When AI takes over (graphical) UI

1 Kudos

I just read a post on Wired about the inventors of Siri who are working on a new kind of Artificial Intelligence [1]. This new kind of AI should be able to break down complex queries into sub-queries and integrate with 3rd party services to find solutions to these sub-queries. From these solutions a “master-solution” to the complex query is re-assembled and then output to the user. One use-case described in the post could be:

  • “Book me a plane seat to Dallas with enough legroom to fit Shaquille O’Neal in”.

The query would be broken down likes this:

  1. Book a plane seat to Dallas
  2. Compare legroom of different plane-types
  3. Find size of Shaquille O’Neal

Now the AI searches solutions to each of these problems. E.g. to find the size of Shaquille O’Neal it would search through a database of some Basketball-website. This database search would be realized using an API to create fast, standardized results.

The vision of the founders is to create a global AI that connects to a multitude of 3rd party services in order to find the right information for each of the sub-queries instantly.

I was quickly thinking about the impact such an AI would have on UI. Imagine a world where you interact with data only through voice: there will simply be no need for graphical UI (except for some edge cases, of course, like professional work). But in general debates like “flat design vs everything else”, “skeumorphism yes or no” and “serif or no-serif” will be obsolete. The same goes for questions like “should I design and build this website in a responsive way?”, “should I bother with accessibility?” or “do I really need to optimize for IE8?”. A myriad of (graphical) UI designers will be out of work as soon as such an AI starts to take over.

Such kind of AI will become the UI between data and brain.

The point I am trying to make here: the kind of profession that we regard today as working on cutting-edge solutions, namely UI for complex software, is already being made obsolete, un-employed.

In the bigger picture this means: the speed with which technology is evolving is growing exponentially. Yes, we all knew that. [2] But the impact it will have in 5, 10, 20 years from now will be much worse than we imagine right now. And it will hit professions that we all regard as safe right now, only because they are in IT.

Disclaimer: I don’t dislike these developments. I believe in the benefits they will eventually bring. I am just trying to think forward in order to make better decisions in which directions to develop personally. If I can make some people think about their own decisions, too, something has been won already.

[1] Wired about Viv in August 2014
[2] http://en.wikipedia.org/wiki/Moore’s_law

Why not everybody has to learn to program

4 Kudos

Every now and then it is proclaimed that “programming” should be a bigger part of education and that every child should learn to program (and every adult should go for it also). Surely this makes sense if you’re a big company and craving for good engineers. It also makes sense for some other cases. But I don’t think that everybody should be able to code.

What I think everybody, and I mean that literally, should understand, is the implication that technology is going to have on our future lives. In the next decades a lot of people will go into unemployment because their jobs are taken over by robots, software and any combination of the two.

I recently read an article by German t3n magazine which listed a few jobs which could be taken over. Even for me, who is well aware of the possible implications of technology, there were some surprises. Taxi-drivers are threatened by Google and their self-driving cars. Just today a prototype of a self-driving “pod” was presented by Google at the CODE conference. But doctors could also be replaced in complex operations by more precise robots. The hypotheses in the article is: every job that consists of just manual work will be replaced, and even some which involve intellectual work. An example could be the in-store sales advisor who would be assisted by some kind of virtual sales advisor that is able to communicate with the customer.

This makes it clear why it’s important to understand technology and what it will mean for our future lives. Someone who has to decide into which job to go, has to make a very crucial decision, because there is no guarantee that his job or even the industry will be there for all of his life. Understanding the general developments in technology can help to make a better decision or – as someone who is already working – put things on the right track in terms of professional development and further education.

Quick review of Fever, the RSS reading app

3 Kudos

I was looking for an application to read RSS feeds and to be able to move some browser bookmarks into RSS reading mode. I also wanted to self-host the application to be in full control. The application should also not only read RSS feeds, but also generate some kind of stream in which more important (= content similar to stuff I read frequently in the past) is put on top and less important RSS stories are automatically removed from my sight. I wanted to stay on top of the the item-flood which is typically piling up in RSS readers.

The application should be capable of replacing my facebook feed. Currently I removed all personal posts from my facebook feed and use it to read content from facebook Pages I subscribed to. facebook determines quite good what interests me based upon what interested me in the past. Recently facebook decided it wanted to make more money from advertisers and the post reach of facebook Pages declined. Thus I can’t rely on facebook’s algorithm anymore to deliver the most engaging stories from all subscribed pages into my feed.

Finding Fever

I found Fever. It costs 30 US$ and fulfills all of the above needs. One can enter a lot of RSS feeds and Fever determines which of the posts are valuable enough to show them and which are not. Fever sounds like it does what it should, but I am writing this review to help others not spend the 30 bucks.

How does it work?

Fever lets you put feeds into two categories. Let’s call the categories “main” and “supportive”. “Supportive” contains feeds from outlets which typically re-blog content from outlets in the “main”-category. Fever assumes that outlets from the “supportive” category do this by back-linking to the original story. Fever now counts the amount of back-links to the original story and uses this count to determine how important the original story is.

The general idea is good and works similar as Googles Page Rank, which uses back-links to assess the relevance of a website. But it also means that you have to add as many normal blogs and websites as possible to make the algorithm work better. If you don’t add any feeds to the “supportive” category, Fever doesn’t work.

Fever is useless

Fever comes up with a good idea: use RSS as an established base and then compare the linking between different items to determine what’s important. No more unread feed items in the thousands. Convenient. But.

But it’s not convenient. Because to achieve the above Fever asks the user to actively research as many blogs as possible which just repost content from the big outlets. That’s pointless. I’m reading the big outlets because I want the original story. And I avoid blogs which just repost content without creating their own value.

All in all I would not recommend Fever. Not in order to find the most valuable content in all your feeds, and not as a pure RSS reader – 30 bucks is too much for a simple RSS reader.

Bad design in Windows 8 vs Mac

0 Kudos

Windows is catching up. In Windows 8 some features were finally implemented to help users achieve things faster – even though they are buried beneath the Metro-UI that no one really knows how to use. One example is the easy PC-wide search that Mac had since ages with Spotlight: on a Windows 8 PC one can now press the windows button on the keyboard and start typing away to search through his PC. The integrated automatic backup solution is also getting better, I heard.

One thing Microsoft doesn’t seem to be able to deliver though: an easy, integrated and fast tool to take complete or partial screenshots. On a Mac I press CMD + Shift + 3 / 4 to take a complete or partial screenshot, which is then saved automatically to my Desktop. From there I can continue to edit, copy and rename it. Easy.

Windows still wants me, in Windows 8 — it’s 2014, guys!, to press the print button on the keyboard to save the screenshot to the clipboard, then open Paint and paste it there, then save the screenshot. Partial screenshots are not even possible without setting up some additional tool.

That’s just one reason for me to keep using Mac for productive stuff. It’s still just so much faster than working with Windows, because it just gets out of the way!

Bad design by BMW

0 Kudos

In meinem Auto ist mir letztens ein Design-Fehler aufgefallen, der aus einer Veränderung der erwarteten Funktionalität resultiert. Das wäre nur ein Grund, warum es sich manchmal lohnt auch im Produkt-Design konservativ vorzugehen.

Am Radio gibt es wie in jedem Auto den An/Aus Knopf. In meinem Auto reagiert der Knopf sehr schnell. D.h. es gibt keine Verzögerung zwischen Drücken und Radio-an. Normalerweise würde man damit rechnen, dass einige Sekunden-Bruchteile die Frequenz gesucht und empfangen und erst dann abgespielt wird.

Mir fiel auf, dass der An/Aus Knopf das Radio gar nicht wirklich An/Aus macht, sondern nur Lautlos/Laut. Das war also die Veränderung der erwarteten Funktionalität. Sicher keine schlechte Idee von BMW, weil es sich eben anfühlt, als würde das Radio “schneller” angehen.

Das Problem: wenn ich ein iPhone anschließe und meine Musik in der “Cloud” habe, dann werden die Lieder gestreamt sobald sie abgespielt werden sollen. Das funktioniert mobil häufig nicht, weil die Netze zu langsam sind oder iTunes Match gerade mal wieder nicht läuft. Daher machte ich mir keine weiteren Gedanken, als beim Anschließen des iPhones der Timer des Lieds bei 0:00 stehen blieb.

Die Erwartung war, dass selbst wenn das Soundsystem durch den Radio-Knopf auf Lautlos wäre, würde der Timer beginnen zu laufen. Das hat sich als falsch herausgestellt. Als ich probeweise zufällig den An/Aus Knopf drückte, begann dann tatsächlich das Soundsystem die Musik abzuspielen.

Der gleiche Knopf hat also in – aus Benutzersicht – vergleichbaren Situationen unterschiedliche Funktionen. Einmal stellt er das Soundsystem nur auf Lautlos, einmal hängt von ihm tatsächlich ab, ob die Musik beginnt zu spielen.

Warum das ganze Gefasel? Ich finde man kann daraus ein spannendes Learning für das Design von komplexen Systemen ziehen: jede Abweichung von einer Konvention an der einen Stelle, kann zu einer Fehlfunktion an anderer Stelle führen. 

Nicht umsonst gibt es in allen möglichen Disziplinen konventionsbasierte Regelwerke oder Methodologien, z.B. in der Software-Entwicklung mit dem Grundsatz “Konvention vor Konfiguration”, Design-Konventionen im Webdesign oder auch im Recht mit der Untergliederung des BGB in die verschiedenen Bücher.

Science Fiction or Future?

1 Kudos

I was just watching an episode of “The Next Great Starship”, a casting show by the makers of “Star Citizen”. In the show multiple teams from around the world compete to let their starship design become “reality” in the game.

A thought occurred to me: I had always been saying that I think that all the stuff from current generation science fiction movies will become reality sometime. Chris Roberts just confirmed that thought when spitting out thoughts on starship-concepts from the casting-show.

When designing futuristic stuff for science-fiction products the designer can simply let his creativity flow without any restrictions by legislation, economic forces or technology / engineering. That results in a product which fits perfectly into its time (in the case of Star-Citizen the year 21xx, I think) doing the job it is designed for in this universe perfectly.

As soon as these restrictions from legislation, economic forces and technology / engineering are overcome in “our world”, we will start to see the stuff which we formerly only knew from science fiction in real life.

And since it can be connected so beautifully: it’s already happening. See for example the attached first video, which features an exoskeleton. Remember Matrix, anyone? Or think about what Space-X does: it’s a private company with the explicit mission to go to Mars in this century.

Such amazing times to live in!

Deutschlands Rüstungsindustrie – rückläufiger Umsatz

0 Kudos

Vorhin habe ich eine interessante Nachricht im DLF gehört: die deutsche Rüstungsindustrie vermeldet für 2013 rückläufige Umsatzzahlen. Die Kunden fragen weniger schweres Gerät nach und darauf sind die Jungs bei ThyssenKrupp und Krauss-Maffei Wegmann wohl noch nicht eingestellt.

Doch woran mag das liegen?

Anscheinend macht sich die Umsatzveränderung seit der langsamen Beendigung des zweiten Irak-Kriegs bemerkbar. Kunden unserer stolzen Industrie führen heute nicht mehr nur Kriege, sie führen asymmetrische Kriege. D.h. man ballert nicht mehr mit Panzern auf einzelne “Terroristen”, sondern mit Drohnen oder WLAN-Scharfschützen Gewehren.

Und sonst so? Naja, die USA machen es vor. Wenn keiner mehr die Panzer kaufen will, genehmigt man halt die Lieferung von leichten Waffen in Krisengebiete. Denn es ist ja nicht so, dass es keine Kriege mehr gäbe, sondern nur veränderte Anforderungen.

The responsibility of the tech-industry? Providing unconditional basic income.

1 Kudos

Just read a post by Calcanis on Pando, one of the best tech-blogs at the moment. He makes a point about which responsibility the tech-industry has for the people it has made redundant. [1]

We are not talking of the Snapchats, Instagrams or facebooks of the world here – these make no one redundant. We are talking of applications in the fields of business management, data mining and automation in general. I always felt uncomfortable with the self-perception of SV that it would in any case change the world and everything in it for the better.

I don’t think it does necessarily. Most of the businesses which offer software tools for companies to do something faster, easier or cheaper ultimately are targeted towards making one or more people inside the company redundant. In fact that would be the ultimate sales-pitch for a software company: “Rent my software for 100€ per month, because it does what employee x does now, but for a fraction of the price!” For the client this would be a no-brainer and happily accepted.

But it doesn’t make the world better. It removes employees from their jobs. It helps accumulate money around people who already have money, because money isn’t spread around society via employees any more.

We are only starting to see the outcome of letting manufacturing businesses move to the East right now: more and more people are  becoming unemployed, unable to keep up with a more services-oriented society and the income-gap growing bigger and bigger. I don’t even think we have reached the peak of that change already – so what do you think will happen as soon as the services-based industries are impacted by automation through software?

In the mid- to long-term future people will become redundant in being productive for a living on a large scale altogether. That will leave a large part of society unemployed while a small portion of society will be richer than ever in human history. In order to keep society stable two ways will unfold: suppressing the uncontent masses violently or proving an unconditional basic income, knowing that the money will accumulate around the distributing class through consumption again anyway.

Of course this will not be the tech-industries obligation alone. Society as a whole will have to commit to this change – but the tech-industry and their leaders will play an important role in that. So instead of preparing to take the first road (= “suppression”) by hating against the poor people living in the streets of SF today, the leaders in tech should start to understand where everything is going and get comfortable with the thought of redistribution – be it as UBI or any other way (see Bill Gates, Warren Buffett etc).

That’s their responsibility for making employees redundant.

[1] http://pando.com/2014/01/05/techbrats-goldberg-shih-and-gopman-do-not-represent-the-tech-industry/

Is Facebook Graph Search a Google Search killer?

157 Kudos

Facebook Graph Search a Google Search killer?
No, because Facebook Graph Search (FGS) goes only through the, well, social graph. In some cases it might refer people to some Bing results, but from what I’ve read so far that will be a comparably rare case.

Facebook wants people to stay on site, while Google (Search) — ultimately — wants people to leave the site as fast as possible with the best possible result waiting at the other end.

I think FGS has been hugely over-estimated today: most media seem to have understood that FGS is something like a real search-engine, where it is really just an improved way to finally find some information on Facebook itself. If Facebook really would want to build a web-wide search-engine, they would not have improved their partnership with Bing, right?

The ultimate goal?
To answer the other question: No, I don’t think it is the ultimate goal. Facebook is in the business of gathering data about their users to be able to target ads better and sells ad-slots at higher prices.

Google is in the business of gathering data about the web itself and analyzing those data in order to make people come back to their site because these people find what they want to find. (Obviously, Google also tries to gather data from other location, such as Google+, to also target ads better).

In general, you see, Facebook and Google have different business models even though they seem to be the same at the first look. That’s why Facebook will never try to build a comparable search-engine like the one Google operates.