This is a banger article that should be referenced in academic papers and conferences about backend. You're on fire.
Looking at your age, kudo's for the article, but I think there is some misleading info in there:
When I say that we don't need a backend, I mean that a lot of applications that we are building right now (SaaS, AI Tools, etc.) don't really need a separate backend.
Separate to what? The frontend? You definitely need something other than the front-end for most applications. Your point seems to be that you don't need a custom built backend, not that you don't need one at all.
So first, How do you really build an application without a backend with Next.js 14? Simply, you use a lot of third-party APIs to do so ...
Wait, you don't need a backend, you only need third-party API's. How is an API not a backend?
Other than that, if you are using disconnected generic solutions you really, really have to pay attention to authorization and sanitization. Great that you can authenticate using a service, how is that going to handle authorization in another? Also, the frontend is not to be trusted, so any solution that puts authentication or sanitization in the frontend is highly flawed.
I have been thinking about this one. It is tempting to go with the all-in-one but what if your api grows and you need to use it outside of your application. I have worked for a few different companies that allowed external access to the API only. I know that it would still be available going this route but managing it may be a bit more complicated.
Do we really need frontend frameworks now, if we have HTMX? - htmx.org
How do plan to run Server Components without a backend?
T Kumagai
idea man
I liked Prisma and actually tried it when it was written in Scala, I think Backend needs to continue to play a role. Logging OpenTeremetry Audit Log AccessLog Visualization of them Scalability Perspectives such as
Other Backup Solutions I would appreciate your opinions on the following.