OpenAI's New AI Agents Just Got Leaked....

Did you know that OpenAI are launching a new tier of AI agents that are probably going to replace many individual workers? This is something that's super interesting, so let's dive into all the details and everything you should know. So if you actually take a look at this article, it talks about how OpenAI is plotting charging $20,000 a month for PhD level agents and the fact that they have told some investors that it had planned to sell low end agents at a cost of 2, $2,000 per month to high income knowledge workers, mid tier agents for software development possibly costing $10,000 a month, and high end agents acting as PhD level research agents which could cost $20,000 per month according to a person who's spoken with executives. Now that seems like an absurd number for an AI tool, but this isn't the first time OpenAI has launched agents and we, when we actually look at these prices, I think if OpenAI can actually deliver on their promise of actually having these agents be successful, then I do think that it is very possible that they could have agents that are at this level. Now when we actually take a look at what these are, of course you can see that there are three different tiers from 2000 to $10,000 to $20,000 per month. And of course companies would have to justify spending that much in terms of, you know, what they're going to do on those agents. So for me, I do think it is possible because Whilst we think $200 is pretty expensive, and yes it is for one person, companies willing to spend ridiculous amounts on AI are definitely going to be scrambling through the doors to get their hands on this kind of software. And remember guys, this kind of software isn't just going to be an AI tool for $10,000 a month or even $20,000 a month. This is going to be an AI agent. Now remember? So this is where we of course have the taste of AI agents. And I'll say that because most people don't realize that Deep Research is actually an agentic workflow behind the scenes where you actually have AI research agents, you have AI evaluation agents all working together behind the scenes to orchestrate a detailed report that you get at the end of your queries. So if you haven't, you know, understood, I didn't actually make a video on this at the time, but, but OpenAI released deep research, which was a really, really impressive tool that allowed you to do probably hours of research work just compiled into around 10 minutes. And this is an invaluable tool for those of you who have Tons of research tasks every week. Maybe you're looking for the latest alpha in the stock market, wondering which stocks are going to be the next best play. Maybe you're looking for some research on your work. I mean there are a billion different things that you could do use deep research for. But this was of course the first taste of agents. Now why did I include this in this video? Well, this first taste of the agents, this was actually received very, very well amongst those who tested it. And I tested it myself and it's definitely something very comprehensive. So look at the individuals who were talking about deep research. They said this is something truly special. It conducted an expert level business and technical analysis on Deep Seq's entire R& D history and made excellent extrapolations. Not only did this person say this, if we look at what other individuals said, this singularity post, you can see that some people said I just used deep research for work and I'm in shock. I've given it three separate queries, arching each one about detail, yada yada yada. And this is pretty crazy. And of course you can see someone else here, he has said that, look, deep research is the oh crap moment for the majority of modern knowledge work. And so the point I'm trying to show you guys here is that this is just OpenAI's first simple research agent and we have to think about it, okay? If this is what these kind of agents can do in terms of research, what happens in, let's say five years or so when they are able to do not just that, but they're able to do many different things on a computer, run applications, use different tools, book things. I mean the future is going to be super, super interesting with as to just how much leverage we're going to have. And this is why I think OpenAI is looking to truly dominate the agent space by offering these crazy, crazy tiers of AI agents. Now remember we spoke about in the beginning that these AI agents are probably going to cost somewhere between 20k per month. Now you might think that that is pretty expensive, but you have to understand that for what you're going to be able to get, I definitely feel like it's going to be able to bring you more value than you actually pay. So for maybe $10,000 a month for the software developer agent, you actually get a bunch of different agents working in your company that are probably 10 times more efficient than an actual software developer. And it can be run at 24, 7 and of course PhD level research agents. I'm guessing that maybe they're going to be selling those to universities, to research labs, and they're able to help scientists find different things that they otherwise would take them a long amount of time. Now, if you're thinking about why they are charging so much, I also do think OpenAI is a company and they of course need to make a lot of revenue. They actually spoke about how their 2024 forecast, well, it actually was 3.7 billion and their 2025 forecast was $12.7 billion. And they actually expect a lot of that to come from agents. Now, crazily, SoftBank actually did buy, I think it was $3 billion on OpenAI's agents. So they are super, super bullish and I can imagine other companies going ahead and spending a large amount of money on OpenAI's agents. Despite the fact that there might be some, you know, open source tools that are going crazy on the market right now, I do think that OpenAI's brand name still holds a lot of weight. Now remember, when we do look at agents, one of the things that many people spoken about is the fact that if OpenAI does launch some kind of coding agent, it's probably going to be very similar. So they actually talk about the fact that OpenAI are going to release a coding agent similar to what Devin has. And I do wonder how OpenAI are going to do this, because one thing that people do forget about OpenAI is that OpenAI aren't actually a model company, but they are more so a product company now, which means that even though they make good models, their bread and butter is actually making these things into products that people actually like. And I think that's why a lot of people use ChatGPT.

In fact, OpenAI is not a model company, it's a product company that happens to have fantastic models at this point and which is great for them and great for us as, you know, both partners of theirs.

And so the reason I'm stating that is because OpenAI, they're going to have a tough problem of making an even better product than Devon and and an even better product than Anthropic. And that does require them to make a really good coding tool that somehow manages to be anthropic. Not only that, but beat the use case of Devon, because if they are launching their, you know, own coding agent, then it's going to have to be better than the Devon coding agent. And if you don't know what that one is, that one is one that essentially it costs around $500 a month and it's mainly pitched towards enterprises as a junior developer. But it is something that people do use, but it is something that's not for personal use. So if we do take a look at what's going on here, you can see that some people have spoke about the fact that Devin may not take off like Cursor, and that's not just because of the $500 a month starting point. It's because Cursor is a lot more easier to adopt and they've got an incremental approach. So this is something that, like I said before, it's where OpenAI may differentiate itself. It's probably going to have an agent that is quite like a software engineer and I do think it's probably going to be a lot more comprehensive than Devon, which may ruffle some feathers, especially in the AI agent space, but I'm definitely going to be excited for that. Now, of course, with OpenAI's coding assistant, and you can see right here in January, the information actually broke the news that OpenAI actually has been working on a coding assistant targeting senior software engineers and explain the kind of tasks it might do. And that's of course quite like the mid tier agent that they discussed with investors. So if OpenAI can deliver on their promise of this and of this actual agent, then it would make sense for them to charge 10k a month for this. Because I think to many different companies that actually might be cheaper than a real software engineer. Now I do think, of course, you know, you're probably going to have a software engineer managing these AI agents and I do think these AI agents are probably going to be a larger stack of what companies are going to be able to do in terms of their output. But I am saying across the board it seems like these are going to be really, really valuable to companies considering just how much they're going to be able to do. Now, in addition, you can see right here that it also points out the fact that like OpenAI is still unprofitable on the $200 a month tier. So it's quite likely that prices are going to be raised. And honestly, when I do think about the amount of value that you do get for $200, I don't think it is that bad if you are using AI all the time. I think the amount of value that you get from advanced voice mode to just talking with the models to have them research your problems, I think if you're someone that maybe works in social media content creation, doing research and you're on the computer all the time, I think $200 a month for the amount of value that you're really getting, which is, let's just say it's like mini AGI in a bottle. I definitely feel like it's quite valuable. But of course, if you're just someone who uses AI to play around with, it's definitely not something that is that valuable for you. But of course everyone is going to be pretty different. So of course you can see it says it stands to reason that customers would be willing to pay a lot more for AI that effectively does the work of a 200,000 per year software engineer or to speed up important medical research, which of course they would now remember. OpenAI actually spoke about super agents and super agents was an article where they were speaking about some kinds of, you know, information that OpenAI employees were really psyched about. And super agents were this thing where they were talking about how AI agents in the future are going to be a lot more advanced than they are now. And the fact that AI tools designed to tackle messy, multi layered real world problems. The human minds struggle to organize and conquer. And these agents don't just respond to a single command, they pursue a goal. Super agents synthesize massive amounts of information, analyze options and deliver products. So that's going to be something that's really interesting. And what happens when we take a look at this? You can see that if we have super agents, these are going to be AI agents that build from scratch. So imagine telling your AI agent build me a new payment software. The AI agent could design, test and deliver a functioning product. Make sense of the chaos for financial analysis of potential investment. Your agent could scour thousands of sources, evaluate risks and compile insights faster and better than a team of humans, which is really true. And master logistics planning an off site retreat. The agent could handle scheduling, travel arrangements, handouts and more, down to booking a big dinner in a private room near the venue. And I think all of these agents, or at least some of them, are probably going to fall into this category. So remember the tiers of AGI or tiers of AI. I think a lot of these AI agents are probably going to fall into level four or five, which are going to be AI agent frameworks that are essentially just so many different AI agents connected together that can do the work of a company. What does the value of the economy look like when we have companies like that that are able to have all these AI agents doing a variety of different tasks? I don't know about you guys, but this is definitely something super interesting. And I'm going to be using AI agents to grow my social media, because that is where I find that they are most effective at creating content on the fly. And that's something I'm doing in my community. So, with that being said, what do you guys think about OpenAI's $20,000 per month agents? Do you think it is all hype and bluff, or do you think OpenAI is really going on? These agents, for me, it shows me where their head is at. It shows me that OpenAI is clearly solving the problem of actually delivering real world products that people would actually use. And I feel like it's going to be a really interesting year because OpenAI definitely has to maintain their lead, and in doing so, they're probably going to release some amazing products. So, that being said, hopefully you guys enjoyed the video and I'll see you in the next one.