But who will collect the garbage?

Well I'm not doing it!

This post is in response to a question on Google+ by a friend “Kevin C”, when talking about a world where we don’t have income inequality to force the majority of people to work for the minority. The question amounted to, “who will collect the garbage?”

I’m going to use the term Adhocracy, by which I mean the adhoc, leaderless, tech and network powered organisations / events / revolutions / actions which taken together form a nascent world system, one which is challenging the old institutional/corporatist world system. Think wikileaks, wikipedia, the Arab Spring, Anonymous, flash mobs, facebook causes, twitter hashtags, the open source movement.

Firstly, one of the really important parts of the adhocracy is the shrinking need for people to do real work. Lots of old industry was replicated geographically, for instance, so needed armies of people doing the same thing, spread out across the globe. eg: people working at newspapers, soon people working at universities. A lot more stuff is/was replicated per org (HR? much software? record keeping/admin? middle management?) What all this amounts to is armies of people doomed to do repeated, endlessly mediocre work, like a world spanning fractal turd.

The adhoc way doesn’t need these armies of the mediocre. Instead, highly competitive reputation competitions arise in every niche, and the top couple of entities wins everything. So for example, being an Exchange administrator is a bad idea when hosted email (eg: gmail!) is taking over.

The technologically mediated adhoc setting is one in which the efforts of the best few are multiplied vastly more than ever before by a technological support structure, so that one great solution run by a few people for a niche problem can support the entire world. In aggregate, we simply no longer need the same number of people to do the things that need doing. What’s more, profits razor down very close to cost, because how much does a tiny group really need to take for their efforts, even if they have hundreds of thousands or millions of users? If they try to take more than the minimum, someone else eats their lunch. That’s something the big corporates struggle with; they need orders of magnitude more profit just to support their bloat, and it’s why the papers or the record industry or the book industry or soon the universities fail to adapt and instead collapse.

Secondly, most of the commercial work we do now is actually entirely unimportant, it’s waste. I don’t want to try to defend that statement here, instead read this paper by Kevin Carson, “The Great Domain of Cost-Plus: The Waste Production Economy” ( http://c4ss.org/wp-content/uploads/2010/12/Political-Economy-of-Waste.pdf ) . It’s an excellent, deep study of how what we do is mostly makework, and how wages are kept artificially low to keep us working longer and harder. But really we all see it; how much of your own work is ultimately productive in some meaningful way? I hope for me the number is higher than zero, but I can’t guarantee it.

So taken together, the first two points say that most of what we do now is totally unnecessary and wasteful, and of what does remain that needs doing, we’ll need less and less people in order to do it as time moves on. So who will collect the garbage? Very few people indeed, with excellent automation.

How do we support ourselves without work for pay? Hard question, just assume there’s a way for now, and I’ll come back to it.

But what of the work that still needs doing? Who oversees those robot garbage trucks? Well, if we don’t all need to work in order to live (as we’ve just assumed) then you’ve now got a lot of people looking for something useful to do, something with meaning. So volunteering will fill the automation gap (the stuff we can’t automate fully).

I’ve been involved with a choir of retirees for the last couple of years, and it turns out they’re an excellent model here. They’re fit and healthy, they don’t need to work to live. But, they don’t just sit around doing nothing; they all do things, they all volunteer in one way or another. But what distinguishes this from paid work is that they chose what to do based on their sense of what needs doing, what is meaningful, rather than what pays well. So, I’ve never seen an instance of a retiree working as a volunteer IP lawyer, for instance, or a volunteer advertising copywriter. But they do things like looking after overseas students who need local friends, or helping out refugees, or running the counter at the local op shop, or looking after their grandkids while mum and dad are out being bankers and plastic surgeons and telephone sanitisers.

If the robot garbage truck needed oversight in the local neighborhood, and it was being run in a volunteer capacity, there’d be a waiting list of retired blokes looking for something solid and meaningful to do, volunteering to do it. After all, we all know the garbage needs collecting, and there’ll be kudos for those who step up.

Now these guys wont work like a full time grunt. They wont overwork, and they wont take orders from a foreman. But, they’ll get what needs doing done, if they have the possibility of Autonomy, Mastery, & Purpose (see Dan Pink). How will they organise? Well, this is an adhocracy, after all; the answer is, frictionlessly and effortlessly.

A society of people no longer bound to paid work wouldn’t face the problem of how to get the thankless jobs done, in my opinion. It’d have a potential problem more like Mega City One; not enough meaningful work for all the people.

But all this hinges on us not needing to work. So I need to justify that.

This is a tough one. This is where we really need the adhocracy to rise up as an alternative to the corporatist system. After all, the corporate system has us working more than ever, doing increasingly meaningless crap. But the Adhocracy is rising to this task even now.

The more adhocracy we have, the more we’ll notice that the things we need are becoming extremely cheap or free. We have that now if we choose to take it with respect to information goods; you can survive with a cheap chinese netbook and a low end broadband connection. You don’t need to buy books, music, movies, software, etc etc, if you choose to use the freely available alternatives.

The challenge is to get some of the more material basics running this way. Food, clothing, shelter, utilities.

Some ways I can think of:

  • The sharing economy is massively underrated, but is slowly become a force through pure tech mediated adhoc. Car sharing, tool sharing, people giving away their unwanted stuff, house sharing, couch surfing, bike sharing, on and on and on it goes.
  • Second hand stuff is crazy cheap and getting cheaper. EBay, et al, have revolutionised this area.
Some more sci-fi stuff, but plausible over the next few decades:
  • Power could drop to ridiculously low prices if the solar economy can get off the ground. How cheaply can we make halfway decent photovoltaics if we really push on it? What else happens to the economy if we get cheap, decentralised power? Adhoc type things.
  • The adhoc solution to food is for everyone to start producing it locally, in all kinds of crazy new ways, ways that can beat the prices of centralised production. If power drops way down in price, then food can too. What can biotech do to help here? Can we invent some self-contained food producing units that just require sunlight, water, air to make food locally (more efficiently and reliably than plants)? If we’ve got decent 3D printing, then we just need some basic local gloop production than can be feedstock for food printers.
  • Clothing, similar to food. Plus, if you’re prepared to go second hand, clothing costs next to nothing already.
  • This also covers a lot of appliances/physical objects. Can we print electronics?
  • Shelter is really, really hard. We have a very strongly self-reinforcing enclosure of all the commons. Might have to wait to see what kind of adhoc internety solutions come up in this space, hopefully not bloody revolution, although that’s becoming increasingly common.
Also most notable here is that this is the perspective of someone from the wealthy western world. We might just get the benefits of the adhocracy last, as we have the most to dismantle. Does some of this stuff emerge more easily in China, or India, or South America?

I think this is all ahead of us, and it’s hard to see how it’ll play out. I do think though, that the adhoc world system will begin to control more of the world’s resources, and the capitalist system less. It’s the more dynamic system, resources will flow into it (just think of the corporates paying money, donating time and equipment, to the open source software world on which they are now dependent). What will it do with those resources? Whatever it is, it wont be financially mediated.

As this happens more, the effect will be that, if you choose to take advantage of it, the cost of living will become lower and lower, because there will be non-monetary ways to get the things you need, ways that don’t currently exist. As the necessary monetary price of a decent life falls, and the need for workers diminishes (so unemployment rises), as some point it’ll be clear that governments can bridge the gap by bringing in a universal basic income to pay for the few remaining things that require money.

Also, as more and more people slowly liberate more and more of their time from the work-for-pay grind, you get increasing cognitive surplus pushing into the Adhoc economy. That’s an accelerating factor; they’ll tend to be pushing more capability and resources into the Adhocracy (making it better). That’s a positive feedback loop, which there’s no equivalent for in the money-based world. Just think of what happens every time an important open source software author manages to get free of paid labour – they help accelerate the process.

So in summary, who will collect the garbage?
– We’ll need less garbage collectors due to tech improvement, and removing garbage collection where it was unnecessary.
– Increasingly people will find themselves “unemployable” but also able to get by. They’ll have time on their hands.
– What still needs doing by people will, if it really clearly does need doing, will be done by those people with time, voluntarily. They’ll fight over it.
– If the job needs doing but is incredibly awful, volunteers will work to modify things so that awful job is no longer necessary.
– As this trend continues, there’ll be increasing amounts of people with increasing time on their hands all clamouring to help collect the garbage.
– There wont be strictly necessary-for-survival work for them all, but there’s more to life than that. Expect more people to live their music/dance/art dream, less to give out parking tickets.

A really important thing right now is to see this coming, shift your values toward accepting and embracing it, and begin to evangelise it. The capitalist system holds us in its thrall largely due to pre-existing memeplexes colonising our minds; ie: we are prisoners of our own unexamined values. If you believe that you’ll always need money to live and there is no other way, then that will be true. But if a lot of us can see the better way that is possible, it come, and it’ll work for us.

But who will collect the garbage?

Calling all Autodidacts…

At the bottom of this post are a bunch of questions for autodidacts. Feel free to skip my blathering and go straight there.

I’ve been writing here and there about supporting deep learning on the web, something that’s not been adequately addressed anywhere to date.

My previous posts:

Today I’m thinking about this part of my second post:

“I realised that this is not an idea that makes sense to teachers, who like carefully curated courses that teach whole areas at once, to students who just accept what they are being taught. Rather, this is a system for autodidacts, which should be constructed by autodidacts. For learners, by learners.”

and this part

“So it begins as a personal learning tool, the autodidact’s friend, and builds out into a crowdsourced deep learning knowledge base. This also satisfies the vision “for learners by learners”.

So, what do autodidacts need? What a tricky question! I would consider myself in this group, but that doesn’t mean I know all about it. Is there even a profile of an autodidact? How similar are we? What kinds of dimensions do we vary along?

One guess: Specificity. Some autodidacts will be extremely general, carving their way through any and all knowledge as their muse takes them. Others will be specific, confined perhaps to a single discipline or two. A lot of software people are in this camp, totally autodidactic within the IT / compsci realms, but much less so outside of that. This will in fact be a continuum; people will fall somewhere on the specific <-> general line.

How about process? Do we all use the same one? I tend to be driven by a project focus, usually containing a question. “How can I understand the class of techniques used in aural digital signal processing, specifically related to the human voice, so that I can make construct my own novel implementations?” or “Why is the internet oriented toward shallow learning” or “what parts of our culture, that we take for granted, are actually supremely weird, and how did they come to be that way?”.

As I think more about this, I realise that my process is mostly unexamined; I’m not really sure how I decide to proceed. I could improve on that. But on reflection, some techniques are:

  • I try to “feel” my way through material. There’s a sense of flowing, like water finding the lowest path. When there’s too much I don’t understand, the flow is obstructed. When I feel that happening, I back up and see if there’s a route around the block. It’s expensive to have to go back through dependencies, learning about something more basic before you can then progress through advanced material. But that’s still better than not realising you need to do this, leading to loss of traction, and often a loss of motivation; that’s a way you can derail yourself, and end up failing.
  • Sometimes I don’t even know the name for the things I’m trying to learn. For instance, it took me *ages* (half a year at least?) to learn the term “digital signal processing”, and that was a giant block to my inital progress on the Esteso Voce. What I do when I’m so ignorant that I don’t even know which field contains the specialists who could point me in the right direction, is to ask around. To that end, I tend to cultivate networks of ridiculously intelligent and well educated people, who know lots of stuff and like to talk about it. Social Networking has been brilliant for this, but prior to that I used the extropian chat list (an intellectual powerhouse). And of course I have friends in rl, too, who I lean on, but you can’t beat the weight of numbers in online fora.
  • I try to read a lot of varied stuff. Sort of priming the pump? You can’t have interesting ideas without raw material to work on.
  • I don’t horde materials, although I know a lot of people do. Rather, I try to collect ways of refinding information that I’ve seen before. Books that I can get on pdfs I tend to upload into Google Docs so I don’t lose them. Probably my best current resource is Google Web History (https://www.google.com/history/ + the chrome extension “Google Web History Updater”) which lets me search only on what I’ve seen before, like a commonplace book but everything goes in, without me needing to think about it or take any action.
  • I write. Writing helps me get my ideas in order, and keeps a log of complex thoughts that I’ve had, so I don’t have to go through the process again. Rereading my blog often gives me ideas, and sends me spinning further down whatever path I had been travelling. So even if no one else ever reads anything here, the blog is incredibly useful.

I can’t think of much more along those lines at the moment.

Another useful question might be, do I fail, and why?

I fail *a lot*. Many big questions are just so hard to penetrate without a background in the right disciplines (whatever they may be, sometimes I can’t even find that out). The less pre-existing relevant background I have, the more likely failure seems to be. Also, if I try to forge ahead through areas I don’t actually understand, it usually ends in failure, as I’ve noted above. If I can identify background knowledge I need, but it’s too onerous to get the bits I need (often true of specialised academic areas, where they structure the knowledge like a fortress to keep out the infidels), that can be failure.

I think I also fail when the things I have to learn have too many unknowns, and the dependencies are too complex. I tend to approach these things a bit at a time; make a bit of progress, drop it for months, come back and try a bit more. If the endeavour is too complex, it can be too hard to do it piecemeal.

Sometimes I fail to penetrate a particular field because I come at it with incompatible cultural assumptions. Even related disciplines can be very far apart culturally. Digital signal processing has been tough, not least because I think like a software developer, but they, even though doing everything in software, think like electronic engineers, hardware people, and to some extent mathematicians. Those ways of thinking are wildly divergent, so it can be very difficult to understand the texts.

The Questions

So that’s me. But I need more input. If you consider yourself an autodidact, whether specific or general or inbetween, I’d love to hear about your experiences and approach. Some specific questions:

  1. Where do you lie on the specific / general continuum? If there are areas you are more comfortable with, what are they? How much difference do you find between your well known areas (perhaps where you have a degree?) and those you don’t know?
  2. What’s your motivation / how do you initiate? I think my motivator is questions in service of a project. Is that true for you? If not, what’s your thing?
  3. When you know your target, what kind of process do you use to get there? Are you aware of it, or is it largely intuitive?
  4. How do you solve the “I don’t know what I don’t know” problem?
  5. Do you record your progress? What sort of tools do you use?
  6. Do you talk to other people much, or confine yourself to written materials?
  7. Do you use esoteric knowledge sources, like academic journals, or is it mostly Google? Books? Blogs? Wikipedia? Anything else?
  8. Do you incorporate structured learning materials? MIT OpenCourseWare? Actual enrolment in courses of study? Or do you find structured courses and materials intolerable?
  9. Are there tools you use to help? Mindmapping? Diary/Commonplace Book? Notebooks? Webpages? Blogs? … Where does this fail you, what would be better?
  10. When do you fail, and why?

I’m awaiting your reponses with baited breath (should have brushed my teeth). Just comment wherever you see this, or if in doubt then comment on my blog. Please feel free to ignore some or all question, suggest and/or answer your questions, or just say whatever.  Thanks!

Calling all Autodidacts…

Bucking the System: The Netusian project

I’m fed up with it.

I love the net, and I love being a netizen. We’ve got access to basically all of the world’s knowledge, most of the people, and can even see out of the eyes of a growing distributed cyborganism; maps, remote sensors, massive datasets. And entertainment, w00t, there’s great stuff!

But, of all the evolving that the net is doing, there’s one path I’m hating; I’m hating the intrusion of geography, especially in the commercial sphere. There’s a lot of talk about the balkanization of the internet, especially in the American press, referring to the limits other countries (particularly China) are placing on their citizen’s internet access. However, my experience is that the US is one of the main creators of this geographical divide.

From where I sit, there are two classes of netizen emerging; people inside the US (first class citizens), and everyone else (second class citizens). As a netizen it really bugs me that I can’t touch things like Google Music, Google Voice, Hulu, Netflix, Spotify, Pandora, Last.fm, Amazon Android Market. As a developer playing in the social space, it’s even more irritating; it leaves me ignorant of technologies and services that I need to understand. Not to mention missing out on stuff like the cr-48 program!!!

In the past I’d sign up to things simply saying I came from the US. I’ve come from Beverly Hills, zip code 90210, a lot of times (hey, I don’t know any other zipcodes!). But that tends not to work these days. Nowadays there’s lots of checking my incoming IP address against geographic ranges, and/or using existing profile information in, for instance, my google account, to lock me out of things.

Why? Why would you bother to do this, people?? What’s so threatening about an Aussie using your stuff? Do we smell or something? (I know the actual answer, btw, it’s Regionalization in the service of Market Segmentation. Australians always get put over a barrel in the name of market segmentation. Marketing dudes, screw you.)

Anyway, I’m not content to be a second class citizen. It’s time to play a little harder.

After giving it a lot of thought, I’ve decided it’s time to create a sock puppet US based netizen identity, a Netusian (Netizen + US? get it? yeah I don’t need your approval, grrr).

My netusian’s codename is Buck System (credit: my darling wife!). Obviously this isn’t the real identity I’ll be using, because there is some amount of subterfuge necessary in doing this.

And subterfuge raises an important question: how far does this go, where is the line? Obviously creating a new identity is something that can venture into areas that are at least legally, what, questionable? What I’m trying to achieve is quite banal, quite minimal; I want to be able to use largely free online services that are only available to people in the US. Buck System isn’t trying to do international drug deals, he’s trying to maybe watch a bit of streaming TV.

So how do I create Buck? Here are my thoughts so far:

Firstly, I need a US IP Address. There are a few easy ways to do this, and they’re unsatisfactory.

  • You can use a free open proxy, route all your traffic through an unknown third party. Awesome way to set yourself up for really bad network performance and some hardcore man-in-the-middle action. No.
  • You can pay for a US based commercial proxy. I’ve used WiTopia in the past, it’s pretty good actually, but the problem is that they are visible and known for what they are doing. So, for instance, Hulu blocks (or at least used to block) WiTopia addresses. That said, I’d recommend it to anyone who’s not seriously technical, it’s a good option.

The way I think I’ll go is to get myself a US based linux (virtual) server and install some VPN software on it, then set myself up to connect to it and route all my traffic through it. Compared to Australian domestic internet prices, server traffic is practically free, so if I can get something with a fairly low setup and ongoing base cost, it’ll work well.

What I’m tempted to do is to make an Amazon EC2 based machine image to do this (VPN). The advantage to this is to only spin the environment up when I want to use it (which should be rarely), so it’ll be really cheap. If I do this, I’ll make the image available on request.

Probably I also want to keep my browser environment clean; cookies need to correspond to Buck’s identity and not conflict with my normal browsing. So, the best approach on my own machine is probably to run a virtual machine (using VirtualBox), running a clean copy of Ubuntu, and only act as Buck through that VM. I could also make that VM available to people.

The second piece of the puzzle is setting myself up a US based online identity. Now this isn’t for real fraud or identity theft or something else horrible, it’s to be able to use some online services. So there’s no need to go to town here; a minimal effort should suffice. I’m thinking Buck will need a google account, a facebook account, etc etc. He might need a little bit of posting history here and there, some online friends perhaps? How far should I go here? Some feedback on this aspect would be helpful, let me know what you think.

To conclude, I’ll just reiterate that I’m not trying to commit fraud here, I’m trying to do some victimless websurfing. I will probably be violating some Terms of Service agreements. Would this put me in legal peril? What are the legal issues if I make a false Google profile, or sign up for Hulu using that profile?

If the answers to those questions aren’t too damning, I’ll forge ahead, and I’ll blog here about my progress, and Buck’s progress. Stay tuned!

Bucking the System: The Netusian project

Crowdsourced Credentialling

We’re in an age where the old institutions are being eaten away from below. They look solid, undefeatable, until they suddenly collapse. The music industry has been fun to watch, the newspapers too. Who’s next?

There are many, many candidates. Here I just want to talk a bit more about Universities, which I think are currently in trouble, what with a higher education bubble, $99/month online universities with credentialing, free online video learning as championed by the Khan Academy, and now a free online AI course from the world’s best researchers via Stanford with some credentialling with 58,000 signups and counting.

Old economy industries are vertically integrated hierarchies rooted in high transaction costs, which tend to be geographically replicated (X many in each city).

One part of why they fall apart is that the cost of being hierarchical and vertical no longer pays, because all the separate pieces of value become able to stand on their own. Usually, there is one thing they do which provides all the revenue for the rest, and when that goes, the industry falls.

For the music industry, it was people buying cheap-to-produce media for high markup. Piracy was the scape goat, but iTunes really smashed them. For the newspapers, it was the demise of classified ads (most famously in the US this was replaced by Craigslist, who destroy a billion in newspaper revenue for every million they make).

For universities, it’s undergraduate degrees; these are about milking the middle class to pay for all the other things a uni does. We’d like to think it all goes on cutting edge research, and maybe some of it even does.

The other part of why they fall is when new solutions turn up that make geographic replication pointless. Examples include CD and DVD stores vs iTunes, local newspapers and their classified ads vs Craigslist/ebay/all the rest of the commercial ‘net, and bookstores vs Amazon et al.

For universities, geographical replication is standard, but all the new competing forms exist online. That competition is in two major camps: upstarts, and the top unis (Stanford, MIT, etc). What happens when geographical replication fails, is that the institutions with value propositions other than locality (eg: excellence, status, uniqueness) can dominate, but all the others fail. That’s what why MIT OCW works; accelerating the geographic replication decline pays for the survivors.

I’m interested here, though, in one of the de-integration aspects of this process, as they relate to universities. As far as I can see, unis provide students roughly these pieces of value:

  • learning (largely replaceable with free online content / study guides)
  • networking (replaceable online, in fact a lot of why nerds built the net in the first place)
  • credentialing – this is still the hard one

Credentialing is the force behind the higher education bubble. People pay more and more to get that piece of paper. It’s an unjustifiable, unproductive, exploitative money pump. If you could route around that, you’d blow this industry to pieces.

Now one way to split credentialing off from the rest of the concerns of “education” is to provide something like “recognition for prior learning”. All kinds of institutions, like TAFE in Australia, or like lots of little accreditation bodies, dabble in this already. But, it’s tough; you have to test people rigorously to figure out if they deserve a credential or not, and you can easily make mistakes. That’s why we prefer the Unis, because we know the person had to more or less sit through X many years of study, so there’s some minimal learning assumable even if everything else fails.

But I’m wondering, can we crowdsource credentialing?

Take a social network, or even better a professional network like LinkedIn. Let people just add “qualifications” they have (“skills”? Is there a more appropriate word?). Then, crucially, get others to rate them.

To make this work, you need some kind of credibility rating for the raters. Credibility should be domain specific, and also person specific:

  • if the rater is an expert in field X, they have higher credibility in field X
  • if the rater has worked closely with person Y, they have higher credibility with respect to rating person Y

But also, a rater should have a general credibility factor – do their ratings match reality, or are they bullshitters? Maybe higher for more ratings, lower for ratings outside area of expertise, modified by how much their ratings match other people’s ratings in the context, maybe lowered for complaints registered about them.

How do we know you’ve worked closely with someone? It’s on the resume. Hard to game this without it being very visible. How do we know you’re an expert in X? Because of your highly rated “skills” in X, calculated as above. And so the loop closes.

The resulting skills above aren’t really credentials per se; they’re a little fuzzier than that. However, we don’t *actually* want credentials. We want a way to pick the best person for the job from a crowd, or determine whether a given person is capable of some given thing. This approach potentially works better than credentials.

That’s all I’ve got for now. I’d love to hear more thoughts on this.

(Massive credit to my darling wife; this emerged from a Saturday morning coffee conversation.)

Crowdsourced Credentialling

You Are My Sunshine – done Jodie and Emlyn style.

Jodie and I have been in the recording studio, today and last Sunday, recording some 2 voice a capella work. We’re working on an album-ish thing called Mullach Abu. (What do you call it these days when it’ll largely be online? Not really a CD??) In any case, it’s coming along pretty well I think!

My favourite so far is this, “You Are My Sunshine”.

We’ve put it into the minor, and moved it rhythmically from 4 to 12. The rhythmic aspect was really crucial.

Next up is Kumbayah.

This again is in the minor. It’s amazing what that does for these otherwise difficult to like songs.

Both Kumbayah and You Are My Sunshine were done in the studio with us facing each other, fairly close together, mics between us and back to back. They’ve been done in one take (there’s really no way to do it piecemeal).

Last week, we were further apart, using headphones, and the result was a bit too careful and the tuning wasn’t quite locked in. Here’s us singing Black is the Colour last week in this fashion.

Neither of us were really happy with this, it’s good but not great. So, this week we did it as described above, no headphones, and we just let it rip.

We watched “O Brother Where Art Thou?” a couple of nights ago, and I was inspired by the Fairfield Four singing Lonesome Valley as gravediggers near the end. Sing big, let the vibrato go wide, don’t be careful, and you can make magic!

You Are My Sunshine – done Jodie and Emlyn style.

Follow up to Deep Learning 2.0

In my previous post Deep Learning 2.0 I outlined a system for enabling deep learning on the ‘net. The system consists roughly of:

Know: A unit of knowledge. A short name and description of something someone can know
Learn: A “recipe” for learning a know. Information might be inline, or references to external sources (eg: MIT OpenCourseWare, Wikipedia, blogs, etc).
Person: Someone with a set of Knows, who can undertake Learns to get more Knows.

Knows have a set of Learns that can lead to them. Learns have a set of prerequisite Knows that you need before you can acquire the Learn.

The collection of Knows and Learns is crowdsourced. The links between Knows and Learns are weighted by some kind of voting system (probably a simple Like system).

Some new thoughts on this:

1 – By Learners, not by Teachers.

In early online discussions about this concept, a lot of objections were raised by professional teachers. How can you crowdsource learning paths? There is no concept of authors being qualified. How can you be sure that people have the knows they say they have (accreditation problem)?

I realised that this is not an idea that makes sense to teachers, who like carefully curated courses that teach whole areas at once, to students who just accept what they are being taught. Rather, this is a system for autodidacts, which should be constructed by autodidacts. For learners, by learners.

2 – Personal learning toolset

An idea like this, as with any online social tool, has the bootstrapping problem; how do you get a minimal amount of content in there, to make it useful? I think the right approach is like that used by the social bookmarking sites. Social bookmarking sites provided a useful service (personal online bookmarking) that became more powerful as more people used it, because they could begin to aggregate the personal collections into useful social collections via tags.

DL2 could provide a toolkit for tracking your own progress through unknown territory, from what you know to what you don’t, through the tools of learns and knows, and if this is data is public then you could draw from and connect with other people’s learns and knows as they come to exist over time. So it begins as a personal learning tool, the autodidact’s friend, and builds out into a crowdsourced deep learning knowledge base. This also satisfies the vision “for learners by learners”.

3 – An idea for enhancing the crowdsourcing of learns

If you have a big chunk of learning material (say the size of a chapter of a book), it could be very difficult to decide what the dependencies are (what Knows should be required). So break it down.

Assume the body of the Learn is available inline. Then you could allow people to mark up on a sentence by sentence basis what knows are required; ie: attach a required Know to a sentence. Also allow at higher levels; paragraph, section, etc.

So a Learn then depends on the union of all these sets of knows.

And then you can apply this value: everyone creating or editing learns should strive to minimise the dependencies. It should depend on what it needs to depend on, and no more.

People can then look at learns, and edit them over time to remove unwarranted dependencies, right down at the low level of sentences.

An example: In the wikipedia article on Distributed Hash Tables (http://en.wikipedia.org/wiki/Distributed_hash_table), we have the following sentence:

“A key technique used to achieve these goals is that any one node needs to coordinate with only a few other nodes in the system – most commonly, O(log n) of the n participants (see below) – so that only a limited amount of work needs to be done for each change in membership.”

If this was in DL2, it’d need to be tagged with a required Know ‘Big O Notation’ (http://en.wikipedia.org/wiki/Big_O_notation), a branch of maths used in computer science to classify an aspect of the behaviour of algorithms.

Using Big O Notation here is succinct, but perhaps it isn’t necessary? If it could be rewritten without it, you’d remove a frankly onerous dependency.

But you might lose depth. Perhaps the basic text of the learn could be marked up with optional flyouts, which could have harsher dependencies, but being optional makes the dependency optional to the learn as a whole. eg:

“A key technique used to achieve these goals is that any one node needs to coordinate with only a few other nodes in the system so that only a limited amount(1) of work needs to be done for each change in membership.”

flyout (1): ” most commonly, O(log n) of the n participants (see below)” (dependency Big-O-Notation)

So the approach is to allow people to refine and simplify at the sentence-by-sentence level, and thereby make it easier to make a useful edit, easier for the crowd to participate.

4 – Crowdsourced Learns & Knows vs Personal Learns & Knows

Section 2 above assumes personal, owned Learns and Knows created through a personal learning tool. Section 3 on the other hand talks about crowd constructed, publicly editable wiki-like learns and knows. How can we use/combine both concepts? I’m not sure, it needs more thought.

Follow up to Deep Learning 2.0