hot take.
00:40:54.239 --> 00:40:55.280
I hope this was helpful.
00:40:55.440 --> 00:40:59.599
If so, please go to your everydayai.com, sign up for the free daily newsletter.
00:40:59.679 --> 00:41:00.239
Uh, thanks.
00:41:00.320 --> 00:41:05.039
We'll see you uh, well, not tomorrow, but Monday and every day for more everyday AI.
00:41:05.119 --> 00:41:06.159
Thanks, y'all.
