I have many thoughts on this (some which agree with Anish’s take, some which don’t): 1. If the dominant format continues to be chat, this is not a great experience for third-party developers because of the lack of discoverability. If you want to invoke an app, you have to already know what it's called. OpenAI can try to solve this through some app listing or marketplace, but using the application in context requires people to already know what it is. This is much different than browsing a list of apps on your home screen as the first step in initiating a workflow. 2. Counterpoint, this assumes that chat stays the dominant interface. Something tells me we haven't seen the end of format evolution for AI, especially if you consider the rumors of OpenAI working on a new hardware device with Jony Ive. It's totally possible that an entirely new paradigm for interacting with AI will emerge. And in that future, this will look like a very smart decision in hindsight. 3. If the interface stays chat-first, this will be very bad for startups and very good for established incumbents. If you want to book a room for your vacation, you'll immediately think of Airbnb. But you won't know to use that unknown startup you haven't heard of before. 4. If you're an incumbent, partnering early with OpenAI on this will be taking a very calculated risk. On one hand, if it works, you'll ensure that you're a dominant app/brand on the next big platform. But you’ll have let them disintermediate you and your relationship with your customer. It still might be worth it for the distribution you gain if OpenAI ends up becoming the operating system of the future. On the other hand, you've handed over the interface and the first-party relationship, and you'll probably never get it back. 5. If you're a startup, you pretty much have no choice but to take a bet on this. Some very smart teams will figure out ways to shortcut distribution in this new model, and they will emerge as the incumbents of the future. Just like in the early days of the App Store, figuring out how to stand out now will pay dividends if this ends up working. But no matter what, you should recognize that by partnering, the only thing that’s a guarantee is that you've made OpenAI stronger in the process. 6. They've tried this before with the GPT store, and it didn't work. That might happen again, or it might not. But if you're OpenAI, you have to keep making this bet because the upside case is far too valuable and consequential. If it doesn't work this time, you can be sure we'll see them try it again in the future. (h/t to a few of my friends who I was discussing all of this with recently, they have decided to remain anonymous)