What comes after serverless? | Hacker News Reader

What comes after serverless?

arcentry.com

24 pointswolframhempel24 comments

I hope we'll see more real serverless with P2P apps instead of servers-in-the-cloud-serverless apps. Maybe call them cloudless?

Maybe call them P2P? Or is that term tainted for some reason?

My guess would be 'languageless' and 'architectureless'. A completely managed, language independent high-level, high-performance (hopefully) bytecode which removes the dependence on communicating with external services VIA IPC/Sockets/APIs/Messages and instead provides native datatypes and service primitives as 'syscalls'.

Not to be dismissive of your idea, but isn't this decade(s) old? QT Creator, Dreamweaver, Wordpress, etc... All are in some form a "drag-n-drop code gen" tool. Could you elaborate on your point and how it differs from the existing codeless tools?

Codeless, as others are saying, is my best guess as to the next big thing. I’m not sure that it’s especially near-term though.

Imagine something like a combination of the following tools: Enterprise Architect for modeling/code generation, FaaS, React for reusable UI components, and SSIS/LEGO interface for hooking it all up.

Clientless. Requests go from your brain to the cloud and responses are translated and assimilated into the brain, sorta like Johnny Mnemonic except wireless! There will be instant authentication - some unique identifier / biometric recognition will make remembering passwords / 2FA redundant. It'll be your brain, logging directly into the cloud.

Instant messaging will be instant. There will be a new speed race among search engines. People will complain about having to wait for Google like they complain about being put on hold by the call center.

There will be a new class of addicts that don't know how to log off. There will be people who rip the chip out and go offline forever. But they might not get the best jobs and government services will be much slower for them.

Clientless is the future! From your brain to the cloud. Expand your mind, maan!

I think codeless comes after serverless.

If the pendulum is swinging back (cloud is mainframe), maybe we'll see something like multics for the datacenter age.

If the pendulum is already reaching its limit with serverless, perhaps, battery materials and 5G will make all the kids rediscover personal computing, and we'll see datacenterless computing with micro services hosted directly on smartphones.

Abacus -> Calculator -> Excel -> Bare Metal Server -> VM -> Container -> Serverless -> Deploymentless -> Codeless -> Humanless -> Machineless -> Godless

Hope this helps, thoughts and prayers!

Hey, if we are not there yet, I volunteer to try the "Godfull" server. Does it work better on Sundays?

Nah, it's always down on Sundays. The other 6 days is 100% uptime.

and customer service is a talking snake.

That only speaks in tongues.

Ah, Perl!

> True decentralization means a necessary cap to the size and interconnectedness of any given node, making growth a cumbersome process, highly vulnerable to disruption from centralized players.

I don't remember this being a problem for Skype's growth.

Combine matrix.org with things like Azure Sphere, AWS IoT Core and Google Cloud IoT Core -- these parallel networks could be in turn be sliced up into aspects. Then, we can go to the next level of networked computation.

Clientless. Nothing is faster than code that isn't run.

my best guess is codeless

Codeless. Zapier style.

Strongly agree with this, why sell a service when you can give your client the tools they need to implement the service themselves with ease. Strapi has been working in this direction, selling-itself as a "Wordpress" for backends.

Something even more buzzwordy/BS-y. Ask Dilbert's boss.

Serverless was never a thing. There are servers, whatever you want to call the them.

Observing coworkers spend weeks setting up serverless stacks, for not needing a server they sure spent a lot of time configuring their tools...

But obviously the next thing is codeless, which is WYSIWYG all over again. Kind of like how serverless was the cloud all over again.

Ofc there are servers; the point is that its server-independent — and ofc theres a setup and configuration to it: certain ideas still have to exist somewhere, if its no longer implicitly known by your server

What’s the point in being pedantic about the marketing naming? If you’re going to lay critiques, don’t swing wildly at the first thing you see; you gotta hit ‘em where it hurts (eg as always, most codebases don’t need that kind of scaling; the code architectural benefits can be had without buying into k8/docker entirely; auto-scaling always offers the risk of a surprise high bill; limited to short-lived functions; server configuration replaced with cloud services/docker configuration; etc)

Who do you imagine you’re trying to convince? A first year cs grad? Your mother? Does you really believe anyone who could make claim to an informed opinion would be so dense as to never think “wait! There must be something underlying all this magic!”

Serverless is not really serverless! And cloud is just someone else’s servers! Have got to rank amongst the most commonly useless statements in tech; it convinces no one and is pure signalling

Thank you :-)

Reply on news.ycombinator.com