Building on unfeasible ground – An Amendment

This post was originally published on the 30th August. Since then I’ve been contacted by x.ai who dispute the claims made by the Bloomberg article and the summary I gave. Having reviewed the article and rebuttal by x.ai I’ve decided to amend this article.

The section below on x.ai has now been updated. More to follow on this fascinating and evolving area.

As a designer it’s important to be aware of the current state of cutting edge technology and experience in your discipline area.

As an interaction designer this increasingly difficult as the line blurs between the technology that underpins the things we experience each day. Looking at the average digital product or service, it’s now all but impossible to distinguish between a public beta, a stable service, one half of an A/B test and an untested prototype.

This makes if difficult to know where the cutting edge really is. Staying up to date with things requires that we’re all aware of what’s technologically feasible. In the pr-software world of design it was easier to assume that real meant feasible. However, it’s not longer safe to assume that seeing and interacting with something in the real world means it’s a safe example of technology that you could suggest in your own work.

Real doesn’t mean feasible any more

Feasibility is the measure of how possible it is to build a design with today’s technology. In the design world it’s a term tied to our industrial heritage: a time when success was all about engineering a solution and mass producing it. Feasibility in the age of software is much murkier, much less easy to define.

As designers we still need to strive for feasibility on behalf of our clients. Although we might not be the ones engineering the impossible, it might be us who’s specifying the impossible. It’s really important to see how some of the most cutting edge experiences we’re inspired by might be entirely unfeasible. Below are a few examples.

1. Kickstarter’s fail rate

A case in point is the very existence of Kickstarter – a service built in the notion of a yet-to-exist product being available for preorder. Listers are encouraged to make it as real as possible to encourage people to invest.

The success rate of fulfilled products is an impressive 91%, however it means that 9% of successfully funded items will never actually see the light of day. This of course doesn’t include the others that simply don’t make their funding goal, the much higher number of 56%.

Kickstarter Fail Rate

Amazon would be quite a different service if half of the products we’re listed as ‘unavailable’ and one in ten products order didn’t actually arrive.

2. Lyft Carpool

San Francisco born ride hailing service Lyft recently announced it was rolling back it’s pooled ride feature ‘Lyft Carpool’ – where car riders were encouraged to pickup someone on the same route.

Lyft Carpool

It’s yet to be seen if the Waze equivalent Waze Carpool will also survive. But either way it’s interesting to see features rolled out and rolled back again.

Waze Carpool

Uber itself is an interesting example as leaked information earlier this month showed it had lost $1.2bn in the arly part this year.

3. x.ai accused of using humans

Automated assistant service x.ai garnered some less than positive press earlier this year when Bloomberg asserted that there were perhaps more humans in the loop that the layman might assume when thinking of and AI system. Bloomberg, it seems, have mis-interpreted the role of the ‘trainer’ of the machine (a common practice in AI systems)

x.ai

Whether you believe the Bloomberg article or x.ai’s counterpoint, the reality is that it’s difficult to know which side to take when the simplified, understandable version may actually be very misleading. When most people assume that the state of AI is defined by services like x.ai, it’s important to realise that all is not what it might seem…

What it means for design

So as designers what can we do? It’s not realistic for us to see inside each the companies we’re looking to for inspiration. Even if we could, few of us could really interrogate their tech or business models (although these are both areas that designers should be making themselves more literate in).

There are a few things we can try to do though…

Firstly we can recognise that the examples above are themes that will be repeated. We can keep reminding ourselves and our clients that real doesn’t mean feasible. We can make ourselves aware that if it looks too good to be true, it might well be.

Secondly we can try to be more diligent. We can try to investigate the latest tech announcements, to see if they are too good to be true. Don’t just quote the tweet about the next big thing, read the article, track down the source, cross reference the claims being made.

Thirdly we can also be inspired by the impossible. Design is not solely about operating within the constraints of technology, it’s about pushing the boundaries in search of what’s best for the people were designing. So don’t take this article as a suggestion to be less ambitious, take it as a recommendation to be more professional.