From an NYT article on Google’s translation services, this excerpt sums up the most critical transition in machine learning that has happened thus far:
Creating a translation machine has long been seen as one of the toughest challenges in artificial intelligence. For decades, computer scientists tried using a rules-based approach — teaching the computer the linguistic rules of two languages and giving it the necessary dictionaries.
But in the mid-1990s, researchers began favoring a so-called statistical approach. They found that if they fed the computer thousands or millions of passages and their human-generated translations, it could learn to make accurate guesses about how to translate new texts.
This is huge. Google has enabled a Java scripting application across it’s online Apps suite. Previously available only to premier subscribers, the scripting capability is now open to anyone with a standard account as well. This competes directly with Microsoft’s cross-Office VBA, allowing users to build their own applications while using the productivity suite as a front end. It’s a very impressive and bold move to allow it over the web – and further solidifies Google’s “the web is the OS” mantra.
Read more at the official Google Enterprise blog.
From the “that’s a feature, not a bug” file: I always thought my iPhone’s ability to continue email searches on the server (as opposed to emails stored on the phone) was broken, since it never returned any results even for emails I knew existed. Today, I learned that remote search is explicitly not supported, according to the Google Mobile Sync fine print:
Searching Gmail for messages that have not been synced to your phone is currently not supported.
As a workaround, I’m experimenting with syncing more emails to the phone…
Engadget has gotten their hands on Google’s Nexus One phone and while further details will be forthcoming at Google’s press event on Tuesday, they have a pretty in depth preview.
The most important takeaway is that despite the iPhone-launch-esque frothing of the technology media at large, this is not a revolutionary phone. A couple weeks ago, I posted my disagreement with TechCrunch et al, saying that the Nexus One (in contrast to the TechCrunch headline) doesn’t change much at all. And now Engadget confirms:
The thing that’s struck us most (so far) about the Nexus One thus far is the fact that it’s really not very different than the Droid in any substantial way…. Don’t get us wrong, the phone cooks — but it’s not some paradigmatic shift for Android.
Furthermore, Android’s chief limitation in the new smartphone arena persists:
One other note: multitouch has not been included here, so while the functionality is supported in Android 2.0 and up, we’re still dealing with a one-finger-at-a-time experience… which leaves something to be desired when you’ve got a beautiful touchscreen like this to play around on.
As expected, Google has not split Android onto a new trunk for Nexus development, though they have advanced the OS somewhat – mostly in terms of small efficiencies aimed at improving overall usability. It remains to be seen, however, if those changes will remain exclusive to the Nexus line or distributed to all of the Android models.
Last and certainly not least, it looks like the Nexus will be a T-Mobile device – so much for the grand unlocked plan. It also won’t be compatible with AT&T’s 3G network, which further compromises its portability.
Engadget will be posting a full review shortly – this is just a collection of their first impressions. Nonetheless, it’s clear that the Nexus One is not a game changer.
Aaron Swartz makes some good points about Google, but I just can’t keep reading once he invokes a tactic I can’t stand (and which sadly seems to be gaining currency):
If Microsoft had Google’s market share in search, is there any doubt that they’d be systematically demoting or even banning their competitors in the search results?
Via Daring Fireball
CrunchGear may be focusing on the hardware, I’m going to focus on the competition: the most salient outcome of Google’s decision not to partner with a carrier is that people will be able to discriminate among carriers based on network quality rather than phone features.
This is big (though lest I sound hypocritical, I dont think it has anything to do with Google per se). If I didn’t have to contend with AT&T’s awful NYC coverage, I would be back to Verizon in a second – but my iPhone requires me to put up with it. So, interestingly, it puts the focus back on network quality - where it should be.
In their usual over-enthusiasm for all things with touchscreens (too soon?), CrunchGear has been gushing over Google’s rumored phone. Google has confirmed that they are working on “a device” without further specifics. That hasn’t stopped CrunchGear from actually writing:
…if and when Google starts selling this thing, prepare for some of the strangest – and coolest – times in mobile we’ve ever experienced.
Why all the excitement? Because – get this – a company is going to design hardware (which so far looks like just about every other smartphone out there) which runs open-source software (that so far is available to just about every other smartphone out there).
Here’s how they put it:
When service providers make hardware, they have a different set of priorities. They have a lower number of SKUs so their products have to be great. They control a lot of the software so the UI is great. They control the distribution so there’s a bit of the “rarity” and “early adopter” factor to consider. This sort of stuff is what CE and PC manufacturers would kill for – after all, when’s the last time you drooled over a desktop?
Just to tick through those points – does CrunchGear really believe that fewer products equal better products? The gadget landscape is littered with discarded hardware whose original premise was to tie a service back to a physical device. The iPod is arguably the first massively successful iteration of the idea – but it certainly wasn’t the first attempt. GPS companies, news services, email providers, devices that sit on your nightstand and deliver content, video game producers - you can throw a dart and find example of a service company that has made the fatal decision, “You know what we need? A piece of hardware. Once we’re in people’s pockets/living rooms/night stands/cars they’ll have to use our service!”
Next up: “they control the software.” The software is Android. Yes, it’s great. But it’s also open source. So there’s no competitive advantage in using it. An exception scenario looks like this: Google designs a new, revolutionary hardware feature. They introduce code into Android which supports it. They don’t tell anybody – that’s the key step. On launch day, the Google Phone has a killer feature which no other phone has, at least until the other manufacturers introduce it. Another scenario could run like this: Google says “screw you” to the Android consortium and starts working on their own proprietary fork. Maybe they just work on their own UI – something HTC has done. I’ll give CrunchGear the benefit of the doubt on this one.
As for the distribution – please. When was the last time someone said “I’m so happy that Apple controls the distribution of the iPhone, because it was so easy to get one when they launched” – and Apple has real physical locations at which to sell their devices! See also: Kindle launch, Nook launch, Xbox launch, etc. The myth of distribution is overblown – there’s benefit in controlling your channel and hawking your own wares, but it’s not a make or break situation. Having a good app store is much more important these days (oh, wait… Android’s Market is available to all handset manufactures).
At the end of the day, CrunchGear is basically hoping that “Google is going to be like Apple!” But in fact, the biggest reason that the Google Phone might actually be a massive success is one that Apple explicitly avoids: the Google Phone is supposed to launch as an unlocked phone available on any GSM carrier. On the one hand, that will open it up to a much wider audience. Furthermore, it frees them of any forced branding issues in the software (although Apple managed to do just fine there). The only downside will be that there may be no carrier to subsidize the phone. Perhaps a few key agreements – or Google’s massive cash reserve – will ameliorate the phone’s cost.
So the Google Phone isn’t revolutionary. It isn’t extraordinary. In fact, it’s not even real yet. Now, I have never used a Google product that I thought was in any way buggy, difficult, or pre-production. I’m a huge fan of everything they do. I have faith that they can produce a handset which, with contracted hardware, an open source OS and no launch partner, is a viable competitor to Apple’s closed iPhone ecosystem. But excuse me if I don’t hold my breath and wait.
Now, I do also love one other line in the CrunchGear article:
But suddenly service providers are doing hardware. Amazon has the Kindle, Barnes&Noble has a lumpen Nook, and now Google has a phone. What’s next? The Credit Suisse Fondue Set?
(Excuse for a second the author’s ignorance – you’d think he’s never heard of a company outside the S&P 500. You’d think he’d never heard of his own company [too soon again?]) You better believe that if CS thought they could profitably securitize little cubes of cheese, a fondue set would assuredly follow (probably through a complicated series of acquisitions). Don’t forget that the single biggest positive offset to Goldman Sach’s Q307 loss was, of all things, wind farming.
Has Mashable implemented a word count quota? In a post about Google’s new search product codenamed Caffeine, they published a brief four-sentence note from the Caffeine development team… and then went on to summarize the note’s “key takeaways” with two bullet points:
Since the launch of the developer preview however, we haven’t heard much about it. It’s quietly been undergoing testing and tweaking. That’s all changing though, as now the developer preview of Google Caffeine has been taken down, replaced with an announcement that Google Caffeine will go live in its first datacenter soon.
The new Caffeine splash page says it all. Here’s the full text:
“Based on the success we’ve seen, we believe Caffeine is ready for a larger audience. Soon we will activate Caffeine more widely, beginning with one data center. This sandbox is no longer necessary and has been retired, but we appreciate the testing and positive input that webmasters and publishers have given. ”
While we admit that there isn’t a lot to go off of, we can discern two key takeaways:
- Caffeine testing was successful. The project is moving forward
- Caffeine is going live on Google.com soon, starting with one datacenter
Congratulations, Mashable. You managed to say in 20 words that which took Google 32. But if we include Mashable’s bold text immediately preceding the quote, their wordcount jumps to 32 as well, making their post an exemplary exercise in unnecessary summarization and redundancy.
Finally.
I’ve been using Google’s contact and calendar syncing increasingly, to the point that they are almost indispensable to me. Most critically, when my last iPhone broke I only had to wait a few seconds for my new one to download all of my information from the cloud. The addition of push email completes Google Sync’s basic exchange functionality and establishes it as a robust (and free!) mobile cloud suite.
Previously: The sound of a million MobileMe subscribers screaming.
Update 9/23: Is this killing anyone else’s battery life?
My Google Reader was filled with a lot of headlines on Google’s new Fast Flip service this morning, but none of them amused me quite as much as Silicon Alley Insider’s confused monologue:
7:45 a.m.: Google’s Fast Flip Makes Reading Print Publications Online Easier
8:40 a.m.: Google FastFlip Is A Gigantic Step Backwards
10:29 a.m.: Google FastFlip Is Latest Attack On Amazon Kindle
12:25 a.m.: Google FastFlip Is Especially Useless On The iPhone
(Yes, SAI decided that “Fast Flip” wasn’t Web 2.0-ish enough and removed the offending space themselves.)
The most recent two were written by Dan Frommer; the previous two are by two other authors. The 8:40 piece is actually cross-posted from another blog. Wouldn’t it be nice if media publications had cohesive narratives?
The truth is, I’m most inclined to agree with the 8:40 post. I don’t think this release is a particularly big deal; it certainly doesn’t deserve the attention being lavished on it. At it’s core, it is simply a very fast way of serving up low-bandwidth screenshots of websites. It fits somewhere below Google News and Google Reader. It has some good search and browsing features but I find it a bit too dumbed-down and the previews take up too much space – I feel like I’m missing content because my screen is filled with pictures rather than information.
I know, the point of the service is to present news in a style more akin to scanning a physical page for interesting items – but in such a case, I’m searching within a single graphical style, not between styles as I must on Fast Flip (I sound like I’m teaching ANOVA!). This is a major difference. If I’m scanning a page of the NYT, or browsing a magazine, I’m looking for headlines that interest me; I have already internalized the design/layout of that publication and essentially ignore it (even if it’s what attracted me to that particular paper in the first place). Imagine if the NYT had a vastly different design on every page – it would be a nightmare trying to browse quickly for interesting content, since with each page turn I’d have to internalize a new layout.
Yet that is the metaphor Fast Flip currently embraces. I’d much rather see them reformat the text of every source onto a single template, and present the result as a giant uniform news service. This wouldn’t be so far off from the “expanded” view of my Google Reader. Also, it would let me truly browse for content rather than design. And let’s face it – once I move to any sort of aggregation, I’m effectively giving up on design anyway, since I’ve moved away from the host site entirely.