Ask HN: When did tech stop being cool?

I think Jonathan Blow made a decent argument that we're in a decadent age of software. In a decadent age, the civilization still does and can do the thing, but has started to lose a grip on the fundamentals, so it's on the way out. A decadent age is a kind of end, doesn't necessarily feel like the end, it kind of feels like a stagnant bumbling.

How many programmers do you know that can write assembly code? How many layers of abstraction between a web app and the silicon? Processor, OS, probably a byte-code VM, probably a framework... leaking a little bit all of the way.

He makes a good point--where's the evidence that we can write good software? It seems possible, but it doesn't appear, so really, the feeling that maybe we could is just a feeling, not evidence.

We have replaced personal computing with ultra-scale industrial computing. Instead of soft, flexible systems we can play with & explore, we have hard, fixed, far off & remote inflexible systems that we are end users to. We're all just chumps with some tiny slice of the mainframe, again.

Computing needs to get back to "soft"-ware. We need to make things malleable & flexible, need to allow people in, let them play with the technology more freely. We need to put actual people back at some kind of helm, let them free, let them explore & learn & fiddle/tweak & build. We have diminishing returns because it's only giant companies exploring & trying things.

We need a new starting place for liberatory/libre technology. Open-source is still kind of stuck inbetween the old FreeDesktop / personal-computer world and the new all-connnected networked age. We haven't really got many future-facing architectures or practices for what software might look like when connected. We have lots of toolkits for building client-server systems, for making ReST & graphql & other... but this kind of software tends to be much harder to run, less hackable/free in form. It risks granting us connectivity, but at the mortal cost of being "hard" not soft, & of asking a lot of our users.

We need both lightweight & fun & hackable, AND connected. How we make connected & hackable viable for end users, how we make it so people can easily get/create persistently available online connected systems- which are the table stakes these days- is very much up in the air, is unknown. More radical p2p is probably smarter/better/essential, for availability & resilience & connectivity/networking reasons. We have some interesting p2p, but it's still so To-Be-Determined what software architecture we'd use in these worlds, it's still so new.

To be honest, the One Laptop Per Child's DBus-over-XMPP system (Telepathy Tubes) is one of the most visionary & still promising interesting dots there is on the grid, to this day. There was a lot of research interest in the past in "mobile ambient" computing & mobile agents... this was one of the easiest to author, most "normal" paths I've seemed that semi-matches up, that was semi-pursuant to what seemed like a clean logical/conceptual framing for online computing.

Personally, today, I still see ActivityPub & ActivityStreams as the thing that's slowly opening the door, that's re-enchanting us & opening possibility. The aughts were a buzz of APIs and connectivity, but the tens were a massive undoing, a collapse, with so much diversity & vanguard efforts falling, collapsing, being rolled back. The internet hasn't had new internet shit in a long time. Getting people connecting with each other in new interesting ways, finding new basis for interoperation, interconnection, internetworking is 100% necessary to break us out of this take-over-by-the-giants, to create something adoptable & growable & enriching again. We need some basis for value creation, and right now, we technically don't have that: everything is hard set, hard fixed, gelled into place, inflexible. New protocols, new systems can help untie us.

But that's just the start- making things re-hackable, making it all soft, making online systems as personable as the personal computer was... that's the real challenge. We need a shared base, with well defined possibility & potential, and we need a landscape of tools/systems/implementations that makes covering this terrain fun & interesting. A new cyberspace will emerge, again, once again connected, but this time not ran/hosted entirely from inside the firewall.

The world was already full of appliances, but computing, for a time, was different, was more engaging: returning to something in which we can genuinely engage & immerse ourselves, learn of it & mold it: that is key. The way we do software right, alas, is focused on corporate platforms for building consumer applications/appliances, and that has made tech uninteresting, uncool, has reduced it to the level of toasters & bread slicers.