Intro: Off-piste Architecture: Exploring more human friendly computers

Hi All,

I’ve been working with computers for a long time, I love them but they are also not the easiest of things for us humans to use. They are brittle, hard to configure and put a lot of the load on humans to get things right.

So I’ve been exploring how to make them easier to use and more secure. I’m working from the ground up, as the idea of a competent system administrator is baked in from the lowest levels. It is a long term project, but should hopefully make people less reliant on big internet services like Facebook and Google.

Blog is here:

The source is here:

I’ll leave this here while I explore the rest of edgeryders community. You look like you have lots of interesting ideas.



Hmm what??! :astonished: Computers that manage themselves with market logic. At the very least, this is a really creative idea, and I like it for that. Haven’t seen even a remotely similar combination of concepts so far.

That said, I am not so sure it can make computers more resilient against malicious code, as it is more complex code than current resource allocation systems and, even if the code contains no hole, then market logic itself could be attacked, (Exploiting market failures happens in the real world all the time.) I’m sure you thought about that though. What is your take on this?

And as a last note: I see you have been working on this for years, but it might be that “the stars aligned” right now for you to get significant funding for this. Computer security and digital currency systems are both hot topic right now, and you might want to capitalize on that by doing a cryptocurrency ICO …

1 Like

Hmm. Looks more like biology than economics. “The bot is given some resources…” could be “the bacterium is given some sugar…”. But then, there are many models where econ and bio borrow from each other (or “exchange genes”?).

Anyway, @eb4890, it looks like worthy work. How did you get the idea?

1 Like

There is plenty of code to allocate resources in current computer architectures (if you include cgroups and the like). The code for my market is pretty simple, It is just a single file.

Every asset is rented so some market failures should be harder in the system than the real world markets. The user also gets to radically alter the market by giving good feedback for different things, which is harder to do than in our economies. There will still be some market failures, but hopefully it will still be a more humane system than the one we have at the moment.

It is part of the reason I am presenting this as trying lots of different types of architectures, I would like us to be trying radically different computer architectures to find the best one, with my architecture just being the first thing to try and something to spark people’s imagination.

I’ve been mainly working on a prototype for the system and a demo of it, so I’ve not been doing much spreading of the idea. They should be ready in a couple of months, I’m planning on making it easy for people to pick up and play with.

I’ve been thinking about ICO, but I am put off by the waste in proof of work cryptocurrencies. Also I think my priority is to get other people interested and thinking about what a new computer should be like (even if the market idea doesn’t pan out), rather than raising funds. I’m not sure an ICO is the best way to do that,

Thanks for your interest!

1 Like

I got the idea a long time ago. I was doing a machine learning course and was dissatisfied with the limited systems. So I got interested in the idea of making something like Tierra (a very free evolutionary artificial life system) but useful for a human. I got interested using bidding and markets and then found agoric computing,

A few iterations later, I think that evolution is less important, we don’t have the patience to sit through lots of iterations of programs. I still think the market still has a place in normal computer systems programmed by humans, it can protect people from malware, bad software updates and other bad programs that don’t do anything useful.


W0000t! Computers, I’m in!
I’m not so advanced (I preffer to be humble than to disappoint you. The fact is that I was “playing” with computers since already more than 3 decades (32 years since I discovered Apple IIc and 23 years since my first 486 with 4MB of ram…

Time flies, I ran an IT Bachelor since already 4 years (still 2 years needed to complete) and last year I got the chance to learn Data Science (6 modules in 7 weeks and 24 more days with teachers)

Well I guess this is the critical moment where mankind have really to learn and understand what actions he will do because its future will depend of any of our choices.

We can use, merge and play with different technologies, Machine Learning, Deep Learning, AI, bots, blockchain, vr, IoT, etc… things are changing fast, and if we could say 20 years ago that timeframe of obsolescence was 5 years, now it’s 6 months, probably 3 months within 5 years.

Now we can find solutions that already exist rather than building one ourselves, there are a lot of opensource alternatives and hundreds of different solutions for any big leader like Google, Apple, Amazon or Facebook. Yet we have to explore it and to keep a clean database (fast fast I said… The solutions that exist now in 2018 could change/merge/disappear in 2020).

For your topic I guess the best will be to make a virtual machine of a configuration, and it could be nice if there is in that configuration, in the back-end, an AI that analyze all the software installed and that could see (and check, updating with security website) all the potential security flaws.
If by instant you run a version 10.1 on a software, and the software website announce an important security update, I guess even we can imagine a procedure that updates automatically this software (creating a restoration point to backwards in case.
Just because all the users doesn’t update automatically their software, some of them may even not have a system to check on internet if they have a future version.

Well I hope this post doesn’t make you afraid, for non-IT guys…
What I saw these last 6 months was amazing. We should really care what we can do with technology and know how to prevent worst case. Because it’s not anymore “black mirror” (cf bees drone ep.), it’s happening right now. I don’t want to make you afraid but I’m sure it’s already the reality. Look at this video:


Thanks for the enthusiasm @oliiive.

I think we do need to understand the landscape of different technologies. I think in the very long term we should aim for things like centaurs. Which will involve AI and machine learning.

We still need to build things, because there is plenty of things that are not open sourced or there is not a huge profit motive to build.The open source speech to text isn’t very good for example (last I tried it)

We do need to think about downsides of all this tech, I’d rather not focus too much on them, I think things currently are over played. We already have hard to trace ways of killing people like poisons, but they are not in everybody’s day to day experience. Nothing about the future is going to make everyone more violent.

There are potential societal downsides, if the benefits aren’t shared equally.

About a program looking at a system to detect out of date software, we already have that. I’m hoping to use things like capability based security a long with the market to make out of date software on your system less of a problem in general. Things will become clearer in time hopefully!

This article roughly covers what I feel about the video. It was very overblown.

1 Like

Thank you for your replies, it’s much than appreciated. Because as we have now the technology, it seems that a lot of people could use this technology for bad purpose, as it was always the case in the history.
Last year I had the chance to follow a Data Science training and it was amazing. I discovered some places where you can know more about all those tech stuff, DSci, AI, VR/XR/AR, Blockchain, …