I’m pleased to announce that Microsoft has entered into a definitive agreement to acquire SwiftKey, whose highly rated, highly engaging software keyboard and SDK powers more than 300 million Android and iOS devices. In this cloud-first, mobile-first world, SwiftKey’s technology aligns with our vision for more personal computing experiences that anticipate our needs versus responding to our commands, and directly supports our ambition to reinvent productivity by leveraging the intelligent cloud. SwiftKey estimates that its users have saved nearly 10 trillion keystrokes, across 100 languages, saving more than 100,000 years in combined typing time. Those are impressive results for an app that launched initially on Android in 2010 and arrived on iOS less than two years ago.
The ‘saved nearly 10 trillion keystrokes’ thing sent shivers down my spine.
Unfortunately, I am not a SwiftKey user. I used to be, but I couldn’t train the program to learn my key strokes well. On top of that, its UI is clumsy when it comes to switching between multiple languages (I have English + German + Dutch + Spanish).
At the end of day, I ended up staying with Google Keyboard, and its prediction result is way better than SwiftKey. And its ability to easily switch between languages is superb.
I know I am in minority on this, as my friend hails SwiftKey as her holy grail.
Edited 2016-02-04 03:48 UTC
With you on that. Never really liked swiftkey. It can’t predict me. Google can.
I keep going back and forth between Google’s keyboard and Swiftkey. I think they both suck about as equally. But then, I’ve always hated touch screen keyboards and avoid typing on them as much as possible. Thank god for Google Voice
I have to agree on the multi-lingual aspect. While I’m not technically fluent in anything but English, I discuss both linguistics and a lot of multi-cultural stuff with a number of my friends, and everything I’ve tried other than Google Keyboard just chokes on multi-language input prediction (especially if you do stuff that involves different writing systems, having something in Greek or Japanese in the middle of an English sentence seriously confuses most prediction systems). The language switching on the Google Keyboard is also better than anything I’ve seen (some IME’s for UNIX like systems come close, but Google Keyboard is still better in my opinion).
Admittedly, I’ve had much better results with Google for even plain English prediction as well, but I think that’s just a result of their methodology working well for my typical speech and typing patterns. I think part of it too though is how much you train it. If I remember correctly, you can feed SwiftKey existing samples of your typing to learn from, but for me this is pointless as most existing examples of my typing on the internet are source code, not regular communication, or are heavily mixed in with other content.
I gave it up because I couldn’t configure it to not autocorrect i to I. Since i is a word in Danish (it means in), any application that autocapitalizes i is completely useless if you ever need to write anything in Danish.
The only way to get rid of the behavior was to disable all auto-capitalization (also initial and after period), but that just make SwitchKey much more cumbersome to use than any other software keyboard.
I liked it before they went free. After that I felt like the quality and innovation fell off.
No problems with prediction. I let it ‘parse’ my Facebook… No doubt all these are also saved. 🙁
I’ll probably switch to Google if the prediction is OK. Google already knows lots, I don’t want to cut Microsoft in.
You do realize the ‘save’ is not the one in the “save as dialog box”, but rather the one “I saved ten bucks in shipping by not buying online”, don’t you?
Edited 2016-02-04 07:54 UTC
I think Thom is referring to Microsoft keeping track of how many keystrokes are being saved, as in, Microsoft is spying on you as usual.
Edited 2016-02-04 20:57 UTC
Hmm? MS have only 2 days ago agreed to acquire SwiftKey – so how could they?
Maybe he thinks the amount of typing SwiftKey’s word completion rendered unnecessary is scary.
We’ll probably never know.
Swype user here. I’ve never used SwiftKey, but I think the Google Keyboard is horrible.
I used Swype for years but switched to Swiftkey because of the cursor keys.
Yesterday I switched back to Swype.
Try the Android Blackberry keyboard, it is pretty good.
Thank you. I will install it now and see how I like it.
It’s fine but I’m not excited about it.
Chances are I will return to Swype but I haven’t given up on it yet.
[Entirely speculative]
The prediction and accuracy of the Microsoft WordFlow Keyboard got noticably worse in Windows 10 Mobile. I am suspecting that there were some patent-issues there that escalated when Microsoft announced that they would put the keyboard on Android and iOS as well. Buying a company like them for 250 million and giving that product away for free makes no sense to me.
Then again, Microsoft did the same with calender and mail apps where they spend a lot of money for aquisition/aquahire and don’t seem to benefit from it financially in any way. Of course “Microsoft should be seen as the productivity company” is a reason, but is that a ‘billion dollar’ per year reason?
avgalen,
Obviously the technology itself is certainly not worth $250M, it was probably was built by a couple college kids. A USPTO search shows 0 patents issued to swiftkey or it’s owners. Maybe they have assets in the UK?
So what’s left? Users! They have 250 million users (according to swiftkey’s app description).
Depending on the industry, it might cost businesses $10-$350 of overhead (advertising, calls, demos, samples, etc) to acquire a single customer.
http://www.entrepreneur.com/article/225415
So even though MS could easily come out with their own tech for $1M, it would likely cost a great deal more to acquire those users, especially on IOS and Android where they don’t have any sort of defacto monopoly control.
Now that they’re MS users, I’d expect to start seeing more integration with MS services. Given the microsoft of late, it’s possible they would resort to the same sort of “forced upgrades” we’re seeing them apply elsewhere.
Edited 2016-02-04 14:26 UTC
Oh, there are lots of patents:
https://swiftkey.com/en/intellectual-property
And of course Microsoft bought Swiftkey, Acomplii, Sunrise and many others for their popularity. But having hundreds of millions of users means nothing if you don’t monetize those users in some way. It doesn’t make sense to spend a billion in a year without any way to get a return on that money. And in that case having lots of users is actually bad because you will have quite some bandwidth and developer costs
Monetization has always been a ‘red face’ issue in this industry. Obviously it exist, and is very good.
avgalen,
I am so fortunate not to have this burden, poor guys. My sympathies go out to all those with too many users
Edited 2016-02-05 16:46 UTC
Like they did with Nokia? And then incurring other losses in addition to the “partner payments”. Only to spend another $8 billion to keep them from going Android and abandoning what was essentially the only sales channel for the phone OS? >(8^D>
My first smart phone was the Nokia N900 (horizontal slide-out physical keyboard). Man, I could type so fast on that thing. I had auto-complete and word prediction disabled because it just slowed me down.
I could type anything. I did texting, email, blogging, programming… Using chorded keypresses with Shift, Ctrl, and Fn gave me instant access to almost every key on a full sized keyboard.
It had arrow keys to move around the text field. Ctrl-arrow to jump between words. Shift-arrow to start highlighting. Ctrl-A, Ctrl-X, Ctrl-C, Ctrl-V, all standard, all working instantaly and precicely.
By having a physical keyboard, I was able to “hold” my thumbs on the keys I was going to press next, look up and engage in a conversation, then look down and pick right up where I left off.
Sorry for the ramble. I don’t really have a point. I just really miss having a physical keyboard.
drcouzelis,
Same. I can type on a keyboard faster than I can write with a pen and paper. Ctrl/shift-arrow keys are second nature to me for text editing. Trying to do the same things on touchscreens is totally painful. I’ve tried many virtual keyboards, but even after several days of training myself, they’re not even remotely close to the productivity of the real thing. If I have to jot down a grocery list or notes in a meeting, I still use the old pen & paper because that’s less tedious.
I’d probably use a stylus if it worked well. Voice input seems to work well for simple dictation tasks, but there are situations when speaking things out loud is not appropriate. I dislike having to use IOS and Android voice recognition online.
I just came across this, maybe there is a way to get android’s voice recognition to work offline?
http://stackoverflow.com/questions/17616994/offline-speech-recognit…
Edited 2016-02-04 16:03 UTC
BlackBerry Priv – Android with an oldskool physical keyboard is here! I won’t use a device without it, despite many attempts
Check out LG F3Q. Somewhat old, but it is available on ebay. Camera sucks, internal memory sucks even harder, but overall it’s a usable phone with a keyboard you’ll like.
FastNote savings and loan me ten banana having a stroke! Thank you, FastNote!
T9 4 lyfe!
Still the best “keyboard” and I can still type quicker than the error prone prodding of a touch screen keyboard.
Edited 2016-02-04 19:54 UTC
Boom, there it is, right there: what went wrong in 2007, and why I inevitably rage at every general purpose computing device I lay fingers on today.
Every computing experience I’ve ever enjoyed has in common that it sits there doing nothing until I tell it what to do, and every computing experience I’ve hated has in common that it guesses what I want, makes suggestions, guides me down an artificially narrow set of options, and makes me fight it to turn all of that garbage off.
🙂
Agree with you, atsureki. On this lack of continuity in between generations, at the coding community, lots of wise is lost periodically.
At this moment Microsoft mobile needing it.
A scripting interface ‘a la’ Hawei would be a lot more useful.
Just one thing I find ‘dangerous’ with predictive tools is my wife sending messages with absolutely different meaning. [Not corrected in the hurry].
When auto-correct is disabled, you can guess adequately the meaning.
In general approving this kind of efforts