OpenAI DevDay: Dots, Agents & $100B Opportunities

0:00I gotta be real with you. I didn't think I'd be making this episode of the podcast, but I just watched almost 60 minutes of Sam Altman up on stage and he launched a few things that I think are going to make people an aggregate billions of dollars. And I don't think that's a stretch. I think that there were 20 launches, feature launches today, but I think there was probably three or four that are so big that I think the people who figure out how to use those features and to grab a slice of those billions of dollars of transactions that are going to be happening within the Chatcheepe T platform. I think they're going to be able to create some really, really big businesses. So I had to come on here and just break it down for you. If you're new here, my name is Greg Eisenberg. I'm the host of the Startup Ideas podcast. I started in sold three venture-backed companies. I was an advisor at a tick-tock. And I come on here because I think that this is an incredible time to be building your ideas.

1:03And I think that in the AJI, you can use tools like this to go and do that. So by the end of this episode, you're going to learn about the most important things that happened today that Sam Altman launched at DevDay 2026. I'm going to give you some ideas and some frameworks for thinking about how you could come up with your own ideas so that you can grab a slice of that pie. And I'm just going to cut through the noise. So by the end of it, I'm just going to really focus on for people interested in building businesses, making money around AI. This is the episode for you. The Startup Ideas podcast. It's just in time, baby. So the bit, yes, like I said, there are 20 plus launches, feature launches today. The most important ones to me, the ones I think are the most interesting, and I think maybe for you,

2:03is let's start with dots. So we have to talk about dots. Dots is the personal agent platform and it's their answer to Muse by Meta and their answer to Instink. And what's really cool about it is they now have this front door where you're going to have access to the 1.2 billion weekly active users, that's what Sam Altman said today. So people are going to go and use their personal agent platform dots and they're going to be asking questions. They're going to say, I need this permit fixed or I have this request or I'm looking for a restaurant that serves this. And chat GBT is going to understand that job. And your plugin that you have made, which is essentially like a mini app, is going to get selected for the work

3:03and we're going to talk about ways to surface that. But they've basically announced that plugins, they're doubling down on plugins. I think Sam said something around, you know, they didn't say it verbatim that it's like plugins were a failure, he didn't say that. But he did say like something to the degree, I'm rather paraphrasing that they had some like mixed reviews on plugins. They're now doubling down on plugins. And I believe Sam that he's going to like make plugins a thing and he's going to try to make chat GBT the app ecosystem for AI. So there and that's a huge announcement of today and we're going to talk about some of the things, we'll go deeper on that. So your business, you want your business to go in, like fix the customer's request, file the permit, you know, if it's an example of like a permit company, and then you're going to want to do the result.

4:04So this is the new app ecosystem. You need to drill this into your brain. You, you know, have to figure out how you could focus on the right customer. And chat GBT is going to understand the job, so they're going to do that. You're going to have to create the right plugin that does the right thing and actually creates a growth result. Because if you can do that, you know, you can be hired before the customer knows your name. You could have access to the 1.2 billion weekly active users. And we're talking about, like, you know, I said in the intro, like billions of dollars of transactions, I'm even being conservative. Like I don't think it's crazy to say that in 10, 15, 20 years, that there's trillions of dollars of transaction volume happening within personal agent platforms. But let's be conservative and say that there are billions of dollars that will be through, you know, in transaction volume. So now that we understand that, let's go back to, you know,

5:06DevDays website. So you can see here's their personal agent platform. The other thing that they announced today is DecisionJPI. I did a video, I think it was last week on Dev. And this is their answer to Jeff. So you saw their answer to Muse. This is their answer to Jeff. DecisionJPI enables real-time decision making by focusing Luna's intelligence on a specific set of user-defined questions with finite predefined answers. So developers supply context using text or images, and they get back answers they can use to classify content, route requests, or choose an agent's next action. So it's in limited preview today. Excited to play around with it. I've been hooked on Jeff. Hooked on Jeff. I think there's a bunch of, you know, use cases for Jeff that are, you know, people are not talking about, underserved. You know, one that I think is really interesting is just,

6:07you know, UI design. How do you use a, essentially, Jeff as a DecisionJPI? How do you use a DecisionJPI to get you the most highly converted, converting, you know, app store screenshots or direct response websites and apps, things like that. So there's a ton of opportunity there. Question here is, will OpenAI's DecisionJPI be worse as good or better than Jeff? TBD we will find out really soon, but I think it's worth paying attention to. The other big launch of today that I think is really interesting is the Agents API is now with computer use. So the Agents API now supports computers. So developers can build agents that interact with software to complete tasks. It also brings codex's multi-agent capabilities, tool search, tool calling, and context compaction into your application. OpenAI runs an underlying infrastructure.

7:08So your team can focus on building applications. This, I've talked about this before around like, you know, vertical harnesses, you can see here, OpenAI, it's a managed codex harness. So, you know, if you have, you know, great data, if you have a great brand, if you understand a vertical better than other people, you know, there's this huge opportunity to build apps. And that's why I would be playing with the Agents API today, like go and figure out, you know, what you can be building, how you can extend yourself, extend your business to this world. And this is just like a new surface area. It's like thinking about how you can create AI native apps. And it's just, I think there's a lot here. I don't think a lot of people are doing it. And obviously that excites me. The other, like, low key, huge announcements, I don't even know if it's, I don't even know

8:10if it's on this page. That's how low key it is. But when I saw it, it sort of blew my mind. But basically, you can now sign in with your chat GVT login and use your token allowance and on those websites. This is going to open up thousands of business opportunities. In the same way that when Facebook launched their login, it opened up thousands of business opportunities. And actually created many billions, billion dollar companies. Think about, you know, Yelp as an example. Like I remember when Yelp came out, it was basically, you know, you can go check out what their early product looked like, multi-billion dollar company that was basically login with your Facebook and you're gonna see local reviews and reviews that your friends have given. And it had that Facebook graph, so it allowed you to do more.

9:12Now, where do I see the opportunity? We're gonna go into depth into this in a bit. But where I see the opportunity with sign in with chat GVT is now you can create a free app. You don't charge people. So you get your top of the funnel becomes bigger. It's connected with your chat GVT and then you can upsell them to other products and services from there. That is a huge, huge, huge thing. Let's go back here. Here's my word dump of chat GVT login. I think a lot of people look at AI apps and they're like, do I really want to spend $20 or $30 a month for another AI app? It's like, no. You can just sign in with chat GVT and start using the plan you already have. So I think a lot of people, if you're smart, you create your core app free and then you charge for advanced features. So you can charge for enterprise stuff. You can charge for sharing with your team. You can charge for more data. But now you have this opportunity to create the core app for free.

10:13In the past, you couldn't create the core app for free because you were paying for inference. But now you can just focus on your business, building the distribution, building the brand, getting the data sets, obviously getting those people from the brand into the app, get them to use the app. And when they start saying value, upsell them to other things. A few ideas around what you can do with the login and agent that cleans architectural CAD files. A researcher that monitors one specific scientific field. A Shopify catalog, clean up desktop app. A video production tool built around one repeatable editing workflow. A contract review tool for one type of franchise agreement. Basically, what's the opportunity? These tiny AI categories that weren't viable, weren't economically viable today as of today are.

11:16So you want to build these products that are too niche to become major AI, open AI features. But you have enough of an audience that you can charge something for upselling and for that workflow. So I think this is a really big deal. And I'm really excited about it. My mind is going absolutely, I can't stop thinking about it actually. And now we can focus on selling workflows, right? Yeah, sell workflows instead of tokens. So if you're ideating, if you want to build a business in this agentic age, what's some proprietary information that you have? A specialist workflow, access to outside systems, collaboration and auditability. A network of human or suppliers, a measurable business result. These are the things you can focus on and we don't need to really focus on tokens.

12:17And I think you're also just going to get a lot more users as your app. You get their emails, you might get their phone numbers. And then you could email them, you can sell ads in those emails. There's just so much more that you can do now that I think it's really, really interesting. I wanted to just talk about like the mental model I have of this new world where we have decisions API, we have login with chat GPT, we have dots, we have the agents API, how am I thinking about building a business in this world? To me, it's this four-step framework, trigger, decision, action, and feedback. So there's going to be a trigger and invoice becomes due, the inventory runs low or it's tax time. Like something happens such that I'm going to chat GPT and I'm asking for some help, or it does it autonomously. Then there's a decision. So chat GPT is like, okay, what should I do next?

13:17Then there's an action. So there's a correct, there's a submit, there's an ability for an application to do something. And then there's feedback. Was it approved, was it rejected? And I think that the strongest businesses are going to own two of three. So they're going to own trigger, they're going to own action and feedback. So when you're thinking in your brainstorming around what you can be building, I would be focused on those three things. I wanted to give you a couple ideas just to get your creative juices flowing around, okay, what other businesses could you build now that Sam Altman and OpenAI team have launched this? One is like a real-world API. So in a world where you're going to have millions of people, basically going to dots and asking it for stuff, some of those stuff are going to be digital. But some of them, some of those are going to be digital.

14:19Some of them are going to be structured jobs. So if you can figure out a way to get vetted specialists in the real world, I think that that's a huge opportunity. So like permanent expeditors, custom brokers, and then you make a transaction fee. So I think my point is there's all this stuff happening in the physical real world. A lot of people are going to be focusing on the digital stuff. How can you create real-world work APIs in specific niches that you can connect dots or connect chat GPT to real-world workers and finish the last mile of the transaction? Huge opportunity throwing it out there. The second is who's going to build the analytics and optimization layer for agent discovery? So I'm not the only one who's thinking about building plugins right now. At this very moment,

15:20there's probably thousands of companies thinking about how the hell do I get? Surface within dots. Not just dots. How do I get services in mues? How do I get service within Gemini Google's platform? I think it's called Spark. There's all these personal agent platforms, instinct. There's all these platforms, Grockbot. How do you get your plugin surface? Some of these have plugins, some of them don't. But I think eventually they all will. If you can create an analytics platform that shows... Actually, I have it here. The Generates Realistic Indirected Proms, which measures which plugin gets selected, which tracks activation share by customer intent, which finds metadata and tool definition problems, which measures whether invocation turns into completion, tests every plugin update. There is a huge opportunity here. In the same way that in the web era,

16:23you had huge companies. Mix panel came out. Amplitude came out. Google Analytics was a huge product. I think there's tons of analytics products that are going to come out around helping you be seen. You actually saw that with AEO Geo. I think profound raised that like a $2 billion valuation or a billion dollar evaluation. Some absurd, you know, large amount. They're very conservative. They're generating a ton of buzz and revenue. And they're solving a real problem. My point is, who's going to be the profound for this personal agent era? Huge opportunity. I hope someone takes that. So, I'll end with they obviously launched other stuff. They've obviously launched GPT 6.1 Soul, which was huge, which is obviously a major upgrade to GPT 6. But we kind of expected that. They obviously launched UltraFast, which is amazing.

17:24If you haven't tried UltraFast, it's absolutely amazing. You're going to love it. It's just way faster. Of course, it's a lot more tokens. Of course, they launched this. They obviously launched private intelligence, which I think a lot of companies are scared to use Frontier AI because their data is really protected. So, I'm not surprised that they're doing things like that and doing stuff like Codex Security Cloud. But I think the point of this whole episode was, there's a few main features. Decision API, Agents API, the login, which is GPT in dots, that I think, you know, if you're listening to this, you're a founder. You're someone who wants to build a company, who wants to bet on yourself. That's going to be huge. So, yeah, and they also launched the spaces, which is basically their version of having cloud slides and docs.

18:25You can bring your AI agent to work with it. So, yeah, it's cool. And it's going to help make people more productive. But for people listening to this, for my people, for the idea junkies, the people who want to build businesses, I thought this would be most interesting to you. Hope it was. Let me know if it was. I read every single comment, YouTube comment. I reply to a lot of them too. And have a creative day, and I'll see you next time.