Comment by miguelrochefort | Hacker News Reader

Comment by miguelrochefort | Hacker News Reader

Parent Full thread

miguelrochefort

I'm rethinking the computer. The hardware, the software, the UI paradigms, everything. Xerox PARC 2.0.

- We must stop writing UI by hand

- We must adopt a new language that's less ambiguous

- We must prefer graphs to trees

- We must not build service-specific apps and websites

- We must stop thinking of computer hardware as a personal device, and make switching from a device to another completely seamless

- Local storage must be nothing but a local cache of sections of the ONE universal knowledge base

- Speech and text must not be the primary way to interact with a machine/AI

- Service providers must never have a say when it comes to UI/UX, as services must be completely decoupled from UI

- Everything is an agent, the system shouldn't make a distinction between human users, AI, smart contracts, APIs, etc.

- There must be no difference between creating software and using software

- Money shall be replaced by a social currency, a form of trust score (based on reliability, honesty and predictability)

- Businesses must adapt their practices and models to software, not the opposite

- Brands must be eliminated and replaced with a trust score

- We must allow for non-precise facts to exist (ranges, set of weighted values, conditional constraints)

- We must allow for non-universal facts (opinions, contradicting values, different weight/credibility based on the user's trust graph)

- "Undo" and "Predict next" functionalities must be ubiquitous and present everywhere

- Intentions and Predictions are communicated by logging events in the future (where the date doesn't need to be precise as allowed above)

- A cooking recipe, GPS directions, task dependencies, a playlist, IKEA directions, a tutorial, must all use the same model

- ...

We have surprisingly identical ambitions. BVic's "Future of Programming" is particularly relevent to several of your points and really puts how terrible our current system is in context.... https://vimeo.com/71278954

I feel like this sentiment (of "everything is wrong") is gaining traction recently (last year/two). Maybe it's just wishful thinking though...

Although some people seem to share our understanding that things must change, it usually stops there.

I haven't seen anyone present a radically new paradigm for computing (especially consumer UI) in the last decade.

I intend to change that.

I intend to do that too. Why don't we all get on a chat room and work at it?

Let's do this.

Feel free to join: https://bettercomputer.slack.com

I've been pondering how to attack the problem of making better UIs for going on 10 years now. Half the time I think I'm completely out of touch and that my ideas are irrelevant (with good reason), the rest of the time I wonder what would happen if I actually tried to implement some of the stuff I've come up with. (I strongly suspect I'd get a rude shock and realize how much iteration would be needed to make it usable, heh.)

I'd love to bounce ideas back and forth, but... agh, a Slack URL. I tried Slack once, felt too stifled, and bolted. The poor team I joined still has no idea where I went or why I left.

I'm torn between asking for an invite and just passing because, ironically, the software being used grates too much (XD)... maybe you could enable the IRC gateway?

One thing at a time though; if you're using Slack, so be it. My public email is in my profile. The private email I'd want to sign up under is different.

Me and a fellow student are working on such a system: https://github.com/Symatem/Symatem. We are interested in joining the Slack group: Alexander.Meissner at student.hpi.de and Niklas.Riekenbrauck at student.hpi.de

I started, and because of this post will now resume, working on Automata: https://github.com/theideasmith/Automata.

Add me to the Slack, aclscientist[AT]gmail[DOT]com

I would also love to be a part of this.

Please add me! william [at] williambernoudy [dot] com

Please add me as well

joelg [at] mit [edu]

yes. invite me if you've already set up one?

Let's get together and work on this. There's nothing more important at this time and era of technology.

beep me at sethjgore@gmail.com

Big and bold challenges.

Does the "no difference between creating and using software" not cover much/most of your considerations?

I'm curious where you've taken to the idea of things like graphs over trees, should a new system be so opinionated? or is there a way to develop something flexible enough that when the next thing after graphs comes along, your system will be able to adapt.

I have lost all motivation for lesser challenges. I can't be excited about a very specific project with a doman-specific purpose.

One could probably summarize my mission as to bridge the gap between consuming and creating software. Sometimes, I tell people that I'm creating a new language (which I am).

A tree is just a unidirectional graph. Most file systems use the tree structure which is crippling. I'd like to know what's better and more flexible than a graph.

I think we can build the trust system using a hypergraph. There's also a couple special kinds of nodes that will shorten the walk between seemingly disconnected items. I've been thinking about this for a while too.

> A simple graph can be considered a special case of the hypergraph, namely the 2-uniform hypergraph.

Hypergraph it is.

Reply on news.ycombinator.com