WEBVTT
00:00:01.280 --> 00:00:06.719
Three and a half years ago, AI was kind of like a party trick.
00:00:06.959 --> 00:00:19.760
I mean, if your entire company or department was using AI in late 2022, people probably would have thought that you were crazy or you could have easily gotten into some hot water.
00:00:20.239 --> 00:00:22.160
Oh, how the times have changed.
00:00:22.399 --> 00:00:28.239
I mean, now if your entire department isn't using AI, you're definitely in hot water.
00:00:28.399 --> 00:00:34.799
That's because there's a good chance that you, your company or your department might be sinking if you're not.
00:00:34.960 --> 00:00:43.119
That's because the rise in prominence and capabilities of AI has been astronomical over the past three years.
00:00:43.200 --> 00:00:53.600
And we've gone from AI chatbots that were kind of fun to play around with to straight up autonomous coworkers that are doing the work for enterprises.
00:00:53.840 --> 00:01:10.799
So that's what we're going to be tackling on today's show, volume 10 of our Start Here series, giving you a quick overview of how AI has gone from a chatbot to an autonomous work coworker and how consumer AI has changed.
00:01:11.040 --> 00:01:12.799
And what's next?
00:01:13.359 --> 00:01:13.920
All right.
00:01:14.239 --> 00:01:17.120
And if you are brand new here, awesome.
00:01:17.280 --> 00:01:18.480
Great time to jump in.
00:01:18.640 --> 00:01:21.760
Uh, this is part of our start here series.
00:01:21.920 --> 00:01:29.200
This is our essential podcast series to both learn the AI basics and to double down on your AI knowledge.
00:01:29.359 --> 00:01:33.519
So, like I said, we are on volume 10 of our Start Here series.
00:01:33.680 --> 00:01:43.200
And if you want to go listen to the entire uh series in order, uh, I actually just dropped a Spotify playlist so you can do just that super easy.
00:01:43.280 --> 00:01:46.640
So make sure you go to starthearseries.com.
00:01:46.799 --> 00:01:51.200
So that's also going to give you free access to our inner circle community.
00:01:51.280 --> 00:01:57.120
So you can go learn and network with other business leaders like yourselves that are learning AI.
00:01:57.200 --> 00:02:01.359
Go listen to every single uh Start Here series, read about it.
00:02:01.439 --> 00:02:03.519
It's all in a dedicated space there.
00:02:03.760 --> 00:02:07.840
And you'll get access to our free Prime Prompt Polish course.
00:02:08.000 --> 00:02:10.639
And hey, talking about starting here, right?
00:02:10.800 --> 00:02:15.360
Because after 700 plus episodes, I didn't have an answer for what people said.
00:02:15.520 --> 00:02:16.400
Where should I start?
00:02:16.560 --> 00:02:18.719
When listening to your podcast, I was like, I don't know.
00:02:18.800 --> 00:02:20.960
Well, listen to the Start Here series in order.
00:02:21.039 --> 00:02:25.680
But along the way, make sure you go listen to episodes 713 and 712.
00:02:25.759 --> 00:02:29.199
That is our 2026 AI prediction and roadmap series.
00:02:29.360 --> 00:02:29.840
All right.
00:02:30.080 --> 00:02:40.240
Last episode in Start Here, uh, we talked about agent risk security and AI sprawl in 2026, why AI that acts changes everything.
00:02:40.319 --> 00:02:44.879
So make sure you go listen to that, the last uh episode of our Start Here series.
00:02:44.960 --> 00:02:50.319
And now let's jump into today's from AI Chatbot to autonomous coworkers.
00:02:50.479 --> 00:02:59.360
So on today's show, all right, I'm gonna go over the five phases of AI enabled work and why most teams are still stuck in phase one or two.
00:02:59.439 --> 00:03:07.120
Uh, I'm gonna tell you how and why every major AI company is suddenly racing to put agents on your desktop.
00:03:07.280 --> 00:03:15.039
Uh, you're gonna learn about the hidden layer that almost no one is talking about that changes how AI works for you or your company entirely.
00:03:15.199 --> 00:03:19.680
And then we're gonna end with my five actionable takeaways for what's next.
00:03:19.759 --> 00:03:28.000
Because yeah, it's been a fast sprint from AI chatbot to autonomous coworker, and we are not done, obviously.
00:03:28.319 --> 00:03:35.199
So let's first start with a zoomed out overview of what the heck has happened.
00:03:35.360 --> 00:03:38.479
Uh, so large language models are not new, right?
00:03:38.560 --> 00:03:46.639
They've been around for a long time, uh, technically well before even the uh Chat GPT moment of November 2022.
00:03:46.719 --> 00:03:48.719
But let's start there, shall we?
00:03:48.960 --> 00:03:58.719
Okay, so that's kind of when the whole right generative AI and large language model phase in the corporate world kicked off was in November 2022, right?
00:03:58.879 --> 00:04:06.960
Uh many people, including NVIDIA um CEO Jensen Wong, call that uh the the line in the sand, right?
00:04:07.520 --> 00:04:10.719
So November 2022, ChatGPT launches.
00:04:10.879 --> 00:04:14.560
Uh, you know, it's a simple QA chatbot in a browser tab.
00:04:14.719 --> 00:04:15.199
All right.
00:04:15.439 --> 00:04:19.839
Then let's fast forward to where we are today, February 2026.
00:04:20.000 --> 00:04:27.360
I mean, we have AI agents that can run recurring tasks on your desk desktop with local file access, right?
00:04:27.600 --> 00:04:34.319
So at that point, they can use your browser, they can move files, they can upload, they can download, right?
00:04:34.720 --> 00:04:38.639
At that point, they are an actual autonomous AI coworker.
00:04:38.800 --> 00:04:40.800
So that's where we started, right?
00:04:40.879 --> 00:04:45.279
This is like one of those uh social media posts, like how it started, right?
00:04:45.360 --> 00:04:49.680
Uh in AI chatbot that was a party trick that didn't really know anything.
00:04:49.839 --> 00:04:52.959
To okay, this thing runs just like a human would.
00:04:53.120 --> 00:05:02.399
It can work on your desktop, it can use your files, it can move your files, upload, download, use your browser, uh, right, access all your data.
00:05:02.560 --> 00:05:04.240
Oh, how the times have changed.
00:05:04.319 --> 00:05:07.600
Uh, in even just the last few days and weeks, right?
00:05:07.839 --> 00:05:13.680
Uh again, you may be listening to this in late 2026 or early 2027.
00:05:13.759 --> 00:05:14.240
I don't know.
00:05:14.399 --> 00:05:26.720
But if you're listening to it in February, uh, right, today, February 27th, I mean, the last 48 hours alone, we've seen a ton of movement that I think signal where this is headed, right?
00:05:26.879 --> 00:05:32.319
We got perplexity computer, uh, co-pilot scheduled tasks from Microsoft.
00:05:32.399 --> 00:05:35.600
Uh, we got remote control and scheduled tasks from Anthropic.
00:05:35.680 --> 00:05:50.399
So, you know, three of the big five players there uh in Perplexity, Microsoft, and Anthropic just released pretty big updates that signal toward this more scheduled autonomous desktop computer worker.
00:05:51.040 --> 00:05:51.600
All right.
00:05:52.959 --> 00:05:55.199
So here's the five phases.
00:05:55.680 --> 00:05:58.879
All right, so uh phase one, online chat bots, right?
00:05:58.959 --> 00:06:01.600
They answer your questions in a browser tab.
00:06:01.759 --> 00:06:03.839
Uh phase two, business chat.
00:06:04.000 --> 00:06:09.920
This is where you know answers started to come with your company's data and using apps.
00:06:10.079 --> 00:06:13.519
Uh phase three, we have the agentic era, right?
00:06:13.600 --> 00:06:19.040
So this is both agenc models and multi-step execution toward an outcome, right?
00:06:19.120 --> 00:06:21.600
So more outcome-based with multiple steps.
00:06:21.759 --> 00:06:24.480
Phase four, this is your AI coworkers, right?
00:06:24.639 --> 00:06:27.120
Delegate a goal and get finished work.
00:06:27.360 --> 00:06:33.040
And then phase five, desktop AI, agents that work with your local files, right?
00:06:33.360 --> 00:06:42.079
And these are kind of in order, but as you'll see, as we talk about each phase a little bit more, technically there's some overlap, right?
00:06:42.240 --> 00:06:47.279
Uh, we're still obviously getting updates to phase two and phase five at the time, right?
00:06:47.360 --> 00:06:51.040
It doesn't mean phase two is over just because phase five has started, right?
00:06:51.120 --> 00:06:51.680
That's not where it is.
00:06:51.759 --> 00:06:56.000
It's it's more of kind of the uh official eras of when things started.
00:06:56.079 --> 00:07:16.480
And I think, you know, for the most part, I'd say phase one is probably phased out, although unfortunately, I still think people use ChatGPT just as an online chatbot to get answers without, you know, using any of the agentic capabilities, without using any of the business capabilities, without using any of the more coworker type capabilities.
00:07:16.639 --> 00:07:22.800
So I think for the most part, people have kind of uh phased out of that era, but not completely.
00:07:23.279 --> 00:07:35.920
And then there's the feature or the layer on top of these five phases that I think might be even more important in really changing how works get how work gets done.
00:07:36.079 --> 00:07:38.000
And that's scheduling, right?
00:07:38.079 --> 00:07:46.879
So I just mentioned that in the last 48 hours, we've seen literally three big updates just in this space.
00:07:47.199 --> 00:07:52.319
But this does live across all phases, so it's not its own phase necessarily, but right.
00:07:52.399 --> 00:07:56.879
Uh, and Google and OpenAI have had this for a pretty long time, uh, right?
00:07:56.959 --> 00:08:02.399
It's about six to 12 months, a little bit longer, but it's rarely talked about until recently, right?
00:08:02.560 --> 00:08:13.279
If you go back to my uh original Chat GPT tasks show, uh and it's a huge bummer that open AI changed how tasks work in uh in uh chat GPT, it's actually weird.
00:08:13.360 --> 00:08:18.639
If you're on the pro plan, you don't even have scheduled tasks anymore, which is one of the things I use most, right?
00:08:18.720 --> 00:08:19.759
Now it's just pulse.
00:08:20.000 --> 00:08:23.680
Um, in scheduled tasks, you can only really use in agent mode, anyways.
00:08:23.839 --> 00:08:26.800
Uh, it's it's rarely been talked about until recently.
00:08:26.879 --> 00:08:34.639
That's because of all the changes in the last 48 hours with Perplexity, Anthropic, and Microsoft coming in with some new offerings there.
00:08:35.200 --> 00:08:41.840
All right, well, now you have the answers, but if you want to dig in a little bit deeper, now's where we're gonna do it.
00:08:41.919 --> 00:08:43.919
So let's talk a little bit about phase one.
00:08:44.080 --> 00:08:51.919
So that started with Chat GPT launching in November of 2022, and it was obviously mega viral, right?
00:08:52.000 --> 00:08:53.440
I remember at the time, right?
00:08:53.519 --> 00:08:58.240
So my company, we had been using the GPT technology since it came out in 2020.
00:08:58.320 --> 00:09:00.080
So a lot of people overlooked that.
00:09:00.240 --> 00:09:08.000
Before there was Chat GPT, OpenAI put their product out there, uh GPT 3, uh, in other pieces of software.
00:09:08.159 --> 00:09:13.600
So when ChatGPT came out, at least for me, and I used it in November, I'm like, this stinks, right?
00:09:13.679 --> 00:09:15.440
It's not wasn't it was terrible.
00:09:15.600 --> 00:09:21.840
It was worse than using the other platforms that had been out for two years, obviously, because they had uh the time to improve their platforms.
00:09:22.000 --> 00:09:27.679
But regardless, this is what put today's AI and large language models on the map.
00:09:27.759 --> 00:09:30.559
And it did technically, I think, change the course.
00:09:30.639 --> 00:09:33.039
And this might sound corny, but it's true.
00:09:33.200 --> 00:09:37.519
I think it's gonna change the course and has changed the course of business history.
00:09:37.759 --> 00:09:40.639
All right, then uh you had a little bit later, right?
00:09:40.720 --> 00:09:48.240
About three, four months later, you had Google launching Bard in March, and then also Anthropic launching Claude in March.
00:09:48.320 --> 00:09:54.159
Although uh Claude was not uh generally available outside of kind of an enterprise pilot, I believe, until that summer.
00:09:54.399 --> 00:09:57.360
But at the time, this is just online chatbot era, right?
00:09:57.519 --> 00:09:58.480
Ask a question.
00:09:58.639 --> 00:10:08.000
So a couple huge reasons why in uh this first phase, it this wasn't suited for the business world, right?
00:10:08.080 --> 00:10:13.200
Uh, one of the reasons large language models at the time did not have access to the internet.
00:10:13.519 --> 00:10:13.840
Okay.
00:10:14.080 --> 00:10:15.919
And that made them extremely dangerous.
00:10:16.080 --> 00:10:29.039
And I think I depending on how you look at it, fortunately or unfortunately, it was really this phase one that I think is still kind of setting the tone for how a lot of people think about AI today, right?
00:10:29.200 --> 00:10:35.759
That's because hallucinations were rampant in phase one, because number one, uh, well, three reasons, I think.
00:10:35.840 --> 00:10:37.120
Number one, you couldn't upload data.
00:10:37.279 --> 00:10:40.320
Number two, it couldn't access the internet and models couldn't reason.
00:10:40.559 --> 00:10:52.799
So, unless you were really good at prompt engineering, there's a good chance that if you were trying to use a large language model in 2022 or early 2023 for any business context, it was probably pretty bad.
00:10:52.960 --> 00:10:53.279
All right.
00:10:53.360 --> 00:10:57.679
And that's why in 2022 and 2023, RAG was all the rage, right?
00:10:57.840 --> 00:11:11.279
Retrieval logmin and generation, because if you wanted trustworthy outputs from AI, you had to build your own rag pipeline because the outputs were absolutely terrible, just riddled with hallucinations because, well, the models were not capable.
00:11:11.360 --> 00:11:13.679
It was at the time, I think, more than a party trick.
00:11:13.759 --> 00:11:17.440
So no file access, no integrations, no real world actions, right?
00:11:17.600 --> 00:11:20.879
And the models were pulling on very old data, right?
00:11:21.039 --> 00:11:24.799
At the time, those first models that came out, right?
00:11:25.279 --> 00:11:29.120
Now the people training the models are legit superstars, right?
00:11:29.200 --> 00:11:34.240
Some of your heads of research are making tens of millions or hundreds of millions of dollars.
00:11:34.399 --> 00:11:34.960
Uh, right.
00:11:35.120 --> 00:11:37.279
At the time, I don't think that was the case, right?
00:11:37.360 --> 00:11:50.159
Obviously, they're very smart researchers and seasoned researchers, but uh, you know, just the emphasis that has been placed on getting good, clean, uh, reliable training data uh and just the work that can be done at inference has completely changed.
00:11:50.240 --> 00:11:51.840
All right, so that was phase one.
00:11:52.480 --> 00:12:04.559
Phase two, uh, I'd say probably started in like 2023, and it kind of, I won't say again, it didn't end in 2025, but I'll give this phase or this era 2023 to 2025.
00:12:04.720 --> 00:12:06.240
So this is team AI.
00:12:06.399 --> 00:12:07.519
This is business chat.
00:12:07.600 --> 00:12:09.440
This is when it was born, uh, right.
00:12:09.519 --> 00:12:12.159
So team AI that's connected to your company's data.
00:12:12.240 --> 00:12:17.679
Uh, so first, uh, I think technically first was Microsoft Copilot in late 2023.
00:12:17.759 --> 00:12:22.799
Um, and then uh also you had ChatGPT Enterprise in August 2023.
00:12:22.879 --> 00:12:31.679
Uh, and then you had uh Gemini for Workspace in February 24, and also Anthropic for Teams and Enterprises came out in 2024 as well.
00:12:32.240 --> 00:12:35.519
But it wasn't just going team first, right?
00:12:35.600 --> 00:12:36.960
That was a big first step.
00:12:37.120 --> 00:12:47.120
So, along with that, companies found out, right, or big AI labs, they're like, okay, this is great to get, you know, a hundred million consumers to check out your AI chatbot.
00:12:47.279 --> 00:12:50.240
But if we want to make this a tool that companies are gonna buy, right?
00:12:50.320 --> 00:12:55.600
Because at the time, I think it was Microsoft that was cleaning house with enterprise AI, uh, right.
00:12:55.679 --> 00:12:59.679
Because when they came out in 2023, uh, there weren't a lot of other options.
00:12:59.840 --> 00:13:04.080
If you wanted to get 10,000 people uh to use a product, right?
00:13:04.399 --> 00:13:08.240
You weren't gonna use a consumer chat bot that didn't have data.
00:13:08.480 --> 00:13:11.360
So this is really the the maturation phase.
00:13:11.440 --> 00:13:25.440
Uh, and I think things like Chat GPT connectors, cloud connectors, uh Google Apps, uh, right, all these things in 2025 uh really I think help solidify the space and also cut down on hallucinations.
00:13:25.519 --> 00:13:34.559
Uh and this is where AI could finally access your email, docs, and drives all in a dynamic way, but it really couldn't act on it, right?
00:13:34.720 --> 00:13:37.120
At least in the earlier parts of 2025.
00:13:37.200 --> 00:13:38.720
That's kind of in the next phase.
00:13:38.799 --> 00:13:46.159
But this is the beginning of uh, you know, AI going from fun party trick to okay, this could work, right?
00:13:46.240 --> 00:13:54.080
Uh kind of the experimentation phase of of 2023 and 2024 to I think by 2025, the business world knew.
00:13:54.240 --> 00:14:08.159
I would say I would I was hoping or assuming it would be in 2024, but I don't think it was maybe until early 2025, maybe late 2024, that everyone realized, oh yeah, the future of work is 100% AI and there's no way around it.
00:14:08.240 --> 00:14:17.360
I think you still had fence sitters uh in 2024 at the enterprise level, which is absolutely nutty to me because now I think those fence sitters are paying for it.
00:14:17.440 --> 00:14:20.879
Uh, but uh phase two really pushed it.
00:14:21.120 --> 00:14:28.159
Phase three, I'll say started in 2024, kind of through 2025, although it's still technically going on, right?
00:14:28.240 --> 00:14:29.279
And it's ever improving.
00:14:29.360 --> 00:14:31.200
But this is the agentic area.
00:14:31.360 --> 00:14:34.879
This is where you have AI that plans and executes, right?
00:14:35.120 --> 00:14:41.519
I think technically you could say 2023 because Microsoft had Copilot Studio at the very end of 2023.
00:14:41.600 --> 00:14:49.200
But I'll say 2024, because I think nothing was really uh, you know, um adapted at March until 2024.
00:14:49.279 --> 00:14:55.679
But this is when you had things like uh OpenAI launched launched Operator in January 2025, right?
00:14:55.759 --> 00:15:00.000
That was a big moment, although Operator was fairly bad and slow.
00:15:00.159 --> 00:15:06.480
Uh, but it it signaled a big shift toward AI that could think and do human work, right?
00:15:06.559 --> 00:15:11.519
The big thing was being able to access a browser, which is what operator could do.
00:15:12.080 --> 00:15:16.720
You know, it could uh it was an agent that could uh navigate the web autonomously.
00:15:16.879 --> 00:15:23.600
Uh, then I think you had this uh right thereafter, uh shortly thereafter, you have this push for agenc browsers, right?
00:15:23.759 --> 00:15:29.360
So different than an agent, it is a browser powered by a reasoning model.
00:15:29.519 --> 00:15:34.799
So, you know, an agentic browser, one of its main features is to agentically browse the web.
00:15:34.960 --> 00:15:44.960
So I think some tools like ChatGPT Atlas, uh Perplexity Comet, uh Google, you know, announced their uh agent mode, Project Mariner.
00:15:45.279 --> 00:15:47.279
Uh, they had their Gemini in Chrome.
00:15:47.519 --> 00:15:50.559
Uh, that was probably more uh, you know, late 2025.
00:15:50.720 --> 00:15:55.759
Uh, but the key shift here thing, uh I think was reasoning models, right?
00:15:55.840 --> 00:15:57.360
They powered all of this.
00:15:57.519 --> 00:16:04.720
You know, you don't have um worthwhile AI agents without reasoning models, all right.
00:16:04.799 --> 00:16:10.960
Models that can think, plan, uh use logic like humans can, all right.
00:16:11.120 --> 00:16:24.720
Um, and that's really what I think separates, you know, at least in in my opinion, it was the agencera um that I think is ultimately going to be the most impactful.
00:16:24.799 --> 00:16:26.399
It wasn't Chat GPT, right?
00:16:26.559 --> 00:16:34.159
Because if we never kind of had models that could reason or which that would mean we probably would never have AI agents, right?
00:16:34.320 --> 00:16:38.559
We would have more uh human duct tape than agentic output.
00:16:38.639 --> 00:16:49.440
So it wouldn't have been worth it for any businesses uh to truly invest in AI over the long term if we never entered into the agentic era or reasoning models, right?
00:16:49.600 --> 00:16:59.120
Um, so some of the models maybe you've heard of these, they're a little older by now, but you know, 01 preview really kicked it all off in September of 2024.
00:16:59.279 --> 00:17:10.640
Uh, then you had Gemini 2.0 Flash in December 24, and then Claude Sonnet37, I think was their first uh reasoning or hybrid model from Anthropic in February 2025.
00:17:10.880 --> 00:17:22.240
So a pretty big like eight-month period there where we kind of left the quote unquote old uh versions, you you know, the non-thinking transformer models, right?
00:17:22.319 --> 00:17:25.279
Even though they're still transformer models, they just think and reason, right?
00:17:25.359 --> 00:17:38.559
But I like I really say that's like the old school AI versus the new school AI, because I think what um agentic models can do and their capabilities, the scaffolding, the harnessing that continues to be improved, right?
00:17:38.720 --> 00:17:44.079
You can make the argument today, and I've talked with very smart people about this, right?
00:17:44.160 --> 00:17:51.039
Like the head of Microsoft Research that's been working in agents for 20 years, uh, the head of agents at Cloudflare, right?
00:17:51.119 --> 00:18:07.839
I've had so many conversations with extremely smart people in the space that have eventually essentially agreed that, yeah, if you're using, you know, uh GPT-5.2 Pro and you have all your business data connected to it, that's an agent, right?
00:18:08.079 --> 00:18:11.119
Especially when you can schedule it and it can act autonomously.
00:18:11.200 --> 00:18:12.640
It's like, yeah, that's an agent, right?
00:18:12.720 --> 00:18:19.839
Or if you're using, you know, Gemini 3.1 Pro and scheduling things and it has access to your data, that's an agent.
00:18:20.319 --> 00:18:20.640
Right.
00:18:20.799 --> 00:18:23.759
So uh, but it really started with the reasoning models.
00:18:23.839 --> 00:18:25.279
Uh, then we have phase four.
00:18:25.359 --> 00:18:26.960
This is AI coworkers.
00:18:27.039 --> 00:18:32.799
Uh, so where you delegate a goal and it just finishes work on its own, right?
00:18:32.960 --> 00:18:37.039
And yes, I know and understand that some of these phases start to blend together.
00:18:37.200 --> 00:18:38.240
I get it, right?
00:18:38.319 --> 00:18:44.880
But in my opinion, the big difference between the agentic era, I would say that's more model and browser based, right?
00:18:44.960 --> 00:18:52.720
And that's kind of the um the foundation or the stepping stones for today's AI coworkers or agent coworkers.
00:18:53.200 --> 00:18:59.599
And this is more or less general AI agents that have a virtual computer and can access your data.
00:18:59.759 --> 00:19:06.000
So, right, you can say that's you know, Chat GPT agent mode, uh, you know, Google Gemini agent mode, obviously.
00:19:06.079 --> 00:19:09.599
Uh, but I'd say things like Manaus, right?
00:19:09.839 --> 00:19:16.240
Huge uh successful launch uh for Manaus, uh, you know, recently uh acquired by Meta.
00:19:16.559 --> 00:19:18.799
Uh same thing with uh Gen Spark, right?
00:19:18.880 --> 00:19:24.559
Uh Gen Spark, uh another kind of general purpose agent that can browse the web, uh, right?
00:19:24.720 --> 00:19:28.079
It can access all your data, but for the most part, you know, scheduling.
00:19:28.319 --> 00:19:33.440
A lot of these AI coworkers are more browser-based, right?
00:19:33.599 --> 00:19:39.839
Uh, but you can give access to your data and then they can go use the web for the most part.
00:19:40.000 --> 00:19:43.359
They can, you can log, even though it can be risky, uh, right.
00:19:43.440 --> 00:19:47.440
But you know, they have a virtual computer, a sandbox, a terminal, all these things.
00:19:47.519 --> 00:19:54.160
Uh, you, you know, they can run code, you know, a lot of them can run different models, uh, you know, use uh sub agents, right?
00:19:54.240 --> 00:20:01.839
So uh this is kind of the uh the version of a virtual AI teammate that works more in the cloud.
00:20:02.079 --> 00:20:12.079
And then we also, the newest entry here, uh, you know, perplexity computer, uh, that was just launched, uh, that uses 19 uh AI models.
00:20:12.240 --> 00:20:12.400
Yeah.
00:20:12.480 --> 00:20:14.960
Let me know if we should do a show on perplexity computer.
00:20:15.039 --> 00:20:16.160
I've been thinking about it.
00:20:16.400 --> 00:20:16.799
All right.
00:20:16.960 --> 00:20:19.279
And then that leads us to phase five.
00:20:19.440 --> 00:20:22.160
And this is kind of where we're at now.
00:20:22.240 --> 00:20:25.200
Although, yes, we're still in phase four with AI coworkers.
00:20:25.279 --> 00:20:26.240
That's not going anywhere.
00:20:26.319 --> 00:20:29.680
We're still in the agentic area uh era, that's not going anywhere.
00:20:29.759 --> 00:20:32.000
But I think where we are today, right?
00:20:32.240 --> 00:20:37.839
The era that is maybe most recently started, that is desktop AI agents, right?
00:20:38.000 --> 00:20:50.880
So this is a little different in a step, uh, both in a um technically a more opportunistic direction, but also in a much more dangerous direction as well, right?
00:20:51.119 --> 00:21:05.599
Um, it is probably in most cases, and I think most experts would agree, somewhat safer uh to use a cloud uh-based agent that maybe can't access your local files, right?
00:21:05.759 --> 00:21:15.200
Even though they can oftentimes, uh if you grant them access, they can access your Gmail and you know your OneDrive, your SharePoint, or right, your Notion, right?
00:21:15.279 --> 00:21:23.839
You can add all these um different uh connectors and apps to the uh kind of virtual AI coworkers.
00:21:23.920 --> 00:21:29.279
But in phase five, this is agents that are running on your actual computer.
00:21:29.440 --> 00:21:34.240
So they can do everything uh that an AI coworker can do, right?
00:21:34.400 --> 00:21:40.799
A virtual uh AI agent, but they can control your actual computer.
00:21:41.039 --> 00:21:41.359
All right.
00:21:41.519 --> 00:21:58.079
So I think uh technically, I you could say, oh, this is more 2026 starting, but technically, I think Claude Code kicked this off in um February 2025, but it really got popular, I would say, in the fourth quarter of 2025, right?
00:21:58.160 --> 00:22:05.200
Uh so Cloud Code was essentially a Desktop coding program, um, that really what it turned into.
00:22:05.359 --> 00:22:16.480
I think some of the popular use cases that really exploded this category was actually people using Claude Code, a terminal tool for developers.
00:22:16.640 --> 00:22:20.960
Well, non-developers started to use it for well, non-technical work.
00:22:21.119 --> 00:22:25.440
And I think Anthropic realized that early on and capitalized on it.
00:22:25.519 --> 00:22:29.279
And that led to the launch of Cowork.
00:22:29.839 --> 00:22:34.160
Claude Co-Work in January 2026, which I absolutely love.
00:22:34.240 --> 00:22:36.240
I'm a big uh Claude Co-Work fan.
00:22:36.319 --> 00:22:42.000
I use Claude Code as well on the desktop version, not in the terminal, not really a terminal by myself.
00:22:42.079 --> 00:22:47.759
Uh, right, but this is now you have in the desktop version of Claude Code and Claude Co-Work.
00:22:48.000 --> 00:22:54.160
This is a desktop AI agent that can control your computer and work for hours, right?
00:22:54.400 --> 00:23:04.000
I'm never one that's intentionally trying to push uh desktop AI agents to go longer and longer, but I got codecs to work for 10 hours months.
00:23:04.079 --> 00:23:04.559
Uh, right.
00:23:04.640 --> 00:23:08.559
I've gotten uh Claude usually is a little faster, sometimes not as thorough.
00:23:08.640 --> 00:23:09.839
I personally prefer Codecs.
00:23:09.920 --> 00:23:11.279
I'm gonna talk about that here in a minute.
00:23:11.440 --> 00:23:21.200
But these are now desktop AI agents that have access to literally everything: your computer, your files, your notes, your your browser, everything.
00:23:21.359 --> 00:23:22.079
Uh right.
00:23:22.319 --> 00:23:31.200
So you can't talk about Claude Code and um uh Claude Cowork without talking about OpenAI's codecs.
00:23:31.279 --> 00:23:31.680
All right.
00:23:31.839 --> 00:23:35.680
And I am crazy bullish on codecs.
00:23:35.839 --> 00:23:36.960
I freaking love it.
00:23:37.039 --> 00:23:44.720
If you've been listening since codecs was released, I probably have it running eight to ten hours a day.
00:23:44.880 --> 00:23:49.359
All right, I'm I'm really getting my money's worth on the uh on the uh pro plan there.
00:23:49.440 --> 00:23:55.039
And I know they've had like double usage, and uh I think that might be going away sometime soon, or maybe it went away sometime soon.
00:23:55.119 --> 00:23:56.640
That's why I'm hitting my limits, right?
00:23:56.720 --> 00:24:04.000
So OpenAI launched Claude, their Codex desktop app in February, and this is more of a multi-agent command center.
00:24:04.160 --> 00:24:11.759
And the great thing here, they've had kind of skills, so uh Anthropics popularized uh skills uh kind of protocol.
00:24:11.839 --> 00:24:16.880
So they've had skills support, automation support, so you can schedule things, right?
00:24:17.039 --> 00:24:31.359
So if you wanted to, right, you could schedule codecs to you know go on a certain website, grab some information, you know, create a Word document on your computer, you know, organize your computer every day, right?
00:24:31.440 --> 00:24:40.880
So any any task that you could do in your local terminal, uh on your local computer with your files and folders and your browser, codecs can do and it can schedule it, right?
00:24:40.960 --> 00:24:52.319
So I think this is um, and also important to note, uh Claude Cowork uh just literally hours ago, uh, added schedule task support as well.
00:24:52.480 --> 00:24:54.960
So I think you know that's something that's kind of like low-key.
00:24:55.039 --> 00:25:02.640
I don't think people are using, I don't think for the most part, codex people are using for non-technical work, and that's the majority of what I'm using it for.
00:25:02.799 --> 00:25:11.359
Um, so that space, the desktop AI agent space that can control your browser, access all your local files, uh, upload, download.
00:25:11.440 --> 00:25:12.400
I mean, it's huge, right?
00:25:12.480 --> 00:25:17.599
And you can't talk about this space, the desktop AI agent, without talking about OpenClaw, right?
00:25:17.759 --> 00:25:21.759
So OpenClaw technically can run uh in a virtual environment.
00:25:21.920 --> 00:25:31.039
So you could say it's a phase four, an AI coworker, but many people are buying uh OpenClaw their own uh or its own uh computer, right?
00:25:31.200 --> 00:25:32.160
And that's all that happens.
00:25:32.240 --> 00:25:34.000
This is OpenClaw's computer, right?
00:25:34.079 --> 00:25:40.000
And this is where uh you know it has its own phone number and its own uh you know email, right?
00:25:40.079 --> 00:25:42.640
But that's where we're at now, right?
00:25:43.119 --> 00:25:52.000
This is the journey from an AI chatbot that hallucinated and was just kind of a fun party trick and no businesses would touch.
00:25:52.319 --> 00:25:59.839
Now, literally, you have companies buying dedicated computers for multiple.
00:26:00.240 --> 00:26:01.680
I've seen stories of this, right?
00:26:01.839 --> 00:26:10.160
People buying multiple dedicated computers for every single employee so they can have you know desktop AI agents running more and more.
00:26:10.960 --> 00:26:14.559
So the hidden kind of phase that impacts it all, right?
00:26:14.640 --> 00:26:17.440
So there's our five phases, but the hidden one is scheduling.
00:26:17.519 --> 00:26:17.920
Uh right.
00:26:18.000 --> 00:26:23.200
And we've seen, like I've already said, um, you know, uh Claude Cowork just added this.
00:26:23.359 --> 00:26:25.599
Microsoft Copilot just added this.
00:26:25.839 --> 00:26:28.720
Codex has always had this since it was announced.
00:26:28.880 --> 00:26:35.279
But I mean, OpenAI released scheduling via task uh more than uh more than a year ago.
00:26:35.440 --> 00:26:40.319
Although, like I said, unfortunately, it's really just uh in ancient mode now.
00:26:40.480 --> 00:26:46.640
Google has it, which I don't know why more people aren't using scheduled actions in uh Google Gemini, it's amazing.
00:26:46.720 --> 00:26:51.519
Uh Perplexity and Grok uh rolled out tasks uh you know about seven, eight months ago.
00:26:51.680 --> 00:26:53.839
Uh OpenAI Pulse, which I'm not a fan of.
00:26:54.000 --> 00:26:56.400
If if they made Pulse better, I think it would be great, right?
00:26:56.559 --> 00:27:02.400
But this is kind of the hidden layer that infiltrates now, you know, phases two through five, right?
00:27:02.640 --> 00:27:05.920
The ability now uh to schedule, right?
00:27:06.000 --> 00:27:10.240
Whether it's a scheduling a desktop agent, which is crazy, right?
00:27:10.400 --> 00:27:15.599
That's why I leave my computer on now all the time, because you know, all of a sudden, oh, it's 2 a.m.
00:27:15.839 --> 00:27:19.039
You know, Codecs is gonna go do a three-hour task for me.
00:27:19.119 --> 00:27:19.680
Uh, right.
00:27:19.839 --> 00:27:34.640
So this is if you aren't paying attention to scheduling, whether it's in the business chat context, whether it's the remote uh you know AI virtual coworker, or whether it's phase five, you have to pay attention to it.
00:27:35.599 --> 00:27:36.079
All right.
00:27:36.480 --> 00:27:39.359
Now let's just go to let's wrap it up here.
00:27:39.519 --> 00:27:42.400
Uh, I told you I'm gonna give you what's next.
00:27:42.559 --> 00:27:46.960
And here's kind of my five facts and strategies for what's next.
00:27:47.200 --> 00:27:47.680
All right.
00:27:47.920 --> 00:27:56.799
Um number one, agents are delivering real artifacts, decks, docs, spreadsheets, everything better than humans.
00:27:56.960 --> 00:28:05.519
I think this is probably what number one of the most overlooked facts and aspects of large language models.
00:28:05.599 --> 00:28:07.039
And this is models by default.
00:28:07.200 --> 00:28:21.599
People don't understand that you can literally, if you know what you're doing, if you give uh you know, Claude or uh, I'd say right now, probably uh Anthropic and um OpenAI are the leaders in this, at least on the you know, business chatbot space.
00:28:21.839 --> 00:28:31.200
You can literally go do your research, uh, contextualize and personalize through your business context and create spreadsheets and uh docs, decks, etc.
00:28:31.440 --> 00:28:31.759
All right.
00:28:31.839 --> 00:28:40.079
Number two, desktop AI creates a security and governance challenge that most IT teams aren't remotely ready for.
00:28:40.160 --> 00:28:43.680
So be ready for agent drift and agent crash.
00:28:43.759 --> 00:28:46.640
That is gonna be a huge trend of 2026 and 2027.
00:28:46.799 --> 00:29:01.599
All right, the next piece of advice here tomorrow's AI winners are gonna come care less about what actual AI systems they're using and they're gonna care more about in spending more time rebuilding how knowledge work works, right?
00:29:01.839 --> 00:29:04.960
I think you know, oh, today's best model, right?
00:29:05.039 --> 00:29:06.480
It becomes a commodity, right?
00:29:06.559 --> 00:29:08.960
They're all getting similar or the same features.
00:29:09.119 --> 00:29:12.720
Yes, I still think there's always gonna be winners and ones that are slightly better.
00:29:12.880 --> 00:29:18.079
But if your team is spending more time on deciding, oh, are we gonna use Gemini or OpenAI?
00:29:18.240 --> 00:29:21.440
Oh, are we gonna use, you know, Claude or Google, right?
00:29:21.519 --> 00:29:28.079
At that point, if you're spending more time doing that than rebuilding how knowledge work works, you're behind.
00:29:28.319 --> 00:29:28.640
All right.
00:29:28.799 --> 00:29:33.599
Number four, the AI assistant era is already over, right?
00:29:33.920 --> 00:29:39.839
The AI worker era is here, both virtually uh and uh on the desktop.
00:29:39.920 --> 00:29:44.000
And most companies haven't noticed because I think it's been Swift, right?
00:29:44.079 --> 00:29:45.920
A lot of these other rollouts, right?
00:29:46.000 --> 00:29:52.559
Even if you look at the uh, you know, the business chat context, you could say that took 18 months, right?
00:29:52.799 --> 00:30:00.480
The uh AI worker and going from AI assistant to AI worker has taken 18 days.
00:30:00.640 --> 00:30:02.160
It is fast, it is here.
00:30:02.240 --> 00:30:03.759
Most companies haven't noticed.
00:30:03.920 --> 00:30:06.160
If you're listening to me, you need to pay attention.
00:30:06.319 --> 00:30:15.519
All right, and then last but not least, scheduled or proactive AI is the sleeper feature that changes the entire relationship between humans and AI, right?
00:30:15.599 --> 00:30:18.960
And it's still one of those things that's flying under the radar for now.
00:30:19.119 --> 00:30:26.640
So if you, your company, uh, your department wants to take advantage, that's where you should be spending your time on right now.
00:30:26.799 --> 00:30:28.720
All right, I hope this one was helpful.
00:30:28.880 --> 00:30:36.720
A quick journey and recap of how we got from AI chatbots to now we have autonomous AI coworkers.
00:30:36.880 --> 00:30:40.640
So if this was helpful, please go to startheseries.com.
00:30:40.799 --> 00:30:43.759
That's going to give you free access to our inner circle community.
00:30:43.839 --> 00:30:52.480
And it's gonna put you right into our Start Here Series channel, where you can listen to now all 10 volumes of our Start Here series.
00:30:52.640 --> 00:30:54.160
Uh so thank you for tuning in.
00:30:54.240 --> 00:30:57.279
I hope to see you back tomorrow and every day for more everyday AI.
00:30:57.359 --> 00:30:58.400
Thanks, y'all.
00:00:01.280 --> 00:00:06.719
Three and a half years ago, AI was kind of like a party trick.
00:00:06.959 --> 00:00:19.760
I mean, if your entire company or department was using AI in late 2022, people probably would have thought that you were crazy or you could have easily gotten into some hot water.
00:00:20.239 --> 00:00:22.160
Oh, how the times have changed.
00:00:22.399 --> 00:00:28.239
I mean, now if your entire department isn't using AI, you're definitely in hot water.
00:00:28.399 --> 00:00:34.799
That's because there's a good chance that you, your company or your department might be sinking if you're not.
00:00:34.960 --> 00:00:43.119
That's because the rise in prominence and capabilities of AI has been astronomical over the past three years.
00:00:43.200 --> 00:00:53.600
And we've gone from AI chatbots that were kind of fun to play around with to straight up autonomous coworkers that are doing the work for enterprises.
00:00:53.840 --> 00:01:10.799
So that's what we're going to be tackling on today's show, volume 10 of our Start Here series, giving you a quick overview of how AI has gone from a chatbot to an autonomous work coworker and how consumer AI has changed.
00:01:11.040 --> 00:01:12.799
And what's next?
00:01:13.359 --> 00:01:13.920
All right.
00:01:14.239 --> 00:01:17.120
And if you are brand new here, awesome.
00:01:17.280 --> 00:01:18.480
Great time to jump in.
00:01:18.640 --> 00:01:21.760
Uh, this is part of our start here series.
00:01:21.920 --> 00:01:29.200
This is our essential podcast series to both learn the AI basics and to double down on your AI knowledge.
00:01:29.359 --> 00:01:33.519
So, like I said, we are on volume 10 of our Start Here series.
00:01:33.680 --> 00:01:43.200
And if you want to go listen to the entire uh series in order, uh, I actually just dropped a Spotify playlist so you can do just that super easy.
00:01:43.280 --> 00:01:46.640
So make sure you go to starthearseries.com.
00:01:46.799 --> 00:01:51.200
So that's also going to give you free access to our inner circle community.
00:01:51.280 --> 00:01:57.120
So you can go learn and network with other business leaders like yourselves that are learning AI.
00:01:57.200 --> 00:02:01.359
Go listen to every single uh Start Here series, read about it.
00:02:01.439 --> 00:02:03.519
It's all in a dedicated space there.
00:02:03.760 --> 00:02:07.840
And you'll get access to our free Prime Prompt Polish course.
00:02:08.000 --> 00:02:10.639
And hey, talking about starting here, right?
00:02:10.800 --> 00:02:15.360
Because after 700 plus episodes, I didn't have an answer for what people said.
00:02:15.520 --> 00:02:16.400
Where should I start?
00:02:16.560 --> 00:02:18.719
When listening to your podcast, I was like, I don't know.
00:02:18.800 --> 00:02:20.960
Well, listen to the Start Here series in order.
00:02:21.039 --> 00:02:25.680
But along the way, make sure you go listen to episodes 713 and 712.
00:02:25.759 --> 00:02:29.199
That is our 2026 AI prediction and roadmap series.
00:02:29.360 --> 00:02:29.840
All right.
00:02:30.080 --> 00:02:40.240
Last episode in Start Here, uh, we talked about agent risk security and AI sprawl in 2026, why AI that acts changes everything.
00:02:40.319 --> 00:02:44.879
So make sure you go listen to that, the last uh episode of our Start Here series.
00:02:44.960 --> 00:02:50.319
And now let's jump into today's from AI Chatbot to autonomous coworkers.
00:02:50.479 --> 00:02:59.360
So on today's show, all right, I'm gonna go over the five phases of AI enabled work and why most teams are still stuck in phase one or two.
00:02:59.439 --> 00:03:07.120
Uh, I'm gonna tell you how and why every major AI company is suddenly racing to put agents on your desktop.
00:03:07.280 --> 00:03:15.039
Uh, you're gonna learn about the hidden layer that almost no one is talking about that changes how AI works for you or your company entirely.
00:03:15.199 --> 00:03:19.680
And then we're gonna end with my five actionable takeaways for what's next.
00:03:19.759 --> 00:03:28.000
Because yeah, it's been a fast sprint from AI chatbot to autonomous coworker, and we are not done, obviously.
00:03:28.319 --> 00:03:35.199
So let's first start with a zoomed out overview of what the heck has happened.
00:03:35.360 --> 00:03:38.479
Uh, so large language models are not new, right?
00:03:38.560 --> 00:03:46.639
They've been around for a long time, uh, technically well before even the uh Chat GPT moment of November 2022.
00:03:46.719 --> 00:03:48.719
But let's start there, shall we?
00:03:48.960 --> 00:03:58.719
Okay, so that's kind of when the whole right generative AI and large language model phase in the corporate world kicked off was in November 2022, right?
00:03:58.879 --> 00:04:06.960
Uh many people, including NVIDIA um CEO Jensen Wong, call that uh the the line in the sand, right?
00:04:07.520 --> 00:04:10.719
So November 2022, ChatGPT launches.
00:04:10.879 --> 00:04:14.560
Uh, you know, it's a simple QA chatbot in a browser tab.
00:04:14.719 --> 00:04:15.199
All right.
00:04:15.439 --> 00:04:19.839
Then let's fast forward to where we are today, February 2026.
00:04:20.000 --> 00:04:27.360
I mean, we have AI agents that can run recurring tasks on your desk desktop with local file access, right?
00:04:27.600 --> 00:04:34.319
So at that point, they can use your browser, they can move files, they can upload, they can download, right?
00:04:34.720 --> 00:04:38.639
At that point, they are an actual autonomous AI coworker.
00:04:38.800 --> 00:04:40.800
So that's where we started, right?
00:04:40.879 --> 00:04:45.279
This is like one of those uh social media posts, like how it started, right?
00:04:45.360 --> 00:04:49.680
Uh in AI chatbot that was a party trick that didn't really know anything.
00:04:49.839 --> 00:04:52.959
To okay, this thing runs just like a human would.
00:04:53.120 --> 00:05:02.399
It can work on your desktop, it can use your files, it can move your files, upload, download, use your browser, uh, right, access all your data.
00:05:02.560 --> 00:05:04.240
Oh, how the times have changed.
00:05:04.319 --> 00:05:07.600
Uh, in even just the last few days and weeks, right?
00:05:07.839 --> 00:05:13.680
Uh again, you may be listening to this in late 2026 or early 2027.
00:05:13.759 --> 00:05:14.240
I don't know.
00:05:14.399 --> 00:05:26.720
But if you're listening to it in February, uh, right, today, February 27th, I mean, the last 48 hours alone, we've seen a ton of movement that I think signal where this is headed, right?
00:05:26.879 --> 00:05:32.319
We got perplexity computer, uh, co-pilot scheduled tasks from Microsoft.
00:05:32.399 --> 00:05:35.600
Uh, we got remote control and scheduled tasks from Anthropic.
00:05:35.680 --> 00:05:50.399
So, you know, three of the big five players there uh in Perplexity, Microsoft, and Anthropic just released pretty big updates that signal toward this more scheduled autonomous desktop computer worker.
00:05:51.040 --> 00:05:51.600
All right.
00:05:52.959 --> 00:05:55.199
So here's the five phases.
00:05:55.680 --> 00:05:58.879
All right, so uh phase one, online chat bots, right?
00:05:58.959 --> 00:06:01.600
They answer your questions in a browser tab.
00:06:01.759 --> 00:06:03.839
Uh phase two, business chat.
00:06:04.000 --> 00:06:09.920
This is where you know answers started to come with your company's data and using apps.
00:06:10.079 --> 00:06:13.519
Uh phase three, we have the agentic era, right?
00:06:13.600 --> 00:06:19.040
So this is both agenc models and multi-step execution toward an outcome, right?
00:06:19.120 --> 00:06:21.600
So more outcome-based with multiple steps.
00:06:21.759 --> 00:06:24.480
Phase four, this is your AI coworkers, right?
00:06:24.639 --> 00:06:27.120
Delegate a goal and get finished work.
00:06:27.360 --> 00:06:33.040
And then phase five, desktop AI, agents that work with your local files, right?
00:06:33.360 --> 00:06:42.079
And these are kind of in order, but as you'll see, as we talk about each phase a little bit more, technically there's some overlap, right?
00:06:42.240 --> 00:06:47.279
Uh, we're still obviously getting updates to phase two and phase five at the time, right?
00:06:47.360 --> 00:06:51.040
It doesn't mean phase two is over just because phase five has started, right?
00:06:51.120 --> 00:06:51.680
That's not where it is.
00:06:51.759 --> 00:06:56.000
It's it's more of kind of the uh official eras of when things started.
00:06:56.079 --> 00:07:16.480
And I think, you know, for the most part, I'd say phase one is probably phased out, although unfortunately, I still think people use ChatGPT just as an online chatbot to get answers without, you know, using any of the agentic capabilities, without using any of the business capabilities, without using any of the more coworker type capabilities.
00:07:16.639 --> 00:07:22.800
So I think for the most part, people have kind of uh phased out of that era, but not completely.
00:07:23.279 --> 00:07:35.920
And then there's the feature or the layer on top of these five phases that I think might be even more important in really changing how works get how work gets done.
00:07:36.079 --> 00:07:38.000
And that's scheduling, right?
00:07:38.079 --> 00:07:46.879
So I just mentioned that in the last 48 hours, we've seen literally three big updates just in this space.
00:07:47.199 --> 00:07:52.319
But this does live across all phases, so it's not its own phase necessarily, but right.
00:07:52.399 --> 00:07:56.879
Uh, and Google and OpenAI have had this for a pretty long time, uh, right?
00:07:56.959 --> 00:08:02.399
It's about six to 12 months, a little bit longer, but it's rarely talked about until recently, right?
00:08:02.560 --> 00:08:13.279
If you go back to my uh original Chat GPT tasks show, uh and it's a huge bummer that open AI changed how tasks work in uh in uh chat GPT, it's actually weird.
00:08:13.360 --> 00:08:18.639
If you're on the pro plan, you don't even have scheduled tasks anymore, which is one of the things I use most, right?
00:08:18.720 --> 00:08:19.759
Now it's just pulse.
00:08:20.000 --> 00:08:23.680
Um, in scheduled tasks, you can only really use in agent mode, anyways.
00:08:23.839 --> 00:08:26.800
Uh, it's it's rarely been talked about until recently.
00:08:26.879 --> 00:08:34.639
That's because of all the changes in the last 48 hours with Perplexity, Anthropic, and Microsoft coming in with some new offerings there.
00:08:35.200 --> 00:08:41.840
All right, well, now you have the answers, but if you want to dig in a little bit deeper, now's where we're gonna do it.
00:08:41.919 --> 00:08:43.919
So let's talk a little bit about phase one.
00:08:44.080 --> 00:08:51.919
So that started with Chat GPT launching in November of 2022, and it was obviously mega viral, right?
00:08:52.000 --> 00:08:53.440
I remember at the time, right?
00:08:53.519 --> 00:08:58.240
So my company, we had been using the GPT technology since it came out in 2020.
00:08:58.320 --> 00:09:00.080
So a lot of people overlooked that.
00:09:00.240 --> 00:09:08.000
Before there was Chat GPT, OpenAI put their product out there, uh GPT 3, uh, in other pieces of software.
00:09:08.159 --> 00:09:13.600
So when ChatGPT came out, at least for me, and I used it in November, I'm like, this stinks, right?
00:09:13.679 --> 00:09:15.440
It's not wasn't it was terrible.
00:09:15.600 --> 00:09:21.840
It was worse than using the other platforms that had been out for two years, obviously, because they had uh the time to improve their platforms.
00:09:22.000 --> 00:09:27.679
But regardless, this is what put today's AI and large language models on the map.
00:09:27.759 --> 00:09:30.559
And it did technically, I think, change the course.
00:09:30.639 --> 00:09:33.039
And this might sound corny, but it's true.
00:09:33.200 --> 00:09:37.519
I think it's gonna change the course and has changed the course of business history.
00:09:37.759 --> 00:09:40.639
All right, then uh you had a little bit later, right?
00:09:40.720 --> 00:09:48.240
About three, four months later, you had Google launching Bard in March, and then also Anthropic launching Claude in March.
00:09:48.320 --> 00:09:54.159
Although uh Claude was not uh generally available outside of kind of an enterprise pilot, I believe, until that summer.
00:09:54.399 --> 00:09:57.360
But at the time, this is just online chatbot era, right?
00:09:57.519 --> 00:09:58.480
Ask a question.
00:09:58.639 --> 00:10:08.000
So a couple huge reasons why in uh this first phase, it this wasn't suited for the business world, right?
00:10:08.080 --> 00:10:13.200
Uh, one of the reasons large language models at the time did not have access to the internet.
00:10:13.519 --> 00:10:13.840
Okay.
00:10:14.080 --> 00:10:15.919
And that made them extremely dangerous.
00:10:16.080 --> 00:10:29.039
And I think I depending on how you look at it, fortunately or unfortunately, it was really this phase one that I think is still kind of setting the tone for how a lot of people think about AI today, right?
00:10:29.200 --> 00:10:35.759
That's because hallucinations were rampant in phase one, because number one, uh, well, three reasons, I think.
00:10:35.840 --> 00:10:37.120
Number one, you couldn't upload data.
00:10:37.279 --> 00:10:40.320
Number two, it couldn't access the internet and models couldn't reason.
00:10:40.559 --> 00:10:52.799
So, unless you were really good at prompt engineering, there's a good chance that if you were trying to use a large language model in 2022 or early 2023 for any business context, it was probably pretty bad.
00:10:52.960 --> 00:10:53.279
All right.
00:10:53.360 --> 00:10:57.679
And that's why in 2022 and 2023, RAG was all the rage, right?
00:10:57.840 --> 00:11:11.279
Retrieval logmin and generation, because if you wanted trustworthy outputs from AI, you had to build your own rag pipeline because the outputs were absolutely terrible, just riddled with hallucinations because, well, the models were not capable.
00:11:11.360 --> 00:11:13.679
It was at the time, I think, more than a party trick.
00:11:13.759 --> 00:11:17.440
So no file access, no integrations, no real world actions, right?
00:11:17.600 --> 00:11:20.879
And the models were pulling on very old data, right?
00:11:21.039 --> 00:11:24.799
At the time, those first models that came out, right?
00:11:25.279 --> 00:11:29.120
Now the people training the models are legit superstars, right?
00:11:29.200 --> 00:11:34.240
Some of your heads of research are making tens of millions or hundreds of millions of dollars.
00:11:34.399 --> 00:11:34.960
Uh, right.
00:11:35.120 --> 00:11:37.279
At the time, I don't think that was the case, right?
00:11:37.360 --> 00:11:50.159
Obviously, they're very smart researchers and seasoned researchers, but uh, you know, just the emphasis that has been placed on getting good, clean, uh, reliable training data uh and just the work that can be done at inference has completely changed.
00:11:50.240 --> 00:11:51.840
All right, so that was phase one.
00:11:52.480 --> 00:12:04.559
Phase two, uh, I'd say probably started in like 2023, and it kind of, I won't say again, it didn't end in 2025, but I'll give this phase or this era 2023 to 2025.
00:12:04.720 --> 00:12:06.240
So this is team AI.
00:12:06.399 --> 00:12:07.519
This is business chat.
00:12:07.600 --> 00:12:09.440
This is when it was born, uh, right.
00:12:09.519 --> 00:12:12.159
So team AI that's connected to your company's data.
00:12:12.240 --> 00:12:17.679
Uh, so first, uh, I think technically first was Microsoft Copilot in late 2023.
00:12:17.759 --> 00:12:22.799
Um, and then uh also you had ChatGPT Enterprise in August 2023.
00:12:22.879 --> 00:12:31.679
Uh, and then you had uh Gemini for Workspace in February 24, and also Anthropic for Teams and Enterprises came out in 2024 as well.
00:12:32.240 --> 00:12:35.519
But it wasn't just going team first, right?
00:12:35.600 --> 00:12:36.960
That was a big first step.
00:12:37.120 --> 00:12:47.120
So, along with that, companies found out, right, or big AI labs, they're like, okay, this is great to get, you know, a hundred million consumers to check out your AI chatbot.
00:12:47.279 --> 00:12:50.240
But if we want to make this a tool that companies are gonna buy, right?
00:12:50.320 --> 00:12:55.600
Because at the time, I think it was Microsoft that was cleaning house with enterprise AI, uh, right.
00:12:55.679 --> 00:12:59.679
Because when they came out in 2023, uh, there weren't a lot of other options.
00:12:59.840 --> 00:13:04.080
If you wanted to get 10,000 people uh to use a product, right?
00:13:04.399 --> 00:13:08.240
You weren't gonna use a consumer chat bot that didn't have data.
00:13:08.480 --> 00:13:11.360
So this is really the the maturation phase.
00:13:11.440 --> 00:13:25.440
Uh, and I think things like Chat GPT connectors, cloud connectors, uh Google Apps, uh, right, all these things in 2025 uh really I think help solidify the space and also cut down on hallucinations.
00:13:25.519 --> 00:13:34.559
Uh and this is where AI could finally access your email, docs, and drives all in a dynamic way, but it really couldn't act on it, right?
00:13:34.720 --> 00:13:37.120
At least in the earlier parts of 2025.
00:13:37.200 --> 00:13:38.720
That's kind of in the next phase.
00:13:38.799 --> 00:13:46.159
But this is the beginning of uh, you know, AI going from fun party trick to okay, this could work, right?
00:13:46.240 --> 00:13:54.080
Uh kind of the experimentation phase of of 2023 and 2024 to I think by 2025, the business world knew.
00:13:54.240 --> 00:14:08.159
I would say I would I was hoping or assuming it would be in 2024, but I don't think it was maybe until early 2025, maybe late 2024, that everyone realized, oh yeah, the future of work is 100% AI and there's no way around it.
00:14:08.240 --> 00:14:17.360
I think you still had fence sitters uh in 2024 at the enterprise level, which is absolutely nutty to me because now I think those fence sitters are paying for it.
00:14:17.440 --> 00:14:20.879
Uh, but uh phase two really pushed it.
00:14:21.120 --> 00:14:28.159
Phase three, I'll say started in 2024, kind of through 2025, although it's still technically going on, right?
00:14:28.240 --> 00:14:29.279
And it's ever improving.
00:14:29.360 --> 00:14:31.200
But this is the agentic area.
00:14:31.360 --> 00:14:34.879
This is where you have AI that plans and executes, right?
00:14:35.120 --> 00:14:41.519
I think technically you could say 2023 because Microsoft had Copilot Studio at the very end of 2023.
00:14:41.600 --> 00:14:49.200
But I'll say 2024, because I think nothing was really uh, you know, um adapted at March until 2024.
00:14:49.279 --> 00:14:55.679
But this is when you had things like uh OpenAI launched launched Operator in January 2025, right?
00:14:55.759 --> 00:15:00.000
That was a big moment, although Operator was fairly bad and slow.
00:15:00.159 --> 00:15:06.480
Uh, but it it signaled a big shift toward AI that could think and do human work, right?
00:15:06.559 --> 00:15:11.519
The big thing was being able to access a browser, which is what operator could do.
00:15:12.080 --> 00:15:16.720
You know, it could uh it was an agent that could uh navigate the web autonomously.
00:15:16.879 --> 00:15:23.600
Uh, then I think you had this uh right thereafter, uh shortly thereafter, you have this push for agenc browsers, right?
00:15:23.759 --> 00:15:29.360
So different than an agent, it is a browser powered by a reasoning model.
00:15:29.519 --> 00:15:34.799
So, you know, an agentic browser, one of its main features is to agentically browse the web.
00:15:34.960 --> 00:15:44.960
So I think some tools like ChatGPT Atlas, uh Perplexity Comet, uh Google, you know, announced their uh agent mode, Project Mariner.
00:15:45.279 --> 00:15:47.279
Uh, they had their Gemini in Chrome.
00:15:47.519 --> 00:15:50.559
Uh, that was probably more uh, you know, late 2025.
00:15:50.720 --> 00:15:55.759
Uh, but the key shift here thing, uh I think was reasoning models, right?
00:15:55.840 --> 00:15:57.360
They powered all of this.
00:15:57.519 --> 00:16:04.720
You know, you don't have um worthwhile AI agents without reasoning models, all right.
00:16:04.799 --> 00:16:10.960
Models that can think, plan, uh use logic like humans can, all right.
00:16:11.120 --> 00:16:24.720
Um, and that's really what I think separates, you know, at least in in my opinion, it was the agencera um that I think is ultimately going to be the most impactful.
00:16:24.799 --> 00:16:26.399
It wasn't Chat GPT, right?
00:16:26.559 --> 00:16:34.159
Because if we never kind of had models that could reason or which that would mean we probably would never have AI agents, right?
00:16:34.320 --> 00:16:38.559
We would have more uh human duct tape than agentic output.
00:16:38.639 --> 00:16:49.440
So it wouldn't have been worth it for any businesses uh to truly invest in AI over the long term if we never entered into the agentic era or reasoning models, right?
00:16:49.600 --> 00:16:59.120
Um, so some of the models maybe you've heard of these, they're a little older by now, but you know, 01 preview really kicked it all off in September of 2024.
00:16:59.279 --> 00:17:10.640
Uh, then you had Gemini 2.0 Flash in December 24, and then Claude Sonnet37, I think was their first uh reasoning or hybrid model from Anthropic in February 2025.
00:17:10.880 --> 00:17:22.240
So a pretty big like eight-month period there where we kind of left the quote unquote old uh versions, you you know, the non-thinking transformer models, right?
00:17:22.319 --> 00:17:25.279
Even though they're still transformer models, they just think and reason, right?
00:17:25.359 --> 00:17:38.559
But I like I really say that's like the old school AI versus the new school AI, because I think what um agentic models can do and their capabilities, the scaffolding, the harnessing that continues to be improved, right?
00:17:38.720 --> 00:17:44.079
You can make the argument today, and I've talked with very smart people about this, right?
00:17:44.160 --> 00:17:51.039
Like the head of Microsoft Research that's been working in agents for 20 years, uh, the head of agents at Cloudflare, right?
00:17:51.119 --> 00:18:07.839
I've had so many conversations with extremely smart people in the space that have eventually essentially agreed that, yeah, if you're using, you know, uh GPT-5.2 Pro and you have all your business data connected to it, that's an agent, right?
00:18:08.079 --> 00:18:11.119
Especially when you can schedule it and it can act autonomously.
00:18:11.200 --> 00:18:12.640
It's like, yeah, that's an agent, right?
00:18:12.720 --> 00:18:19.839
Or if you're using, you know, Gemini 3.1 Pro and scheduling things and it has access to your data, that's an agent.
00:18:20.319 --> 00:18:20.640
Right.
00:18:20.799 --> 00:18:23.759
So uh, but it really started with the reasoning models.
00:18:23.839 --> 00:18:25.279
Uh, then we have phase four.
00:18:25.359 --> 00:18:26.960
This is AI coworkers.
00:18:27.039 --> 00:18:32.799
Uh, so where you delegate a goal and it just finishes work on its own, right?
00:18:32.960 --> 00:18:37.039
And yes, I know and understand that some of these phases start to blend together.
00:18:37.200 --> 00:18:38.240
I get it, right?
00:18:38.319 --> 00:18:44.880
But in my opinion, the big difference between the agentic era, I would say that's more model and browser based, right?
00:18:44.960 --> 00:18:52.720
And that's kind of the um the foundation or the stepping stones for today's AI coworkers or agent coworkers.
00:18:53.200 --> 00:18:59.599
And this is more or less general AI agents that have a virtual computer and can access your data.
00:18:59.759 --> 00:19:06.000
So, right, you can say that's you know, Chat GPT agent mode, uh, you know, Google Gemini agent mode, obviously.
00:19:06.079 --> 00:19:09.599
Uh, but I'd say things like Manaus, right?
00:19:09.839 --> 00:19:16.240
Huge uh successful launch uh for Manaus, uh, you know, recently uh acquired by Meta.
00:19:16.559 --> 00:19:18.799
Uh same thing with uh Gen Spark, right?
00:19:18.880 --> 00:19:24.559
Uh Gen Spark, uh another kind of general purpose agent that can browse the web, uh, right?
00:19:24.720 --> 00:19:28.079
It can access all your data, but for the most part, you know, scheduling.
00:19:28.319 --> 00:19:33.440
A lot of these AI coworkers are more browser-based, right?
00:19:33.599 --> 00:19:39.839
Uh, but you can give access to your data and then they can go use the web for the most part.
00:19:40.000 --> 00:19:43.359
They can, you can log, even though it can be risky, uh, right.
00:19:43.440 --> 00:19:47.440
But you know, they have a virtual computer, a sandbox, a terminal, all these things.
00:19:47.519 --> 00:19:54.160
Uh, you, you know, they can run code, you know, a lot of them can run different models, uh, you know, use uh sub agents, right?
00:19:54.240 --> 00:20:01.839
So uh this is kind of the uh the version of a virtual AI teammate that works more in the cloud.
00:20:02.079 --> 00:20:12.079
And then we also, the newest entry here, uh, you know, perplexity computer, uh, that was just launched, uh, that uses 19 uh AI models.
00:20:12.240 --> 00:20:12.400
Yeah.
00:20:12.480 --> 00:20:14.960
Let me know if we should do a show on perplexity computer.
00:20:15.039 --> 00:20:16.160
I've been thinking about it.
00:20:16.400 --> 00:20:16.799
All right.
00:20:16.960 --> 00:20:19.279
And then that leads us to phase five.
00:20:19.440 --> 00:20:22.160
And this is kind of where we're at now.
00:20:22.240 --> 00:20:25.200
Although, yes, we're still in phase four with AI coworkers.
00:20:25.279 --> 00:20:26.240
That's not going anywhere.
00:20:26.319 --> 00:20:29.680
We're still in the agentic area uh era, that's not going anywhere.
00:20:29.759 --> 00:20:32.000
But I think where we are today, right?
00:20:32.240 --> 00:20:37.839
The era that is maybe most recently started, that is desktop AI agents, right?
00:20:38.000 --> 00:20:50.880
So this is a little different in a step, uh, both in a um technically a more opportunistic direction, but also in a much more dangerous direction as well, right?
00:20:51.119 --> 00:21:05.599
Um, it is probably in most cases, and I think most experts would agree, somewhat safer uh to use a cloud uh-based agent that maybe can't access your local files, right?
00:21:05.759 --> 00:21:15.200
Even though they can oftentimes, uh if you grant them access, they can access your Gmail and you know your OneDrive, your SharePoint, or right, your Notion, right?
00:21:15.279 --> 00:21:23.839
You can add all these um different uh connectors and apps to the uh kind of virtual AI coworkers.
00:21:23.920 --> 00:21:29.279
But in phase five, this is agents that are running on your actual computer.
00:21:29.440 --> 00:21:34.240
So they can do everything uh that an AI coworker can do, right?
00:21:34.400 --> 00:21:40.799
A virtual uh AI agent, but they can control your actual computer.
00:21:41.039 --> 00:21:41.359
All right.
00:21:41.519 --> 00:21:58.079
So I think uh technically, I you could say, oh, this is more 2026 starting, but technically, I think Claude Code kicked this off in um February 2025, but it really got popular, I would say, in the fourth quarter of 2025, right?
00:21:58.160 --> 00:22:05.200
Uh so Cloud Code was essentially a Desktop coding program, um, that really what it turned into.
00:22:05.359 --> 00:22:16.480
I think some of the popular use cases that really exploded this category was actually people using Claude Code, a terminal tool for developers.
00:22:16.640 --> 00:22:20.960
Well, non-developers started to use it for well, non-technical work.
00:22:21.119 --> 00:22:25.440
And I think Anthropic realized that early on and capitalized on it.
00:22:25.519 --> 00:22:29.279
And that led to the launch of Cowork.
00:22:29.839 --> 00:22:34.160
Claude Co-Work in January 2026, which I absolutely love.
00:22:34.240 --> 00:22:36.240
I'm a big uh Claude Co-Work fan.
00:22:36.319 --> 00:22:42.000
I use Claude Code as well on the desktop version, not in the terminal, not really a terminal by myself.
00:22:42.079 --> 00:22:47.759
Uh, right, but this is now you have in the desktop version of Claude Code and Claude Co-Work.
00:22:48.000 --> 00:22:54.160
This is a desktop AI agent that can control your computer and work for hours, right?
00:22:54.400 --> 00:23:04.000
I'm never one that's intentionally trying to push uh desktop AI agents to go longer and longer, but I got codecs to work for 10 hours months.
00:23:04.079 --> 00:23:04.559
Uh, right.
00:23:04.640 --> 00:23:08.559
I've gotten uh Claude usually is a little faster, sometimes not as thorough.
00:23:08.640 --> 00:23:09.839
I personally prefer Codecs.
00:23:09.920 --> 00:23:11.279
I'm gonna talk about that here in a minute.
00:23:11.440 --> 00:23:21.200
But these are now desktop AI agents that have access to literally everything: your computer, your files, your notes, your your browser, everything.
00:23:21.359 --> 00:23:22.079
Uh right.
00:23:22.319 --> 00:23:31.200
So you can't talk about Claude Code and um uh Claude Cowork without talking about OpenAI's codecs.
00:23:31.279 --> 00:23:31.680
All right.
00:23:31.839 --> 00:23:35.680
And I am crazy bullish on codecs.
00:23:35.839 --> 00:23:36.960
I freaking love it.
00:23:37.039 --> 00:23:44.720
If you've been listening since codecs was released, I probably have it running eight to ten hours a day.
00:23:44.880 --> 00:23:49.359
All right, I'm I'm really getting my money's worth on the uh on the uh pro plan there.
00:23:49.440 --> 00:23:55.039
And I know they've had like double usage, and uh I think that might be going away sometime soon, or maybe it went away sometime soon.
00:23:55.119 --> 00:23:56.640
That's why I'm hitting my limits, right?
00:23:56.720 --> 00:24:04.000
So OpenAI launched Claude, their Codex desktop app in February, and this is more of a multi-agent command center.
00:24:04.160 --> 00:24:11.759
And the great thing here, they've had kind of skills, so uh Anthropics popularized uh skills uh kind of protocol.
00:24:11.839 --> 00:24:16.880
So they've had skills support, automation support, so you can schedule things, right?
00:24:17.039 --> 00:24:31.359
So if you wanted to, right, you could schedule codecs to you know go on a certain website, grab some information, you know, create a Word document on your computer, you know, organize your computer every day, right?
00:24:31.440 --> 00:24:40.880
So any any task that you could do in your local terminal, uh on your local computer with your files and folders and your browser, codecs can do and it can schedule it, right?
00:24:40.960 --> 00:24:52.319
So I think this is um, and also important to note, uh Claude Cowork uh just literally hours ago, uh, added schedule task support as well.
00:24:52.480 --> 00:24:54.960
So I think you know that's something that's kind of like low-key.
00:24:55.039 --> 00:25:02.640
I don't think people are using, I don't think for the most part, codex people are using for non-technical work, and that's the majority of what I'm using it for.
00:25:02.799 --> 00:25:11.359
Um, so that space, the desktop AI agent space that can control your browser, access all your local files, uh, upload, download.
00:25:11.440 --> 00:25:12.400
I mean, it's huge, right?
00:25:12.480 --> 00:25:17.599
And you can't talk about this space, the desktop AI agent, without talking about OpenClaw, right?
00:25:17.759 --> 00:25:21.759
So OpenClaw technically can run uh in a virtual environment.
00:25:21.920 --> 00:25:31.039
So you could say it's a phase four, an AI coworker, but many people are buying uh OpenClaw their own uh or its own uh computer, right?
00:25:31.200 --> 00:25:32.160
And that's all that happens.
00:25:32.240 --> 00:25:34.000
This is OpenClaw's computer, right?
00:25:34.079 --> 00:25:40.000
And this is where uh you know it has its own phone number and its own uh you know email, right?
00:25:40.079 --> 00:25:42.640
But that's where we're at now, right?
00:25:43.119 --> 00:25:52.000
This is the journey from an AI chatbot that hallucinated and was just kind of a fun party trick and no businesses would touch.
00:25:52.319 --> 00:25:59.839
Now, literally, you have companies buying dedicated computers for multiple.
00:26:00.240 --> 00:26:01.680
I've seen stories of this, right?
00:26:01.839 --> 00:26:10.160
People buying multiple dedicated computers for every single employee so they can have you know desktop AI agents running more and more.
00:26:10.960 --> 00:26:14.559
So the hidden kind of phase that impacts it all, right?
00:26:14.640 --> 00:26:17.440
So there's our five phases, but the hidden one is scheduling.
00:26:17.519 --> 00:26:17.920
Uh right.
00:26:18.000 --> 00:26:23.200
And we've seen, like I've already said, um, you know, uh Claude Cowork just added this.
00:26:23.359 --> 00:26:25.599
Microsoft Copilot just added this.
00:26:25.839 --> 00:26:28.720
Codex has always had this since it was announced.
00:26:28.880 --> 00:26:35.279
But I mean, OpenAI released scheduling via task uh more than uh more than a year ago.
00:26:35.440 --> 00:26:40.319
Although, like I said, unfortunately, it's really just uh in ancient mode now.
00:26:40.480 --> 00:26:46.640
Google has it, which I don't know why more people aren't using scheduled actions in uh Google Gemini, it's amazing.
00:26:46.720 --> 00:26:51.519
Uh Perplexity and Grok uh rolled out tasks uh you know about seven, eight months ago.
00:26:51.680 --> 00:26:53.839
Uh OpenAI Pulse, which I'm not a fan of.
00:26:54.000 --> 00:26:56.400
If if they made Pulse better, I think it would be great, right?
00:26:56.559 --> 00:27:02.400
But this is kind of the hidden layer that infiltrates now, you know, phases two through five, right?
00:27:02.640 --> 00:27:05.920
The ability now uh to schedule, right?
00:27:06.000 --> 00:27:10.240
Whether it's a scheduling a desktop agent, which is crazy, right?
00:27:10.400 --> 00:27:15.599
That's why I leave my computer on now all the time, because you know, all of a sudden, oh, it's 2 a.m.
00:27:15.839 --> 00:27:19.039
You know, Codecs is gonna go do a three-hour task for me.
00:27:19.119 --> 00:27:19.680
Uh, right.
00:27:19.839 --> 00:27:34.640
So this is if you aren't paying attention to scheduling, whether it's in the business chat context, whether it's the remote uh you know AI virtual coworker, or whether it's phase five, you have to pay attention to it.
00:27:35.599 --> 00:27:36.079
All right.
00:27:36.480 --> 00:27:39.359
Now let's just go to let's wrap it up here.
00:27:39.519 --> 00:27:42.400
Uh, I told you I'm gonna give you what's next.
00:27:42.559 --> 00:27:46.960
And here's kind of my five facts and strategies for what's next.
00:27:47.200 --> 00:27:47.680
All right.
00:27:47.920 --> 00:27:56.799
Um number one, agents are delivering real artifacts, decks, docs, spreadsheets, everything better than humans.
00:27:56.960 --> 00:28:05.519
I think this is probably what number one of the most overlooked facts and aspects of large language models.
00:28:05.599 --> 00:28:07.039
And this is models by default.
00:28:07.200 --> 00:28:21.599
People don't understand that you can literally, if you know what you're doing, if you give uh you know, Claude or uh, I'd say right now, probably uh Anthropic and um OpenAI are the leaders in this, at least on the you know, business chatbot space.
00:28:21.839 --> 00:28:31.200
You can literally go do your research, uh, contextualize and personalize through your business context and create spreadsheets and uh docs, decks, etc.
00:28:31.440 --> 00:28:31.759
All right.
00:28:31.839 --> 00:28:40.079
Number two, desktop AI creates a security and governance challenge that most IT teams aren't remotely ready for.
00:28:40.160 --> 00:28:43.680
So be ready for agent drift and agent crash.
00:28:43.759 --> 00:28:46.640
That is gonna be a huge trend of 2026 and 2027.
00:28:46.799 --> 00:29:01.599
All right, the next piece of advice here tomorrow's AI winners are gonna come care less about what actual AI systems they're using and they're gonna care more about in spending more time rebuilding how knowledge work works, right?
00:29:01.839 --> 00:29:04.960
I think you know, oh, today's best model, right?
00:29:05.039 --> 00:29:06.480
It becomes a commodity, right?
00:29:06.559 --> 00:29:08.960
They're all getting similar or the same features.
00:29:09.119 --> 00:29:12.720
Yes, I still think there's always gonna be winners and ones that are slightly better.
00:29:12.880 --> 00:29:18.079
But if your team is spending more time on deciding, oh, are we gonna use Gemini or OpenAI?
00:29:18.240 --> 00:29:21.440
Oh, are we gonna use, you know, Claude or Google, right?
00:29:21.519 --> 00:29:28.079
At that point, if you're spending more time doing that than rebuilding how knowledge work works, you're behind.
00:29:28.319 --> 00:29:28.640
All right.
00:29:28.799 --> 00:29:33.599
Number four, the AI assistant era is already over, right?
00:29:33.920 --> 00:29:39.839
The AI worker era is here, both virtually uh and uh on the desktop.
00:29:39.920 --> 00:29:44.000
And most companies haven't noticed because I think it's been Swift, right?
00:29:44.079 --> 00:29:45.920
A lot of these other rollouts, right?
00:29:46.000 --> 00:29:52.559
Even if you look at the uh, you know, the business chat context, you could say that took 18 months, right?
00:29:52.799 --> 00:30:00.480
The uh AI worker and going from AI assistant to AI worker has taken 18 days.
00:30:00.640 --> 00:30:02.160
It is fast, it is here.
00:30:02.240 --> 00:30:03.759
Most companies haven't noticed.
00:30:03.920 --> 00:30:06.160
If you're listening to me, you need to pay attention.
00:30:06.319 --> 00:30:15.519
All right, and then last but not least, scheduled or proactive AI is the sleeper feature that changes the entire relationship between humans and AI, right?
00:30:15.599 --> 00:30:18.960
And it's still one of those things that's flying under the radar for now.
00:30:19.119 --> 00:30:26.640
So if you, your company, uh, your department wants to take advantage, that's where you should be spending your time on right now.
00:30:26.799 --> 00:30:28.720
All right, I hope this one was helpful.
00:30:28.880 --> 00:30:36.720
A quick journey and recap of how we got from AI chatbots to now we have autonomous AI coworkers.
00:30:36.880 --> 00:30:40.640
So if this was helpful, please go to startheseries.com.
00:30:40.799 --> 00:30:43.759
That's going to give you free access to our inner circle community.
00:30:43.839 --> 00:30:52.480
And it's gonna put you right into our Start Here Series channel, where you can listen to now all 10 volumes of our Start Here series.
00:30:52.640 --> 00:30:54.160
Uh so thank you for tuning in.
00:30:54.240 --> 00:30:57.279
I hope to see you back tomorrow and every day for more everyday AI.
00:30:57.359 --> 00:30:58.400
Thanks, y'all.
