Season 1 Episode 14: The AI Agent Invasion

AI Lens is your Focused view on the emerging hot topics in the Age of AI!! We provide AI news, hot topics, advancements and discussions about how AI is reshaping business and society.
 
 Today, we're talking about something that's been dominating conversations in tech circles for the past few weeks: the AI agent invasion.
 
 In February 2026, we're witnessing an unprecedented surge in AI agent development and deployment. Companies are racing to build autonomous agents that can perform complex tasks, make decisions, and interact with systems on behalf of humans.
 
 But what does this mean for the market? What opportunities are emerging? And what challenges do we need to be aware of?
 
 Today, we're diving deep into the AI agent phenomenon, exploring the market dynamics, the key players, and what this means for the future of work and business. 

Transcript

SPEAKER_00 0:01

You're listening to AI Lens, your focused view on the emerging hot topics in the age of AI. We provide AI news, hot topics, advancements, and discussions about how AI is reshaping business and society. Today we're diving into one of the most talked-about stories. We're going to look at the latest developments in artificial intelligence. We're going to be talking about something that's been dominating conversations in tech circles for the past few weeks. It's the AI agent invasion.

SPEAKER_01 0:44

In February 2026, we're witnessing an unprecedented surge in AI agent development and deployment. Companies are now racing to build autonomous agents that can perform complex tasks, make decisions, and interact with systems on behalf of humans. But what does that mean for the market? What opportunities are emerging and what challenges do we need to be aware of?

SPEAKER_00 1:09

Today we're diving deep into the AI agent phenomenon, exploring the market dynamics, the key players, and what this means for the future of work and businesses. A lot has happened since our last episode. I know we took a couple of weeks break while we were moving offices. So we are back, and so many things happen so fast in today's world. It's incredible.

SPEAKER_01 1:41

It really is. And we're probably doing a series of multiple podcasts here over the next few weeks per day as we try and catch up on everything that's been happening just even recently. It's just been insane. But with that said, I think it's good to put this into some context. So when AI first was introduced, we all got exposed to Chat GPT, and it was basically prompting, giving us back tremendous information that was far that would just blow us away, right?

SPEAKER_00 2:09

And so it was a lot like having an upscale Google.

SPEAKER_01 2:16

It started to do a little bit more than just that. Then toward the end of 2025, agents really started hitting the ground running. And that's certainly picked up speed and momentum. We saw it at CS and coming out of CS, and certainly now agents are like quite the quite the topic. We also think a big trend in 2026 is going to be recursive self-improvement with these models. But that's a whole different podcast for down the road, right now, today, gonna be talking about AI.

SPEAKER_00 2:44

Well, let's let's for some of the listeners who don't understand what that means, it it it I'm gonna take a step back. I don't think we should just put that off. Basically, what John is saying is that agents are gonna be looking over themselves and other agents, other AI agents, and correcting their coding or correcting the path they're going through.

SPEAKER_01 3:11

Yeah, agents, the models themselves, the coding that's done. You hear about Cloud Code, for instance, and there's others. Open Chat GPT has their codex as well. All of that's gonna start improving on itself. So it'll be able to direct itself, make improvements without with less intervention by humans, the human in the loop is gonna become less significant. Again, we're gonna be more orchestrators and less doers, and that's gonna continue to evolve.

SPEAKER_00 3:37

You know, speaking of that, what's come across like what's really booming, I think, in the on in the AI world over the last couple of weeks are what they're calling lobsters, which is kind of another name for an AI agent. And really what's going on is people are going, they're getting these mini-Macs, they're creating their own agents. And I do it, but I don't need to use a mini-Mac because I'll do it through like Lindy or other programs that can make it safe. And they're creating these AI agents, but you have to be careful, is what I found when you do it. You I tend to have an AI agent do too many tasks. So I think we were talking about this earlier, where you have one agent who kind of manages sub-agents, and you have the subagents do more like individual tasks, and the orchestrator agent kind of manages all of them.

SPEAKER_01 4:45

Aaron Ross Powell Yeah, what happens is these agents have what's called the context window, and that's limited. So if you sit there and have your AI agent doing too much, that context window gets overwhelmed and it starts to break down. So you have this orchestrator who can sit there and DV out specialized tasks to subagents. Those subagents can be more focused, therefore not taking up a huge amount of uh of that context window. The orchestrator then gets feedback, the results back to them, so that subagents' work does not mess up their context window. So it's a way of like dispersing the work so that it's able to have the ability to flow through the system properly.

SPEAKER_00 5:24

And and to give you an idea, when I say too many tasks, I was loading up an agent that was bookkeeper and a CFO and it was doing way too many things. And so I had to break it out. Like if it's reading my bank statements and creating a general ledger, maybe from that point it can do reports. But I need a separate agent who's gonna read the reports and then make recommendations and things like that.

SPEAKER_01 6:00

It's like a traditional corporate structure that you'd see. You'd have agents again specializing in these different tasks. They can actually interact and co-work. It's kind of crazy to watch. And the nice thing is they'll do it 24-7. And there's this element of them getting better over time, too. So you're starting to see that happen as well. But let's back up. Let's let's start with the basics. So, what exactly is an AI agent? An AI agent is going to basically run processes for you. So, another way of looking at it, it's like a software system. It can scan its environment, it can make its own decisions, it can take its own actions in order to achieve specific goals, and it can run 24-7. So, unlike traditional software that follows, like say a pre-programmed instruction, agents are actually much more dynamic. They can interact and they can actually behave in a way that's independent. So they can operate autonomously, you know, they don't have to have constant human supervision, and again, they can make decisions based on all that information.

SPEAKER_00 7:02

Right. And in today's landscape, and this is being recorded on February 22nd, 2026. Right now, in today's world, it is characterized by rapid innovation and intense competition when it's coming to these AI agents. It seems like almost every day, if not every day, there are advancements on what they're doing and how they're being deployed. It's it's pretty crazy right now.

SPEAKER_01 7:40

Claude has their co-work that you know you can create agents through along with Claude Code itself if you want to go more of the programming route, which most people don't necessarily need or want to do.

SPEAKER_00 7:50

But then you don't need to code anymore.

SPEAKER_01 7:52

Well, but in claude code, you still kind of can, is what I'm saying.

SPEAKER_00 7:54

It codes itself.

SPEAKER_01 7:55

It can, but but you still can do some coding in there. But but the big story, you know, the last couple weeks has been this open claw that was previously called Moltbot, and then was called ClaudeBot.

SPEAKER_00 8:07

And he's saying Claw, C-L-A-W. But that's where lobster comes from.

SPEAKER_01 8:13

Yeah, they had to basically change their name a couple times because they were getting nasty grams from like Anthropic and other places because the name likeness was too similar to some of their names. But open claws, the name has been settled on now for about a good week, which feels like two months in AI time. And what that was able to do is it took these, these your ability to make agents and made it much simpler using that program. And so what used to be you had to understand like N8N and how to plug things together through different software systems, it kind of stepped in and did that. But the cool thing about it is it makes it so that's interactive with your normal like iMessage or telegram. It can call you, it can email you, and you can actually respond back to those modalities and have it actually pick up your instructions or your interactions and move forward with it. So you don't actually go into a specific software system. It's like interacting with a coworker.

SPEAKER_00 9:10

Yeah, one of my agents texts me even when I tell him not to. Right. I know I might not be getting a restraining order against an AI agent soon. It's I'm joking, but in all fairness, security has been an issue with these AI agents, and that's why I mentioned earlier a lot of people are getting mini-max, which is a computer where it can kind of be it'll have boundaries on what it can do.

SPEAKER_01 9:44

Yeah, Snelly from a security perspective too, it also does so that it limits what the AI agent has access to, because basically you know these AI agents can go off and do some pretty amazing things, but there's also some risk associated with that that has access to things that maybe you don't want to have access to. So keep it in a contained environment, either in Mac Mini on its own or in a in a you know hosted with a with a but with a cloud service that actually has some structure and some firewalls built around it. So but but OpenClaw does does have some of those issues, right? But the cool thing about it is is that you can actually feed it certain information through MD files and that basically kind of help give it some context so that it actually makes it a much better agent for you. In addition to that, you can also create skills files for these different agents so that now they can actually learn and actually even update the skills files themselves so that going forward they actually become more skilled at whatever it is that they're doing for you. I think a big development to talk about too with the agents is you've seen the last game's gonna change, they become even simpler to use as we go forward. And a lot of tech companies are now investing heavily in AI agents and the development of them. So OpenAI, for instance, just struck a deal with OpenClaw, so they're bringing that on board now. You know, they're gonna enhance that. Google, with their latest Gemini release, is making it much more agent-friendly. Plus, it's tying into all of the Google network with, you know, Google Docs, Sheets, YouTube, and all the other things that Google owns.

SPEAKER_00 11:23

You know, well, an anthropic Claude, the Claude Code has a new security system out, I think, with an AI agent that actually defends against AI agents. So there are things coming out to make sure that security is gonna be better. But for right now, I wouldn't just let it loose on your computer.

SPEAKER_01 11:47

That actually highlights another point, too, and that it's gonna radically change cybersecurity going forward, too, right? So agents through like what Anthropic's doing now is gonna end up replacing typical cybersecurity as well as we go forward.

SPEAKER_00 12:01

So Yeah, you'll have like an AI agent that watches everything. Uh you'll have your own cybersecurity AI agent. That's what it'll be assigned.

SPEAKER_01 12:12

What'll be fascinating is there's gonna be bad actors that will use AI combating against the people that do the security of AI with AI agents. It's gonna be like this battle of AI agents, and they're gonna constantly like one up each other, going back and forth, kind of like how it exists now with hackers and cybersecurity professionals.

SPEAKER_00 12:30

Sure. So well, with all of this development, the market is moving so fast. And new AI agent platforms and tools, like we said, they're being released regularly. And when I say regularly, every day there's something new. And companies are experimenting with AI agents in so many different applications, you know. And I encourage you, if you can, you know, do it in a safe way, but things like customer service, data uh analysis, software development, you can create an app or a website with lovable, and there's just so much more. There's a sense of urgency, it really feels like in the uh in the industry to do as much as possible now as an early adapter. And I encourage everybody to just learn because AI agents, they're going to be our subordinates now that we're gonna manage to get things done. It's really something that's going to, I think, especially this year, be a huge tool that you're gonna have to learn to use.

SPEAKER_01 13:49

Many companies now, and if not now, well soon, be expecting employees to be able to leverage AI agents to make their work much more productive, to get more done, et cetera. So for those that were concerned about, like say losing their jobs, chances are you're gonna end up keeping your job, but you're gonna have to get much more efficient with AI to become more productive.

SPEAKER_00 14:11

It's like when we got the internet and people who knew how to use it were much more productive. Yes. But just companies are really starting to realize that AI agents are transformative. They really are. And you don't want to get left behind. This is driving rapid development and deployment in just about all aspects of business right now.

SPEAKER_01 14:39

That's a great point. You know, another thing too is the market opportunity for AI agents really is enormous. So first consider the labor market. There are millions of jobs that involve routine repetitive tasks. AI agents can automate many of these tasks, freeing up humans to focus on higher value work. A great case in point was I attended a seminar at CES and IBM was there, individual from IBM, talking about how IBM made that the strategic focus of their entire organization was to implement and drive productivity gains via AI throughout their entire organization. Their goal was to actually drive $2 billion worth of AI productivity gains, AI-related productivity gains within two years. They actually ended up doing $4.5 billion of productivity gains within a two and a half to three year period, blowing away the $2 billion goal set by their CEO.

SPEAKER_00 15:35

Yeah. I mean, these productivity gains are crazy. And you think about it though, AI agents can work 24-7 without getting tired or needing breaks or anything. They can process information much faster than humans. They can handle multiple tasks simultaneously, unless Liz is overloading them. These productivity gains could translate into a huge economic value for just about any type of company. Something else, you need to consider the new applications that are out there. AI agents could enable entirely new types of services and products that weren't even possible before. For example, AI agents could provide personalized education, healthcare, financial advice, so much more.

SPEAKER_01 16:33

Absolutely so much more. And remember also, it's only going to get simpler to use, more powerful as we move forward.

SPEAKER_00 16:40

So and it's even on the hardware and supply chain. What you were telling me about the ceramic from the toilet manufacturer.

SPEAKER_01 16:50

Well, that's okay, so we'll talk about that. So there's a lot of issues that are challenges that AI industry is facing, right? And it's not so much the development of the models and agents, it's the infrastructure to support all of it that's gonna become a bottleneck. So two major concerns. One is energy, right? So the amount of energy that we need to run these data centers is enormous. It's so significant. And we we're we're gonna have a major bottleneck in developing sufficient power to provide the energy needed for the growth that's going to occur. On the other side, there's also the hardware issue. It's it's growing so rapidly, and demands on hardware are so significant. It's not just like say chips. So there's these GPUs are the chips that are used in AI machines to run the to run the data. So there's a shortage of those, but but they can they're gonna add more manufacturing capacity. Grok and XAI with Elon Musk, they're actually gonna build their own fabs themselves to capture the amount of chips they're gonna need going forward because it's not enough just to outsource it through the existing uh players. But in addition to that, there's all the ancillary players that have to supply the the silicone, for instance, the the uh the machines to make it, the turbines to run, the coolant systems for these data centers. All of these are hitting bottlenecks, and so we have to figure out creative ways to to to work through those. And so one example is there's a toilet manufacturer out of Japan called Toto. And in America, you see some of those and in restrooms here, in uh restaurants and those sorts of things, industrial and commercial uses. They understand ceramics, which are important in building chips. So an activist investor actually, because they've had such demand for their products for chip-related manufacturing, that an activist investor just invested a significant amount of money into this Japanese company called Toto to kind of compel them to force all to focus all their resources and and grow that section of their business because the the demand is gonna just skyrocket and completely like blow away their existing revenue from the toilet manufacturing.

SPEAKER_00 19:12

So right. Some part in this chip needs ceramic. And it's probably dealing with the handle the heat. So what they're doing is they're trying to get this toilet manufacturer to use its manufacturing for the ceramic of the chips and uh and you know, away from toilets, right?

SPEAKER_01 19:35

Yeah, it's like it's like the car industry. You're gonna build a car, right? We've got to have all these like parts manufacturers that supply the different parts of the car. This is that car manufacturing suddenly was gonna, you know, grow exponentially. We got bottlenecks throughout the whole system. So you gotta work through those.

SPEAKER_00 19:49

But this isn't just a small toilet company. This is a huge toilet company.

SPEAKER_01 19:55

But so we talked about the labor market and that a lot of jobs involve repetitive tasks. We talked about productivity gains and how the AI agents can work 24-7, et cetera. Now, the third thing to consider are new applications. So AI agents can enable entirely new types of services or products that weren't possible before. We talked about that. So the fourth thing now is cost savings. Automating tasks with AI agents can significantly reduce labor costs for companies. So that can lead to lower prices for consumers and higher profits for companies at the same time.

SPEAKER_00 20:30

Yeah, I think that's a huge benefit, is that things are going to be much less expensive to lower prices for everyone. The market opportunity is so large that it's kind of difficult to quantify it at this time. Some analysts, though, are estimating that AI agents could create trillions, not billions, not millions, but trillions of dollars in economic value over the next decade. Let's look at some of the key players in the AI agent space and their strategies. Because as you can tell by the numbers we're talking, this is something you need to pay attention to.

SPEAKER_01 21:14

Well, better yet, leverage, right? Make sure that you guys are starting to use your agents and if not, start training on it soon so you can learn how to do it. OpenAI's position itself as a leader, I think, in AI agent development. We talked about how they brought in Open Coop. They're also building advanced AI agents and providing tools for developers to build their own agents. Their strategies to become the platform of choice for AI agent development, which I think is a very good move.

SPEAKER_00 21:40

Right. And Google is leveraging its existing AI capabilities and infrastructure to also develop AI agents. They're integrating AI agents into their existing products and services. Google's strategy is to make AI agents a core part of. Their whole ecosystem. So and I'm sure a lot of you who have been using Google have noticed these changes. I know I have. Yeah.

SPEAKER_01 22:09

And then Anthropic is focusing on building safe and reliable AI agents. They're emphasizing the importance of alignment and safety in AI agent deployment, like we talked about earlier. The strategy is to become the trusted provider of AI agents.

SPEAKER_00 22:22

One difference with anthropic OpenAI and Google are directed really at the consumer. Anthropic is more geared towards B2B. They're looking at businesses, is really what anthropic is more geared towards.

SPEAKER_01 22:42

Software development as well. So startups are emerging now also with specialized solutions. Some are building, you know, like we talked about OpenClaw, for instance. So some are building AI agents for specific industries like healthcare or finance. Others are building like broad-based things to develop agents in general. So the strategies to find a niche and become the leader in that niche.

SPEAKER_00 23:04

Right. And AI agents are being applied to a wide range of use cases. So the world is your oyster here. Customer service is one of the most common applications right now. AI agents can handle customer inquiries, resolve issues, and provide support. Just think if you upload all of your, you know, facts, FAQs, and you upload all of your procedures and all the questions, blog posts, everything, you're going to have the most knowledgeable customer service rep out there who can pretty much answer anything.

SPEAKER_01 23:49

And the voice of these agents is getting better and better and harder and harder to detect. There's no awkward pauses. They have inflections that elicit emotion out of the people they're talking with. You know, just their ability to string sentences together, very persuasive and empathetic and all those things. It's pretty remarkable. Like 11 Labs, again, a great example of what they've been able to do recently with their advances, just staggering. Really impressive.

SPEAKER_00 24:18

Another area where it's booming is data analysis. I talked about this earlier. It's really an important application. AI agents can analyze large data sets, identify patterns, and they'll generate insights. This can help companies make better decisions. They can make recommendations based on what they see. It's incredible.

SPEAKER_01 24:43

You're seeing across the life sciences, for instance, as well as you know, any business that has any type of data, but they're able to again like to spot these trends that maybe others can't as easily pick, as easily figure out. Software development is another one that's just crazy what's happened just in the last couple of months. So AI agents can now write code, debug, debug programs, optimize software. I I mean, we're I I just heard a podcast about, you know, the guy that ran or developed Claude Code for Anthropic, and he doesn't eat most of the people at Anthropic don't even code hardly at all anymore. So most of the coding is now done by Cloud Code. That wasn't the case two months ago. So hardly anybody codes in Anthropic anymore. They they edit things, they give some direction, you know, they kind of monitor things. But not only does the Cloud Code do the coding for them, it also is trying to come up with some ideas and some ways of doing things that the developers didn't even realize at first. So it's becoming even more efficient and better coding than what they were doing on their own. So you're seeing you're starting to see that developer productivity is just skyrocketing now. And with this recursive talk about it again, but this recursive self-improvement, as that continues to evolve and get better, the coach is gonna be able to just do it all on its own and make advances in ways that are just gonna blow people away. Like in a short period of time, it'll be able to code things and get better and better and better right away on its own.

SPEAKER_00 26:12

Right. But the old coders are now managing this and they need to make sure that it's being done correctly and verify.

SPEAKER_01 26:21

But the accuracy is improving radically. Sorry.

SPEAKER_00 26:24

Yeah, no, no, no. Exactly. That's exactly what I'm saying. It's making that their job so much faster. So, and research is another application. AI agents can conduct literature reviews, analyze research data, and they can generate your hypothesis. This could accelerate scientific discovery, like we can't even imagine. It's gonna be incredible what's going on.

SPEAKER_01 26:56

Absolutely. And business process automation is another important application. Some simple examples are AI agents can automate, you know, like routine business processes, like invoicing, expense management, and scheduling. I think the real key differentiator here is gonna be not so much like having AI come in and automate an existing process, but having the ability to rethink how you even do a process, because if you use AI agents, there might be a more efficient way of doing it than the way you're doing it now. So that's gonna be a key differentiator to pay attention to.

SPEAKER_00 27:30

Efficient and probably cost effective, right? Yeah. So, and that's what you really need to think about. The most significant impact of AI agents is gonna be on productivity. They can perform tasks faster than we can. They don't get tired, they don't get breaks, and they don't make mistakes due to fatigue. Now, they do have hallucinations still, and I have had to redirect my agents a few times, but you need that's why we need people to stay on top of them. And but overall, they work so much more efficiently than humans do.

SPEAKER_01 28:16

And not only the 24-7 thing, they can actually handle multiple tasks. Again, there is a limit, it needs some structure to it, but you can sit there and throw AI agents at multiple tasks, and you know, they can process these tasks in parallel. So it makes them even more productive.

SPEAKER_00 28:32

And they they learn and improve over time. As they perform tasks, they can learn from their experiences, and they'll learn more and more information, and it improves with their performance. Again, that's why you want somebody overseeing them to make sure what they're learning aligns with what you want them to learn.

SPEAKER_01 28:54

Aaron Powell So these models are getting better. The agents themselves are getting better, and and on top of just in general getting better, they also can learn specific skills to become even better at specific tasks on top of all of that.

SPEAKER_00 29:09

The whole economic impact of AI agents is going to be substantial. First, there's a direct impact on labor, as we've discussed. There's also an indirect impact on labor. The direct impact on labor is that they can automate many tasks and there will be less demand for the human labor. That's mostly entry-level areas. That could lead to some job displacement in some sectors, but remember you do need to keep people who are knowledgeable to manage them. Now, the indirect impact is going to be AI agents increasing productivity and companies being more profitable. This can lead to higher wages and more hiring and the ability to do more with your company, which will be a huge, huge economic growth.

SPEAKER_01 30:08

Again, there's also this impact on consumer prices. We've alluded to, I talked about it earlier. If AI agents can reduce labor costs, companies could lower prices for consumers, this could increase consumer purchasing power and stimulate further economic growth and sort of a flywheel effect.

SPEAKER_00 30:26

Yeah, and with everything we're seeing, you know, there's the impact on the innovation, they're going to enable new types of products and services, and this can create entirely new industries and new job categories. I mean, I I was in high school way before the internet, and since I was in high school, I think most of the jobs today didn't even exist back then. It's going to be like tenfold with what's happening with these AI agents.

SPEAKER_01 31:03

Yeah, I start to see it too. I mean, you have prompt engineers, for instance, you know, they're getting there's jobs out there for that. You know, strategic AI uh uh positions I've seen uh advertised for as well.

SPEAKER_00 31:18

And and one thing, you know, and we've talked about this before, a big economic impact could be such a state of abundance that we could end up having a universal income where, you know, in a form of socialism out of abundance. It's it's a possibility. I know Elon Musk has thrown that around as an idea.

SPEAKER_01 31:46

Pierre Diamond has done talked about it as well. He kind of also talks about there's gonna be the creators and the couch potatoes, right? So AI has become this like democratization of entrepreneurship. So anybody can come in and be an entrepreneur, employ agents, solve a problem, and benefit from that. And there's others that just because the productivity gains and the social understanding that's gonna happen initially as jobs get uprooted, given the pace and how fast things are gonna change and how disruptive it might be in the short term, that they're gonna probably come up with some sort of like way to bridge that gap for people. Well, people just might just not want to work, right? But they're gonna be, I think, locked in a lower level, is my take on it. They won't be as successful as those that are actually out there creating as well.

SPEAKER_00 32:31

So well, and it's all also what you want to do with it. If if the couch potatoes may have some depression and psychological crisis going on, just like what happened during COVID when people couldn't work, you you really feed your soul by being productive in life and creating progress and contribution to this world one way or another. And I think a lot of people need to think about do they want to be couch potatoes or do they want to be creative with this amazing new technology?

SPEAKER_01 33:14

I think you should consider that as your new uh AI agent, a therapist.

SPEAKER_00 33:22

I think I'd get in trouble with some boards there if I try to do that.

SPEAKER_01 33:27

Well, with that said, I think the net economic impact is likely to be positive, but there will be winners and losers. So some sectors will benefit greatly from AI agents while others may struggle. Some businesses within a sector will benefit greatly because they'll embrace it, while others who don't are going to struggle.

SPEAKER_00 33:44

Yeah, I think we're gonna see a lot of social disruption initially that we really need to be careful. And there's the challenge of alignment. AI agents need to be aligned with human values and goals. They need to do what we want them to do, not what we don't want them to do. And I sure have had some arm wrestling a couple of times with my AI agents on this. It's a fundamental challenge in AI development. So the human input on this and oversight, I think, is critical, at least in these beginning stages.

SPEAKER_01 34:28

Clearly. No doubt about that whatsoever. You know, so we talked about this challenge of alignment, we've talked about safety and reliability. There's also this challenge of transparency and explainability. You know, AI agents often make decisions in ways that are difficult for humans to even understand, right? So this lack of transparency, transparency can be problematic, especially in high-stake applications like healthcare and finance. So it's important that we build in guardrails to make sure that we get a good understanding of what it is that AI is doing in terms of decisions.

SPEAKER_00 35:00

And remember what we've discussed before. There is the challenge of security. AI agents can be vulnerable to attacks or manipulation. So assuring that the AI agents your business develops or you develop are secure is really a challenge you need to take seriously.

SPEAKER_01 35:23

I think another challenge is regulation, right? I mean, governments are still trying to figure out how to regulate AI agents. There's uncertainty about what regulations will be implemented, which could slow down development and deployment. There's also, well, I I think give I think there's some protections built in because sort of the prisoner's dilemma. You've got different, you know, countries that are trying to win the AI arms race, if you will. So China, I think India is going to become a player in this in the very near future.

SPEAKER_00 35:53

Oh, they already are. Yeah, they really are already are.

SPEAKER_01 35:57

I think another challenge too is initial expenses, right? If you don't manage your AI agents properly, and and you can actually run up a significant bill in terms of tokens, and that in the short term could actually be a barrier for some businesses as well. So you have to understand the there is a cost associated with this. It's not all free at this point. So you have to make sure that you manage that side of it as well.

SPEAKER_00 36:20

Okay, so let's talk about the competitive landscape and and how people should be approaching it. The competitive landscape for AI agents is getting intense already. And large companies do have a significant advantage. They have access to capital, talent, data, and they can invest heavily in AI agent develop. However, startups can compete by focusing on specific niches, making sure that they develop, the startup develops on their own an AI agent for a specific industry or for a specific use. If you can develop this and become a leader in that niche, a startup can compete effectively against larger companies and possibly even sell what they develop to a larger company.

SPEAKER_01 37:19

Absolutely. Right. I can agree more. And there's also this question of open source. You know, open source AI agent platforms could level the playing field, allowing smaller companies and individuals to develop AI agents in a way that competes with these large, you know, open AI model or model companies. So Anthropic, OpenAI, XAI, you know, Gemini, Google, and all those people. So I think the bottom line is the competitive landscape is gonna be dynamic. It already is. Everything in AI right now. New players will emerge quickly. The leaders of today might not be the leaders of tomorrow, right? So it's gonna be a lot of change, I think, is the one thing that is guaranteed.

SPEAKER_00 38:02

Well, how long ago did you first hear about Anthropic?

SPEAKER_01 38:06

Yeah, or you know, right. So first it was Chat GPT, right? Anthropic was not that far behind for me personally. But then Google got into it. Look how fast now Google Gemini, the latest version of Gemini, today, as we speak on February 2022, Google Gemini is like the top AI model right now. So it just changes constantly, right? You've got also, you know, the the Chinese open source players, you know, DeepSeek and others, right? They're doing phenomenal work. So uh it's it's just it's just so much is happening so fast.

SPEAKER_00 38:38

So you know what? Here's what I really want to get across. AI agents are a really significant and probably critical opportunity for all businesses and companies. If you're developing a successful AI agent, you can see unbelievable growth. Investors are betting on AI agents as a major trend.

SPEAKER_01 39:07

Yeah, we saw it with IBM, like my example earlier. You know, millions of dollars of productivity, measurable productivity gains within a few year period. And that's before the AI agents and the models became even more powerful, like they have recently. Now, there is also significant risk. Many AI agent companies may fail, right? The technology is still evolving, and it's unclear which approaches will ultimately be the winners and be successful.

SPEAKER_00 39:33

Right. So be careful in which companies you invest in. You need to look for companies with strong technology, experienced teams, and clear business models.

SPEAKER_01 39:47

So let's take a look ahead, Liz. And as we look ahead, I think we can expect to see continued rapid development in AI agents. That's clear. It's already been going on. We're we're we're getting deep into it. That that that train's not going to stop. It's going full steam ahead.

SPEAKER_00 40:04

And they're only going to get more sophisticated, right? What is everyone's saying this is the lowest end of AI agent technology and the most expensive that it will be. They're all saying it's amazing now, and this is gonna be the worst it'll ever be in terms of its capabilities. We're only gonna see more sophisticated AI agents who can handle more complex tasks, even the ones that I give it. And it's we're gonna see agents being deployed in more age more industries and use cases, and we'll see new business models emerging around AI agents.

SPEAKER_01 40:50

I think one point to add to that is it's also become simpler over time. Yeah. So being able to get an AI agent in place and doing the work you want it to do will become easier and easier, even though it's at the time, or over time getting more and more robust and handling more complex tasks and come up with better and better results.

SPEAKER_00 41:12

One big problem is we will most likely also see a labor market disruption. As jobs become automated, new jobs though should get created. Every time we've had a disruption, the industrial revolution is a big example. The jobs created were greater than the jobs that there were before. Same thing when the internet was developed. There are more jobs now because of the internet. And the same thing is going to eventually happen, but there will be a disruption and it's gonna happen fast.

SPEAKER_01 41:48

I'm curious too if that you know, when that does happen, because it's more of a when, not an if, is that gonna prompt governments to increase regulations, for instance? And what kind of effect is that gonna have? Because again, some countries might instill instill regulations, others might not. Is that gonna change the competitive dynamics of all these companies as well? So kind of curious to see how that's gonna evolve too. Or let's say that there's gonna be at some point some AI that goes sideways somewhere. Is that gonna lead to more government regulations? And if so, will that change the dynamics too?

SPEAKER_00 42:22

So either way, the AI invasion, AI agent invasion is real. Very real. We're witnessing rapid development and deployment of AI agents across just about every industry.

SPEAKER_01 42:37

Yeah, but you know, there's also these challenges too, right? So the next few years will be critical. How we develop, deploy, and regulate AI agents will have significant implications for the economy, the labor market, and society as a whole.

SPEAKER_00 42:52

So until next time, thanks for listening to AI Lens. We're your focused view on the emerging hot topics in the age of AI. We provide AI news, advancements, and discussions about how AI is reshaping business and society. If this episode about the AI agent invasion made you curious or skeptical, make sure to follow, subscribe, and share this episode with someone who is curious about where AI is headed next. Until next time, stay curious, stay informed, and keep your lens focused on the future.