Transcription
The internet is abuzz with conversations about Satya Nadella's statement that "SaaS is dead." I'd love for you to tell me how that shift happens from SaaS all the way to agents. I have no idea who he might be referencing with this whole "SaaS is dead" thing, but Satya Nadella, the CEO of Microsoft, gets asked directly the question: Is SaaS dead?
What does the transition look like from traditional software as a service applications to agents on top of databases? It's exactly what he said last time; he doubles down on it. So to me, yes, SaaS is dead, or at least it's dying. Not only that, Satya talks about a future where people are hired based on their agents.
Imagine your LinkedIn profile or your resume—rather than saying all the places you've worked, it says all the agents that you've created, all the workflows that you have in your tool belt. That is a likely future of what a workforce will look like.
So let's watch, and I'm going to give you my thoughts. This video is brought to you by Together.AI, the leading AI acceleration cloud for developers. More on that later.
Whenever there's been a real platform shift, the core application architectures have changed. Right? I mean, if you look back, let's go all the way back to when the relational database was born. That was the first time we really said, "Oh wow, I can separate out my data tier from my application."
Before that, we were building these essentially ISAM databases right into the app. Then we said, "No, no, no, let's have relational algebra. I have SQL, I have data, and then I can build my business logic on top of it." Of course, other platforms came, like the web and what have you, and then we said, "Oh, what's an end-tier way to write applications? How should I restructure my business logic?"
I think that something of that scale, or if not more, is going to happen again. All right, so let's just pause it right there because he's already saying there is a fundamental shift in how applications are being built.
I actually did not know this way back in the day. Apparently, business logic and database logic were all mixed together, and application developers actually had to write the database interaction logic themselves rather than just using a standardized database like Postgres or some other kind of database application.
Now he's saying we're going through another shift with the application logic, and that's really what I said in my previous video: the application logic is going to shift entirely to agents. There is no point in hardcoding a UI anymore.
So when I think about SaaS companies, especially very verticalized SaaS companies, it's essentially a layer of business logic and a UI on top of a relational database. But if agents are handling the business logic and they're handling the UI creation, what is left? Just agents and just the relational database.
So then who owns the agents? Well, it might be a new company, it might be traditional SaaS companies, but most likely it's going to be only the largest of them, like Salesforce, and then everything else just goes away.
In my mind, this is actually an intermediary step. There might be another step after that in which everything is just predicted, just like we saw with the AI Doom demo. Essentially, every frame is going to be predicted, and then there's going to be the ground truth of whatever is in the database.
Let's keep watching. This time around, the thing about agents is they are not going to be bound to essentially anyone's SaaS application and its data. Right? So I'll have an AI-tic sort of view where the task, the intent, and I will go operate and orchestrate all the logic across multiple SaaS applications.
I'll go call a bunch of APIs through tools. I'll call a bunch of APIs. In fact, I'll post-train my model to know about multiple SaaS applications in the agent tier, and so that's what's going to happen.
All right, so he also clarified something else interesting. By the way, again, he basically didn't refute that SaaS is dead. He is saying it's going to change dramatically, and yeah, SaaS is basically in its current form dead or dying.
The interesting thing he said is agents are not going to be specific to any one application or even database. These agents are going to be able to write their own code, write their own UI, and interact with any API, any database, any application, any business logic.
That's what makes this a fundamental shift in how software, software as a service at least, has and will be written in the future. You are going to create your own agent or agents, and they are going to do everything on your behalf.
Because they are so generalizable, you don't need to create agents that are specific to any one industry or business logic or even database. You can work across any database because these agents are just going to be so good.
And stick around because, speaking of how good agents are going to be, OpenAI apparently is working on something called super agents, which I'll touch on in a little bit.
Now, SaaS might be dying, but you know what is very much alive? The partner of today's video. Thank you to the partner of today's video, Together.AI. Very happy to be working with them. Together is an incredible company. They are the leading AI acceleration cloud for developers.
Whether you need fast inference, fine-tuning, or training of AI models, you can do that all through a simple API. They have industry-leading inference speed with the proprietary Together inference engine, four times faster than VM.
They have seamless access to 200-plus models from Llama, Quen, Mistral, Flux, DeepSeek, and that spans text, images, audio, code, and embeddings—all through a single platform. Together is already used by 300,000 developers and major companies like Salesforce, The Washington Post, Zato, and Zoom.
Test out the brand new incredible DeepSea R1 thinking model directly on Together with insane inference speeds. Not only do you get incredible performance, but you can also get better privacy options, honestly, than using DeepSeek directly.
You also get the full 160k context window as well as multiple flexible deployment options. So get started instantly with serverless model endpoints, or you can spin up your own GPUs. Make sure you click through my link together.ai/Burman so they know I sent you, which of course helps us.
If you do that, you unlock free access for a limited time to Llama 3.2 and Flux model endpoints. So start building together on Together. And thanks again to Together for partnering with me on this video.
Now back to the video. All right, so let's hear a deeper dive into his thoughts about how these agents are going to interact with CRUD databases and the future of SaaS. Let's keep watching.
I think what'll happen is these CRUD—I mean, SaaS applications are a CRUD database with a lot of business logic. All right, real quick, let me just explain what a CRUD database is just in case you don't know. Create, read, update, delete—basically the core interactions between business logic and a database.
Create: you create a new record in the database. Read: you just read from it. Update: you update it. And delete: exactly what it sounds like. So let's say I put in a database, "Matt has three apples," and then all of a sudden I get another apple. So update: "Matt has four apples."
That's all CRUD. The database will then get orchestrated outside of the business logic tier of just the SaaS application. What I mean is going to happen—like right now, in my own use case, I go to Copilot. I say, "At sales," which is actually touching Dynamic CRM, brings back whatever the account information.
Then it brings back information from Office 365. I put it into pages, I share it with people—the entire workflow. Right? I never—I mean, everybody talks about their CRM database, but nobody uses it because, you know, when was the last time I logged into CRM? Never!
Except now, I'm every day querying my CRM database because it's so much easier. It's one agent away, and it's working with all the other agents. So that is what's going to be the change.
Imagine this future in which you just have your agent or team of agents, your workflows, and you simply just say what you need, and they go out, they abstract away any technologies, any databases, any business logic, and they just get you what you need.
That is the promise of artificial intelligence, and we're already seeing a lot of that, but it's only going to get better. Agents are only getting better; AI is only getting better. This is the best it's ever going to be.
In this next section, they are going to talk about hiring humans and how hiring humans is going to be very different because it's not going to be necessarily based on your skill set as much as what agents and what workflows you have created and which ones you are going to bring along with you.
Let's watch.
When you're hiring people, I think in the future you'll now hire people plus, you know, their workflows. That's exactly right. In fact, that's a way to think about it. It's a swarm of agents. I mean, in a practical way, kind of look at it and say it's not like when you hire a data analyst. You don't just hire them; you hire them and their spreadsheets.
Yeah, that's kind of what it is, right? So it's like agents are going to—I think two years from now, we're going to say, "Yeah, agents? Yeah, I build them like all day, like I build docs and spreadsheets."
And I think that's—and I come with a basket of them. There you go. I have a basket of my agents.
All right, so he made a really good analogy. When you're hiring a data analyst or a financial analyst, you're basically hiring them and their spreadsheets, their knowledge, and their templates—what they've built up over time and actually formalized into some kind of documentation.
That's going to be the same thing with agents. You're going to be hired based on the agents that you build up over time.
Now here's something interesting to think about that they did not cover. Let's say you're going out for a new job, and you join a company with your group of agents. One of the things that I really don't see people talking about is onboarding.
How do you onboard your agent into that organization? Because just like a human, you need to provide the agent or agents with business rules that are specific to that company. Memory—all of the internal documentation—all of that needs to be onboarded into the agent so they can hit the ground running right away.
Rather than that taking days, weeks, or months like a human would, agents could potentially do it in a matter of minutes or hours. So I really believe that onboarding process might actually be a pretty big business or at least a big piece of the agent stack.
By the way, if you've heard of somebody doing this really well, let me know because I haven't.
All right, let's keep watching.
You know, like, I already see that, right? Like, you know, every SharePoint—like, then I have a leadership meeting, and there's like all these documents. The best grounding data is all in there.
So I just have a simple agent now, which is a SharePoint agent that I'm always addressing, and it's fantastic not to just have to go to a separate entity and, you know, query it, but to have it right there.
So just imagine that you, as an employee, are as valuable as your agents. Agents are joining the workforce in 2025. This is a theme that I've been talking about for a little while now, so make sure you're staying up to date on all of the latest agentic frameworks, the best practices.
Make sure you're getting in the weeds and actually getting your hands dirty building them yourself, seeing what works, seeing what doesn't, because that'll keep you really valuable in the workforce for the foreseeable future.
And I'm not the only one saying this. 2025 is the year of agents. Here's Greg Brockman, one of the founders of OpenAI: "2025 is the year of agents."
Baby steps in that direction. Now what he's referring to is the ChatGPT task feature, and it is really a baby step. It's essentially reminders, but done with AI, and it's more powerful than traditional reminders, but still, that's the basis of what it is.
The reason why he says this is agentic is because it is AI that is able to accomplish things on your behalf in a scheduled way.
On this note, OpenAI apparently is going to be releasing super agents soon, so let's quickly take a look at that.
This article on Axios came out over the weekend, and they talk about PhD-level super agents coming soon. In the article, we've learned that OpenAI CEO Sam Altman, who in September dubbed this the intelligence age and is in Washington this weekend for the inauguration, has scheduled a closed-door briefing for U.S. government officials in Washington on January 30th.
Whenever this happens—and this has happened before with the GPT-1 model—there is something new coming. Sam Altman shows the government first.
It's funny because they also reference in this article Mark Zuckerberg saying AI is going to be replacing mid-level software engineers and other human jobs this year. They already said that as part of Meta.
OpenAI has released the economic blueprint, which I didn't really cover on this channel, but basically shows what the future of the economy might look like when we have millions, billions of agents.
According to the article, several OpenAI staff have been telling friends that they are both jazzed and spooked by recent progress, and that's what we've been seeing. I've seen a ton of tweets from OpenAI employees talking about, "Can you feel the AGI? Can you feel the ASI?" talking about agents and everything.
So something is coming. They've seen something. According to the article, super agents are AI tools designed to tackle messy, multi-layered real-world problems that human minds struggle to organize and conquer.
They don't just respond to a single command; they pursue a goal. Super agents synthesize massive amounts of information, analyze options, and deliver products.
What I really believe this is is GPT-1 as the core intelligence layer mixed with some kind of agent swarm framework and with memories and tools and everything, and that is a super agent.
I've talked about Salesforce before, and they're not hiring software engineers this year. Take a look at this clip in which Salesforce CEO Marc Benioff details why they are not hiring any more software engineers this year.
"Our head of engineering came to me about three weeks ago and said, 'You know, we've been deploying this technology now over the last basically two years. We're writing code, delivering technology and capabilities faster than we've ever done, and we've got tens of thousands of engineers. We have 75,000 people total, and he's like, "I don't want any more engineers this year."
I'm like, 'What did you say?' 'I don't want any more engineers this year.' I'm like, 'You've never said that before.' He goes, 'I know, but we are just in the last year 30% more productive, and we're doing great, and we just want to focus and get all this done.'"
So between SaaS dying, software engineers either becoming insanely more productive or getting replaced by agents, things, especially in the software industry, are changing drastically right now.
It's an exciting, slightly scary time to be a software engineer and just generally in the world of technology.
Thank you one more time to Together.AI for partnering with me on this video. Go to together.ai/Burman. All the links will be down below. If you enjoyed this video, please consider giving a like and subscribe, and I'll see you in the next one.