Big Tech's New AI Obsession: Agents That Do Your Work for You
AI agents can automate complex tasks like sending emails and making calls, unlike simple chatbots. Companies see efficiency gains and cost cuts; workers worry about job displacement. The technology is still unreliable for critical work, but experts say to experiment now and develop skills machines can't replicate.
Chapters
Greeting from post-yoga bliss. Today's video is about the rise of AI agents and what they mean for employees and employers. So AI agents are the new hot thing coming out of Silicon Valley. And if you pay attention to kind of how tech works, once every couple years a new hotness comes out that threatens to take over the world and change everything, whether that's Uber, mobile phones, internet, the Segway, if any of you are old enough to remember that. And AI agents are the new hotness and what everybody's talking about. So an AI agent is kind of like, if you've used ChatGPT or Perplexity, or any of these interfaces where you get to talk to an AI and it edits a document for you or whatever, and it's all just kind of this powerless thing. Imagine an AI agent is something that you're able to give instructions to, but it's able to connect and actually take actions digitally. So it can do things like write emails and actually send those emails as opposed to you having to cut and paste them. And so the idea would be that you could create these agents, give them tasks, and kind of like a college sophomore working for you, uh, you have to watch them, but they can go do things that are projects that can take hours, weeks, days, uh, and go from there. So let's explain this with like an example. So imagine you wanted to create an agent that would be a travel agent, right? If a travel agent was, say, a real person, what would they do, right? That person would ask you some questions about what you want to do for your vacation. Uh, they would go put together some research, they would go identify potential resorts and stuff like that that you would want to go visit. Uh, they would call those resorts and see what's available. They would talk to those people, see what the prices were, get quotes for you. They would analyze all that, come back and give you a recommendation on what sort of vacation they would recommend for you, give you a couple options of fully planned itineraries, uh, that then you could look at and make a decision on based on their recommendation. So imagine instead of a person doing that, software does that. And that's kind of the promise of an AI agent. And you can see where this goes, right? The difference of a person doing that kind of travel agent job versus a piece of software sitting on a piece of computer is hugely different, right? The person can only work on one of those at a time. The software could potentially work on hundreds of them. So imagine suddenly, instead of having an entire fleet and a call center full of people that are travel agents taking phone calls, doing all that work, it all just becomes a single computer that works for basically for free, right? You have to pay so much money in terms of making all that happen. And this, to some extent, is why companies are so excited about this, right? It has the potential to make them incredibly more efficient and create entire workforces that could basically make them provide so much more value to customers without having to hire a bunch more employees to do that. And of course, if you're an employer, you're pretty stoked about something like that, but potential for employees, not as great, right? If you have one of those jobs that is somebody that could be replaced by one of these agents, uh, this news actually gets pretty scary, right? Could your job be done by a computer software like for pennies on the dollar, right? Who cares about worrying about if your job's going to get shipped overseas when you have to worry about a piece of software taking your livelihood from you? So several thoughts about how to think about this. So number one is every piece of technology that goes and comes out of Silicon Valley has a very similar kind of curve. It goes through what's called the hype cycle, and this was produced by a research firm called Gartner. And basically it says that early on there's this huge promise of visions of the world being changed. And like a great example of that is autonomous cars. 10 20 years ago, everybody was promising that cars would be driving themselves. Guess what I'm doing? Driving a Subaru today by myself without any software. So the reality is most pieces and almost every single piece of technology goes through this hype cycle where there's this big promise of changing the world. It crests, people start to see what it really can do and the limitations of the technology. We all get disillusioned and then it takes years, if not decades, for that software to truly come in and change the world. So the good news is like these changes, they're not gonna come immediately. Like we're not gonna lose 80 million jobs next year because of AI. But the interesting question will be, does the software actually get there based on how we're trying to do things today, right? There are some very interesting shortcuts we're taking in terms of how we're creating these AIs that leaves a big question as to what will be the actual future end state of their ability to truly do things that matter. Because right now with AI, it's kind of like hiring a college sophomore. They know a lot, they've read a lot of books, but they don't have a lot of experience, and you really don't trust what they end up telling you. So everybody I know using AI for different stuff. They're frankly using it just for stuff that isn't mission critical. Who cares if somebody chooses a wrong word in something that they're editing for you, or they find a few things on the internet that happen not to be true? Again, you're not using AI and these tools for things that really, really are life or death, and really, really that are important. So I think that's where these agents will go, right? I, I do believe they're going to be something that will truly turn out to be valuable in society, but it's going to be years until they truly start to make the kind of impact that the Visionary is talking about right now. Envision? So I think for us, like, actually the situation leads us to a pretty good place, which is number one, like, if you want to go play with these things, they're actually pretty fun. You can go out and try to build agents that will try to do different things and send emails on your behalf and answer phone calls. Or I've seen articles about people creating them, pretending to be you, right? The AI agent will pretend to be like a Girdley bot, which to me is like super cool. Like, why the hell not? But I think if you're a business owner, The good news is we have time, right, to see these things truly change the world, just like it's going to take time for autonomous cars and humanoid robots and all that kind of stuff to truly hit the streets and be mainstream and meaningful and things that you would trust around your kids and trust around things that are truly important. But I mean, if you're an employee, I think things are a bit scarier, right? I think we're seeing in the near term a situation in which people have to worry about their jobs getting offshored, right? If your job can be done across the internet, then why are they going to put it here in San Antonio when they can put it offshore in India or Argentina? And there's a future where even those jobs are going to get eaten by software and taken over. And so I think that's the important thing if you're an employee is can your job be done in the future by software or somebody overseas? And if so, how do you make sure that you create skills and a place and a set of experience and all that kind of stuff to where you don't have to worry about AI taking your job. And, you know, that way it's easier to feed your family. But in the meantime, it'll be super fun to watch. That's what you can say about America. Nothing's ever boring.
