An invitation to Compute In The Future

Following on from this post which questions one of the ways forum stickers are used, it reminded me of some sage advice from Alan Kay: “Compute in the future”. This means imagining what computing will be like in the future and do your best to try to live that reality now despite the limitations.

I highly recommend reading his discussion of this concept 7 Steps For Inventing The Future.

In this frame of mind, I block all stickers on the forum using my adblocker because the stickers make reading very difficult for me by affecting the vertical positioning of text - this is very similar to how content moderation will have to work on SAFE. I have made the list available if you run AdBlock etc you can import this list to remove stickers from the forum. The url for the list is

I will try to keep it updated as I notice new stickers. Feel free to submit pull requests or issues with updates.

This feels to me a lot like how moderation is going to have to work in SAFE network, building an experience that works best for you by composing bits and pieces together from others on the network who want to experience it in ways you happen to agree with. I was going to keep the list private, I understand it steps on some toes just by making this list public, but in the end I agree with Alan Kay and feel starting an experiment along this path is worth trying.

On the more general point about Compute In The Future, do you have any ideas or suggestions about how we can put SAFE style practices to use right now on the old internet?

One of the main limitations of this solution to the stickers is it only works for those who are subscribed to the list, so the default unsubscribed experience of the forum may not be what the majority might expect. When to use opt-in-removal of content rather than opt-in-addition of content is a tough marketing/moderation/ux question.


This is very useful. One of the biggest problems with human experience is that one gets used to everything. If you eat only chocolate every day, you stop wanting to eat only chocolate …

One of the best ways I’ve found to fight this is deprivation. This tool is useful for me in that when I get used to the stickers and they are no longer making me happy, I can remove them for 1-2 weeks and then return them. So the joy of them will be with me again.

Thank you Mav!


I wish I could do the same with my girlfriend…


I wish I could do the same with my girlfriend…

You can! Women understand that relationships have natural cycles, an ebb and flow. If you want to rest from the grueling grind of it all, most of all she wants to know how you really feel. And you want to tell her, don’t you? Tell her you want a few weeks off because things are getting very routine. Then in a few weeks, you will have a much stronger relationship because you two will realize how much you missed each other. Plus you know you have a woman you can be totally real with.


Friend, that was just a joke. I follow the MGTOW philosophy and I don’t have exactly this problem :secretariat415:

1 Like

New to me but sounds dodgy from Wikipedia entry. Let’s not derail this topic though!!!

I think @mav’s idea about trying to live as we imagine the future we are creating is very interesting so let’s talk about ways to do that.

I’m not personally bothered by stickers, but now I’m spending more time doing than ‘forummimg’ I am following fewer topics.

I do though think that the principle of decentralised curation is important and can be explored with adblock filters and maybe other things. Not long ago I began following Daniel Cuthbert on Twitter when he began sharing block lists (see github) he was making for his home network when he discovered how much tracking was going on. (I recommend following him BTW, he’s is a good guy, a tinkerer, lots of interesting technical stuff to share etc. He also may be interested in creating a system to allow controlled sharing of personal data, but I’m going off topic!)

So decentralised curation is a really good start, and maybe for more than adblock lists.

I’m not sure what other areas we can try to live in the SAFE future but will have a think. This is quite a good way to do thought experiments.


For me the word curation implies primarily an opt-in rather than an opt-out culture, and the more I’ve thought about, the more I feel that is the way I’d like to see the SAFE network going, as opposed to the internet, which I’d characterise as being more opt-out.

Interestingly enough, I also feel like curation is starting to become a bit of a buzzword, like decentralisation. Hopefully those two things will come together and help to promote SAFE!


Perhaps we need a better word than ‘curation’. In this context, for me at least, curation implies manual work, such as librarians carefully filing index cards, or the early days of Yahoo where people categorised content by hand. Looking into the future and at the current exponentials it will be all about automation and massive scale and current curation methods won’t scale to match the need.

I also think it will be about opting in, in this case opting into the algorithms that do the work for us. Hopefully these algorithms will be more transparent than the ones around today which just serve to intensify filter bubbles for commerial or political purposes. Ideally there’ll be room for the personal and also the global in the same paradigm, but that will be about the incentives that can be brought to bear. I think the incentives structure will be key.


I was actually using ‘curation’ to mean exactly what you define it as meaning!

I think algorithms etc. can have a role to play in helping human curation, but I think there is a (perhaps latent, and perhaps not entirely identified) appetite for reducing the amount of sensory information available to us by putting trust in humans rather than maths. I feel like the trust in algorithms is deceiving us into believing that they are objective, or that we as an individual can be objective.


Warning. braindump incoming…

My 30 year view is something like this. We probably won’t use keyboards, PCs, notebooks etc at all as our main way of interacting with the digital universe will be with some sort of wearable, maybe for the sake of argument something like Google Glass with AR.

Assuming our brains aren’t wired to the cloud by then we’ll still face the same problems of limited brain capacity and attention span (we can only really do one thing requiring conscious effort at a time) and a tendency to use mental shortcuts. Assuming the UI is primarily audiovisual + touch, the I/O capacity of us plus our devices will also be pretty limited.

At the same time the amount of information out there will be increasing exponentially, certainly far faster than any increase in our ability to process it as human beings, so the imbalance between the data that controls and shapes the world and our ability to understand it will increase, which is potentially very bad for personal autonomy, particularly if Google or whatever giant corp is around then has a virtual monopoly on what we get to see.

Like it or not, given this imbalance we’re going to depend on some sort of AI to sift through all these mountains of data to find what we need (just as we do already but on steroids), and in the context and timeframe that we need it. Information then will be far more than just data. It will be parts of the physical world.

I was thinking as I was cycling today, what would be really useful for me to know now? First, is that car going to pull out on me, is there broken glass in the road, how fast am I going, what hills are coming up, what’s the weather going to do in the next two hours, and maybe ‘am I enjoying this ride or would I be better cutting it short and doing something else’, and if so what?

This is all realtime contextual information I can take in via my limited brain and UI. Any more info while I’m cycling is overload, and systems should realise that and prioritise showing me what I need when I need it.

Choosing what information I want to see will likely involve choosing an algorithm to find it for me rather than a source, because on a highly decentralised system there may be millions of data sources and I won’t know which ones to trust. Also, because I don’t want to be stuck in a cycling filter bubble or any other bubble which ignores the common ground required for a society to function, I’ll want other baseline information to hand, like a common shared platform of rights and obligations, otherwise the danger is we’ll all disappear up our own digital arses.

There may be 100s of competing algorithms to choose from for each part of life, and sharing the effective ones will be the curation part. The algorithms’ job will be working out which data and sources are trustworthy, accurate and up to date. The tricky incentivisation part will reward providers of good data sources and providers of good algorithms, creating a common ground while avoiding creating new monopolies. That’s something that will require a lot of thought!


I think contact lens are the gadget in the works that could literally take over us.

I will personally create an algorithm that will compute how much of your life has been wasted by listening to algorithms instead of just living life in the physical world among the delights and rough edges of the physical matter and its surprises. I’ll scare people back to jungle and up the nearest tree. Or maybe not, but I’ll try.


You may well be right @jpl, but I’m with @anon28980808!


A couple of other things I do to try to compute in the future.

I use apis to condense and customise the information I regularly seek into a custom UI. I found I was doing the same five things over and over, so I made my own interface from the apis of those services and now the info that used to take 3m to check takes 3s and I have total (and private customisable) control over what is shown and how it’s displayed. It’s like the opposite of the facebook feed algorithm. Is it just a different type of bubble? Maybe. Works for me though. In the future when data and display is less tightly coupled this sort of thing will hopefully become quite normal. What we choose (or don’t choose) to expose ourselves to (or not expose ourselves to) is an important part of who we become and how we engage in the world, I hope more people are able to take this into their own hands rather than leave it at the mercy of profit hungry companies.

The other thing I do is I use extensively. I use it to create historical records of frequently changing pages and data, I use it to link to pages that might change or be lost in several years, I use it for historical analysis of information that would otherwise not be available. The web archive service is incredibly useful and I feel that SAFE will make history-of-data seem natural and obvious in a way that it currently is not. Will that be better than the relatively temporary nature of information today? Hard to say but in certain cases it will be a good thing. Living in a more history-rich internet is something I think we’ll take for granted in 15 years and I try to live a history-rich internet lifestyle in the present.

As for a 30 year vision…?! I’ll have to think a bit more on that!!


This morning I was thinking about your comment earlier about the log analyser when - again - wondering how to display network stats and thought…

Maybe use a console pipeline something like tail -f <logfile> | mavsthing --<select features> | rustplotthing --<show time series of 'chunks, storage'>

I recall there is a rust console plot thing which David mentioned, or one could be coded up using some JS libs and an HTML/CSS component.

Would be a nice little project and easily built upon.


Regarding the tail -f aspect, I had started down a route using ptail to watch vault log files for updates, parse the lines and then push new info via a websocket server to a locally hosted web interface. But I’m still not sure if that’s the cleanest way to do stats. It has the advantage of being fairly interoperable if/when a jsonrpc interface to vaults comes along.

vault with jsonrpc seems ideal for getting / viewing stats, but needs features added to vault. Second best option seems adding more detailed log lines to be parsed by some other tool.

I like the unix simplicity of piping, but also like the simplicity of a ‘click on this exe to view stats’ solution too. Not sure yet on the direction to take for the stats aspect of vaults. But it definitely seems like a useful project.

1 Like

This is a very interesting point. The web, or at least the attention economy, creates a massive bias towards current events at the expense of historical context. It’s not that past information is necessarily unavailable (although often this is a problem) but that it’s relatively hard to access compared with just scanning the surface what’s going on right now, and we’re basically lazy animals most of the time.

SAFE saves all versions of information automatically and if a way could be found to easily attach this history to the present version and present it as a new contextual dimension to everything we do and see, and do so in a way that’s easy and intuitive to consume that could be a very valuable antidote to cultural amnesia.

There are times when real-time information is essential and others where it’s a curse.


I remember one of the early IPFS demos was a photo album that was ‘forkable’. I haven’t been able to find the video of the demo to share here, but the idea was as you are browsing someone’s photo album, you could edit it as you like (removing photos, adding new photos, etc). As you did this, you’d get a new content hash representing your new album. The content hash of the original album of course remains the same. Anyway, just something that stuck with me showing the power and implications of content addressable networks. And I guess since IPFS already exists, it lets us do this future computing today!