That is why I am betting the next big unifying platform will be an “infra-app” that streamlines the DX across multiple vendors versus a “super-app” that consolidates the UX across multiple services.
I feel there are control point possibilities at both layers, at the workflow layer if you own the core decision or at the integration layer if you build a solid operating layer with good DX.
Always insightful posts. I think we can read this phenomena, as a general consequence of reduced transaction cost in composing composable software. In a sense, more users are becoming developers, and therefore developer experience becomes more important. But the dynamic that see one platform create a solid core UX around a certain workflow, and then use a distribution play to attract third parties to create extensions because the two products work well togheter I think will coexist.
I think you've got the zeitgeist though which is:
1) vertical integration will have harder times to serve large scale (composing niche solution is cheaper)
2) increasingly DIY integration is competitive in delivering a great UX
As a result, will become more difficult to extract capital (for example from VCs). Web3 protocols (and hypertructures) go exactly in this direction, I was talking about this a few days ago with Jesse Walden - a podcast episode is upcoming.
Yes, the two will absolutely coexist. I just think that a lot more of developer opportunity will shift to developer-led. Getting developer-led right is difficult. You need to find an initial use case the way Stripe did of facilitating two-sided payments in marketplaces.
I buy that.
That is why I am betting the next big unifying platform will be an “infra-app” that streamlines the DX across multiple vendors versus a “super-app” that consolidates the UX across multiple services.
I believe there are possible plays at both layers.
I understand your reaction to the term super-app, my point there is a workflow control point. I explained it here: https://platforms.substack.com/p/how-to-win-at-generative-ai
I feel there are control point possibilities at both layers, at the workflow layer if you own the core decision or at the integration layer if you build a solid operating layer with good DX.
Always insightful posts. I think we can read this phenomena, as a general consequence of reduced transaction cost in composing composable software. In a sense, more users are becoming developers, and therefore developer experience becomes more important. But the dynamic that see one platform create a solid core UX around a certain workflow, and then use a distribution play to attract third parties to create extensions because the two products work well togheter I think will coexist.
I think you've got the zeitgeist though which is:
1) vertical integration will have harder times to serve large scale (composing niche solution is cheaper)
2) increasingly DIY integration is competitive in delivering a great UX
As a result, will become more difficult to extract capital (for example from VCs). Web3 protocols (and hypertructures) go exactly in this direction, I was talking about this a few days ago with Jesse Walden - a podcast episode is upcoming.
Yes, the two will absolutely coexist. I just think that a lot more of developer opportunity will shift to developer-led. Getting developer-led right is difficult. You need to find an initial use case the way Stripe did of facilitating two-sided payments in marketplaces.