#local maxima

When Humans are Parasites on Algorithms

You know, I’d hoped to completely forget about these things, but apparently they’ve gotten worse--or I just wasn’t fully exposed to the true horror.

If you’re not aware, writer and artist James Bridle recently wrote a post called Something is wrong on the internet.

Someone or something or some combination of people and things is using YouTube to systematically frighten, traumatise, and abuse children, automatically and at scale, and it forces me to question my own beliefs about the internet, at every level. Much of what I am going to describe next has been covered elsewhere, although none of the mainstream coverage I’ve seen has really grasped the implications of what seems to be occurring.

I’d suggest reading it, though maybe don’t watch the linked videos. Which are comparatively mild examples.

It’s still not entirely clear to what extent the creation of these things is automated. Some involve real humans singing finger family songs to match bizzare combinations of keywords. Others are clearly either automated or produced in sweatshops--the dirty secret of AI is the extent to which a lot of it is just an update of the Mechanical Turk.

But the crucial point that I missed is that there is an obvious automated system behind all of this: YouTube’s algorithms. They’re not very smart, even as algorithms go. Though they’d be classified as AI, it’s more like the intelligence of a bacteria, or perhaps a very dim sort of worm, like the kind that crawls on a sidewalk after the rain and dies when the sun comes out.

This is an important point: when we say “Artificial Intelligence” we often think of a complex, almost magical intellect...but most of the real AI work is merely blindly optimizing for a particular goal. AlphaGo is very good at playing Go, and absolutely useless at anything else. The interesting part, for researchers, is figuring out how to transfer the principles behind the AI to solve other problems.

In YouTube’s case, the algorithm is blindly trying to find content that’s vaguely associated with the thing you just watched. As it turns out, when you combine it with human toddlers, that leads down a dark, dark hole as their uncritical viewing reinforces the algorithm. The wet sidewalk has way more water than the damp earth, the worm doesn’t realize the danger, we get a YouTube run by undead zombie worms showing children content they’ll never forget, much like this metaphor.

So we’ve created an economic ecosystem where humans are literally performing incantations to fulfill what they think an AI wants because an interaction between an AI and babies has gotten stuck in a local maxima.

Mike Cook has a good response for this: Better Living Through Generativity:

Photography’s slow walk to ubiquity also had a darker side. As it became better-known, photography was understood to be a way to record real events, but this imperfect understanding enabled a lot of people to do fairly awful things with it. One of its first proponents faked his own death. People used it to prove the existence of fairies, to show that they could summon spirits, to verify the existence of monsters and mythical animals. People knew enough about photography to benefit from it, but not enough to protect themselves from it.

What solved this problem? Well, a bunch of things, but undoubtedly one of those things was helping people understand the processes by which these images were made, and giving them the power to make them. Cameras and development became more commonplace, people understood how to overlay images or touch them up after the fact. We see the same cycle today with Photoshop: first, it caused chaos; then people understood it; finally, people harnessed the power for themselves. Now we edit photos on our phones as we take them, before sharing them with others.

I see PROCJAM as part of an effort to enact this change for generative software. By bringing people the resources, the tools and the space to make generative systems, they can take ownership of the concept and understand their strengths and their weaknesses. Right now only a few hundred people enter PROCJAM, but ultimately we should all be working to make these ideas accessible and fun for people to try. In doing so, we popularise these ideas and rob them of some of their incapacitating power.

Go read the whole thing.

This won’t solve the problem of YouTube--in the end, Google is responsible for what their algorithm has done. But I do think that one of the best personal defenses against this is to learn what procedural generation and AI are really capable of and--more importantly--what they can’t do.

William Fields - Local Maxima [Conditional]

Local Maxima is released on April 30th via Conditional Records.

William Fields is an audio-visual artist like few others, conjuring unforeseen shapes that break apart and fuse together in sync with unconventional rhythmic structures. We are very excited to bring his latest EP, Local Maxima, to Conditional.

Local Maxima creates an entirely unique sound world. Time breathes deeply and rhythm slips out of control, bordering on chaos but always retaining a swinging pulse. Icy harmonies stretch and contract atop rolling drums and warm bass tones.

I want to offer an alternative view to Don Norman’s article “Technology First, Needs Last” where he argues that “design research is great when it comes to improving existing product categories, but essentially useless when it comes to new, innovative breakthroughs” and that those breakthroughs are largely technological ones.

This blog entry got long. Great writers are concise. As I am a mediocre writer, I'm leaving it longer than I'd like it to be. Try not to yawn out loud.

The limitation of a local maximum (background)

When user-centered design is implemented, it usually goes something like this: NonInnovation, Inc., an operations-oriented product firm realizes that they need to make a big advance in a product. They hire BouncyBallChair Inc, an innovation company which offices in a loft space where the walls are pastels and everyone has nurf guns.  Oh, and all of their employees wear Warby-Parker glasses.

So BouncyBallChair Inc goes through the user-centered design process, working with NonInnovation to study the target customer or consumer, identify the core problem or need, and create a super-cool product to solve that problem.  All throughout, they’re using User-Centered Design. The product is released…it’s the best new product in that category.

But what did we really create? It’s a beautifully designed shopping cart with removable boxes, each a different primary color.  If the Google logo were a shopping cart, this is what it would look like. So we have solved the problem of how to make the best possible shopping cart. Local maximum.

Now let’s take a real life company, Twilio. They didn’t have some major technology breakthrough. They used Asterisk, an existing open-source platform, and turned it into a SaaS product.  I don’t know anyone at that company, but I imagine at some point in the company’s formation, someone had this conversation: “how cool would it be if we could text a website…or if a website could call a phone…with, like, one line of code.”  They would be correct, that’s really cool.

Now, to be fair, I’m sure they had specific use cases in mind, but really what they were doing was creating a platform for an infinite number of uses. 

To illustrate this coolness, their pitch in front of a large audience at the New York Tech Meetup went something like this: “Look, with one line of php code, I can create this phone number that you can now call and hear a message.”  The presenter had people call the number. Cool. But what was mind blowing was that he had a second webpage, which looped through all the numbers that had called in, and one by one, people’s phones started to ring in the audience.  It was visceral. It was surprising. A computer shouldn’t be able to do this. Certainly not with a couple of lines of PHP code.

The underlying issue is that user-centered design can produce breakthroughs, but it’s how it’s being used and to some degree who is using it which prevent it from doing so.  I believe the underlying difference is that you’re solving for two different problems: One I will call Solving for a use case and the other Solving for surprise.

Solving For A Use Case vs. Solving For Surprise, Part 2: Single Use Cases vs. Surprise

When we contrast the two examples from part 1 (the shopping cart vs. the grocery store), the key differentiator for me is that user-centered design tends to focus on solving a single problem – a single use case.  However, the most exciting and innovative products do something surprising. Something that shouldn’t be possible.  As a result, designers, programmers, and everyone else can immediately think of their own use cases for these products. In some sense, this is the difference between a product and a platform.

This is not to say that those innovations don’t solve problems – they absolutely do. But they solve problems beyond a single use case or two. At their best, they are infinitely scalable.

So if you agree that the most innovative products will be surprising, potentially solving hundreds of use cases, whereas user-centered design solves specific problems, sometimes using these innovations, two questions remain. First, must these innovations be massive leaps forward in technology, only invented by mad scientists or criminal masterminds?  Second, can user-centered design be used to identify these types of innovations?

Do these innovations have to be massive leaps forward in technology?  

Twilio is a technology company, but they did not make the underlying technology, it just made it super-easy to use. Very few of the truly innovative products right now are technological breakthrous -- not Twitter, not Facebook (especially at first), certainly not Groupon or AirBnB or Birchbox.  

Can user-centered design still be used to identify and create these types of innovations?

Yes, of course! The most entrepreneurs I know in the tech space observe a problem people have through some sort of formal or informal user-centered design. The difference is that they then go out and look for other use cases.  We need two design constraints in our normal user-centered design process.  First, a criterion must be added: is my solution surprising? I don’t mean “oh, that’s clever” I mean “woah, I (almost) can’t believe that’s even possible.”  Not “why didn’t I think of that” but “I never would have thought of that.”

Second, we must ask ourselves “okay, so this is a surprisingly great solution for this one use-case, is it generalizable to potential other use cases?” Or, more likely “what are the characteristics of this solution’s applicability to this use case that might make it a good fit for other use cases? What are the charactaristics that make Facebook great for distributing images? Is it that I know my friends, but not strangers, will automatically see pictures they might find interesting? What else might I want to distribute that way? My location (Foursquare)?  Invitations to private games (Farmville)?    

The Point solved a specific problem for social action that was so generalizable that it cnot only encompassed multiple social actions, but it also included the Groupon use case. Groupon itself is infinitely scalable. If it solved a problem for a local Pizza place (which I believe was their first use case), what other places have that similar problem? Stores that have high lifetime customer values and high loyalty rates are willing to pay to acquire new customers, in this case by offering deep discounts for a first customer's visit. Clearly that scales across lots of different types of retail, but as it grew, it had its own value proposition -- get your company's name in front of millions of people and only pay if people buy the coupon.  This is a second use case of the product that increases the pool of retailers to whom the service is appealing.

Who's funding the designing matters

It may be the case that it’s difficult to be massively innovative using user-centered design on someone else's behalf (i.e., as a consultant). Imagine if in the course of doing consumer research for a shopping cart manufacturer, you (correctly) identify the need for Grocery Subscriptions, a truly innovative concept. It is unlikely that a shopping cart manufacturer can/should take on something that far out of their core competency.

Uber identified a market need and had no constraints as to how to implement it. They basically created NetJets for private drivers. Should they buy our own cars? Pool together existing limo services? Distribute on the iPhone? No constraints = the ability to make it work.  Had a taxi company hired a design consulting firm to identify new consumer needs, even if the firm identified a need for Uber, I doubt the taxi company would have been in a position to create this type of service.  

Conclusion