With the Replit Agent making fully fleged front-end apps from scratch and Claude 3.7 one-shotting landing pages that would have cost $1000 and two weeks, it is increasingly obvious that we may live in a post-app world. Application layer mass genocide is incoming. Pinterest, Twitter, Instagram, Instacart, and Facebook—born out of the late 2000s/early 2010s—define billions of homescreens. But this may change very quickly and few realize this unless you are addicted to the chronic pain of X.
It is hard to believe that all apps might disappear since they seem so entrenched in our daily lives. But extrapolating current trends make it seem like we will be able to turn any idea into a complete app with a frontend, backend, database, API connections, etc. with the click of a button. If I was the CEO of DropBox right now, I would be freaking the fuck out (to put it tamely).
Will all apps be replaced? Like DropBox, Docusign, with a lowest plan of $10/month, can be replaced since an app to sign PDFs has no moat. There is no intrinsic value in the platform. People would find it much easier to prompt an LLM to one-shot the app than pay $10. These sort of one-function apps may even be created and destroyed in seconds—in response to functionality the user is seeking.
For example, if I want my business partner to e-sign some Delaware C corp paperwork, I'd just text my virtual assistant on my phone. That LLM agent would take the paperwork from my desktop, confirm with me that it is the right paperwork, spin up an app for secure e-signatures with a date, email it to my partner... once it is filled out, text me that the paperwork is filled out and updated in my desktop. This is a major step up from any workflow now yet it seems possible to build over the weekend with a large enough Celsius budget.
I don't think all apps will be replaced though. I expect apps with social networks to be resilient. Although making an Instagram will be super easy very soon—it's just a Google Photos with sharing—how would I convince all my friends to get on the app that I vibe-prompted on my phone? How would I look up a girl I met at the climbing gym? Instagram's intrinsic value is a derivative of the volume of its lake of social capital. Uber, Instacart, Facebook (if Gen-Z miraculously begin to use it after older generations begin to drop off), Snap, VSCO seem to have resilient value, entirely a function of their social capability/infrastructure. Idk about DropBox, Docusign, Wix, or Trello though.
Some people argue that in this world of abundance, we will treasure taste not tools. But the taste they refer to is just mimetic mess, like every landing page nowadays with Serif and Inter... maybe Montserrat if the designer thinks they are quirky. Claude is already spitting out landing pages that the self-proclaimed "taste-makers" and "thought-leaders" wish they made. Imagine it a year from now.
The lower-level protocols and APIs, like Stripe or IP, will be extremely durable and even more valuable with this transition. If the e-signature app my LLM assistant spun up needed a $50 signing fee from my partner, it could just add Stripe functionality leveraging their API, so I get the utility of the service without the pain of reading documentation or signing into their dashboard. Agents will entirely rely on APIs and lower-level protocols, like roads between cities, to be able to transport any task to execution, spinning up any service necessary. All abstracted from the user. I'd even give my assistant a $200 budget to get me a Christmas gift since it knows more about me than myself or my family. Only if I hit $10K MRR.