No longer can you seek him here or there. Photo by abrinsky on Flickr.
You can now sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.
A selection of 10 links for you. Use them wisely. I’m charlesarthur on Twitter. Observations and links welcome.
Google search chief [Amit] Singhal to retire, replaced by AI exec » Bloomberg Business
“When I started [at Google in 2000], who would have imagined that in a short period of fifteen years, we would tap a button, ask Google anything and get the answer,” Singhal wrote in a Google+ post announcing his retirement. “My dream Star Trek computer is becoming a reality, and it is far better than what I ever imagined.”
With Giannandrea’s appointment, the technology may get smarter. The executive has overseen recent artificial intelligence efforts, including RankBrain, which saw Google plug an AI technology called a neural network into its search engine to boost the accuracy of results and an e-mail service called Smart Reply that automatically writes responses. Other work he has managed include efforts in image recognition and technologies that fetch information based on what users are doing with their devices, rather than what they’re explicitly searching for.
[John] Giannandrea joined Google in 2010 when it acquired a company he co-founded called Metaweb Technologies. Those assets became the basis for Google’s knowledge graph, a vast store of information on hundreds of millions of entities that helps the search engine present factual data in response to certain queries. Singhal’s last day is scheduled to be Feb. 26.
The elevation of Giannandrea represents a further emphasis on the importance of artificial intelligence to Google, a unit of Alphabet Inc. Chief executive officer Sundar Pichai said the technology has been key to recent efforts in search on mobile devices and personal assistant technologies.
Speaking of search..
link to this extract
Garth Gibbs: ‘The archetypal diary writer’ » Press Gazette
Garth also managed to spend much of his time chasing various ‘sightings’ of ‘Lucky’ Lord Lucan, who was thought to have fled abroad after apparently mistaking his nanny for his wife and bludgeoning the ‘wrong’ woman to death. Of this colourful period in an almost continually helter-skelter career, Garth himself wrote: ‘As that brilliantly bigoted and crusty old columnist John Junor once cannily observed: ‘Laddie, you don’t ever want to shoot the fox. Once the fox is dead there is nothing left to chase.'”
With a wonderfully fertile imagination – a prerequisite of any good tabloid journalist – plus a good deal of chutzpah, Garth relished the challenge of keeping Lord Lucan alive, but never finding him.
‘I regard not finding Lord Lucan as my most spectacular success in journalism,’he said. ‘Of course, many of my colleagues have also been fairly successful in not finding Lord Lucan. But I have successfully not found him in more exotic spots than anybody else.
‘I spent three glorious weeks not finding him in Cape Town, magical days and nights not finding him in the Black Mountains of Wales, and wonderful and successful short breaks not finding him in Macau either, or in Hong Kong or even in Green Turtle Cay in the Bahamas where you can find anyone.”
Lucan was finally declared dead – though never found – on Wednesday. Not finding him was indeed a splendid task allotted to many journalists down the years. Speaking of search…
link to this extract
#SEO for sale?! Exposing Google loopholes in light of FTC native guidelines » aimClear® Blog
Mashable, a respected global media company focused on informing and entertaining “the digital generation,” was our inspiration. Mashable has joined the swelling ranks of websites selling native content articles to advertisers. Initially we were interested in participating in the program and reached out to Mashable regarding their native post advertising, which is called BrandSpeak or BrandLab.
As the conversation progressed, we were curious as to how Mashable native posts show up in Google search results and disclosure verbiage in light of new FTC native advertising guidelines. After we corresponded with a Mashable sales associate and researched BrandSpeak/BrandLab in detail, we were motivated to share our findings with the community as a point of learning about native content.
Those findings surprised (and astonished) us. Aimclear analysts identified a Google SEO loophole, which is perhaps the greatest ranking algorithm gap in years, allowing marketers to literally buy their way into Google search results with paid content…
…At best, allowing paid SEO tilts the playing field, making it even harder for smaller, perhaps more relevant players to compete for free Web Search results.
Google’s Webmaster Guidelines governing native content and Web Search are firmly rooted in 2013.
Tricky; this stuff is low-quality, but sites are desperate to generate revenue somehow. Speaking of revenue…
link to this extract
Google boots ad blockers from Google Play » TechCrunch
According to Rockship Apps founder and CEO Brian Kennish, maker of Adblock Fast, Google’s app reviews team informed him the app was being removed for violating “Section 4.4” of the Android Developer Distribution Agreement.
This is the section that informs developers they can’t release apps that interfere with “the devices, servers, networks, or other properties or services of any third-party including, but not limited to, Android users, Google or any mobile network operator.”
If that text sounds a little broad-reaching and vague, that’s because it is. It’s also what allows Google to react to changes in the industry, like this one, on the fly.
Kennish says that Google’s app reviews team informed him that he could resubmit after modifying his app so it didn’t “interfere with another app, service or product in an unauthorized manner.”
“We’ve been trying to contact Google through their public channels since Monday, and I tried through private ones all day yesterday…but we haven’t gotten any official response from a human – just autoresponders,” notes Kennish.
He suspects that Adblock Fast was the first to be pulled from Google’s app store because it had climbed the charts so quickly and had achieved a 4.25 rating. Kennish says that the app had around 50,000 installs at the time of its removal.
In addition, the company could have gotten on Google’s radar by pushing out an update that offered a better user experience. (Some people didn’t realize it only worked on Samsung’s 4.0 browser and left 1-star reviews. The update was meant to better highlight the app’s requirements.)
Meanwhile, as of the time of writing, other ad blockers are still live, including Crystal and Adblock Plus (Samsung Browser). However, that may not be the case for long.
Crystal’s developer Dean Murphy also just submitted an update that’s just been declined by Google’s app review team for the same reason cited above. Again, Google references section 4.4 of the Developer Agreement as the reason for stopping the update from going live.
“I have appealed the update rejection, as I assume that I am rejected for ‘interfering’ with Samsung Internet Browser, citing the developer documentation that Samsung have for the content blocking feature,” explains Murphy. “I’m still awaiting their reply.”
Wow, that was fast. Crystal was still there on Wednesday. This is going to ratchet up tensions between Google and Samsung (again); in the comments on the Verge article on this topic (which has less detail) there are people who switched to iOS because of adblocking, or are considering moving back because they can’t get it on Android. A small but possibly significant group.
Google has clearly set its face against adblocking on mobile, but the pressure is starting to build up behind the dam.
link to this extract
About » DeepDetect
DeepDetect (http://www.deepdetect.com/) is a machine learning API and server written in C++11. It makes state of the art machine learning (such as deep learning) easy to work with and integrate into existing applications. Its goal is to simplify and secure both the development and production phases by using possibly different servers and passing models from one to the other.
It originates from the need for industries, businesses and researchers to quickly fit a machine learning pipeline into existing applications, starting with well-known models, and moving toward more targeted ones while measuring accuracy.
DeepDetect allows this by coupling a generic API and a server with high performance machine learning libraries. At the moment it has support for the deep learning library Caffe. More libraries are to be supported in order to span over a larger set of common use cases.
There are free (even for commercial use) models that are downloadable from the site. This lies just over my event horizon for understanding – but reading the details about “getting started” puts me in mind of people feeding a giant brain, or disembodied intelligence, and that gives me pause.
But this stuff is going to be everywhere in two years.
link to this extract
Would you be sad to see Sony withdraw from the tablet market? » Xperia Blog
given the challenging smartphone market, as evidenced by last week’s results there is no guarantee that Sony will continue to cater for the tablet market. A recent Japanese blog post by a Sony store manager speculated that the company may withdraw from the tablet market after receiving marketing material suggesting so.
The news would not surprise us, after all, we know that tablets made just 5% of Sony Mobile’s revenues back in 2013 and that was expected to shrink even further. Given the R&D costs of developing and supporting new devices, Sony may feel that producing another tablet for 2016 might not be commercially viable.
I didn’t know that about the tablet revenues; apparently they’re meant to be down to 3-4% now. The question is whether they generate more than 0% in profit – because they must be eating up R+D time and money, which is opportunity cost that Sony probably can’t afford.
link to this extract
The joy of shortcuts » Allen Pike
Next January, Parse is shutting down. The successful Parse apps will get moved to a custom backend like ours was, perhaps using Parse’s excellent open-source server and migration tool. The unsuccessful Parse apps will die. Hundreds of thousands of unsuccessful Parse apps will perish. Like links to long-dead Geocities pages, dead mobile apps that relied on Parse will linger in the App Stores for years, slowly accumulating one-star reviews.
As much as Parse will try to get the word out that they’re shutting down, many apps’ owners don’t even know that they’re reliant on Parse. Parse’s overly generous free plan made them popular with freelancers and consultants building quick app backends for their clients. Many of those clients don’t know what Parse is, let alone that the little app they commissioned a couple years ago is a ticking time bomb.
How many iOS apps, how many Android apps relied on Parse? There needs to be an enumeration.
link to this extract
How the iPhone 6 ruined Apple » All this
While it’s certainly possible that the great days of iPhone sales growth are over, I wouldn’t make that prediction just yet. In fact, I was surprised to learn that iPhone sales were merely flat. I was expecting a decline—not because the iPhone is losing popularity, but because the iPhone 6’s first quarter of sales was such a gigantic leap upward. The pent-up demand for a larger iPhone caused sales to increase nearly 50% year over year, to 74.47m from 51.03m the year before. This was the biggest percentage jump in year-over-year sales since the introduction of the 4S (which was goosed a bit because the 4S was delayed). I just didn’t think the 6S could keep up with that. And maybe it won’t.
But look at how things were going before the iPhone 6. Had the trend of 2012–2014 continued through 2015, iPhone sales last quarter would have been 65–70 million. Instead they were just under 75m. It’s only in comparison to the huge holiday quarter of 2014 that last quarter looks dull.
I’m reminded of the devotion climate change deniers had to the year 1998. Because of an intense El Niño that year, global temperatures rose well above the trend line, and it remained the hottest year on record for several years. Deniers hit upon this fact, and claimed that global warming had stopped, even though the overall warming trend had continued. The iPhone 6 was Apple’s El Niño.
Magic Leap Just Landed an Astounding Amount of VC Money » WIRED
Jessi Hempel on the company which has just raised $793.5m in a C round from Alibaba and others (Google and Qualcomm were already on board):
Many believe Magic Leap’s technology—along with a handful of competing virtual and augmented reality products—will usher in a sea change in how we use computers. By placing sensors everywhere and processing the volumes of data they produce, it’s possible to create better immersive environments and believable layers of digital images on top of the physical world. Facebook, Samsung, and Microsoft are creating competing technology and have chosen to make their headsets available even as they’re engineering the products. Google is also beefing up its virtual reality team, and Apple is also reportedly getting into the action. Magic Leap claims to be using a different technology to achieve its effect, and it’s keeping its efforts mostly secret.
The company has made converts out of many of those who have seen demos. New Zealand design studio Weta Workshops has teamed up with Magic Leap to build games. Science fiction writer Neal Stephenson joined the company as its chief futurist. Last fall, Google led a $542m investment, bringing its previous funding total to $592m. But so far, it hasn’t been clear when Magic Leap’s tech will be available for consumers.
Expectations around this are going to be huge, which usually leads to disappointments of the same size. Shipping product matters; having that much of a cash pillow can’t be good, because it won’t help the financial discipline needed to make things (of whatever sort) to a price, for a user, to a specification. Don’t forget the lesson of Leap Motion – big hype, big letdown.
link to this extract
Startup lands $100m to challenge smartphone superpowers Apple and Google » CNET
Acadine, which CNET previously reported was initially known by the placeholder name Gone Fishing, plans to build an operating system for smartphones, tablets, wearable devicess and the Internet of Things.
That’ll be a tough challenge. But [fouder Li] Gong believes Acadine’s generous funding [from a Chinese state-controlled company], fast development and international reach will mean consumers finally will see the alternative to Apple and Google that so many other companies have failed to build.
And of course that means his startup and its investors will reap the rewards. “Owning an OS is extremely important if you can do it,” Gong said. “It’s very profitable if you can do it.”
Potential sources of money, Gong said, include being paid to promote services like search, storage, music streaming and e-commerce; revenue sharing from those services when customers pay to use them; and fees generated by advertising and game sales. All of those, though, depend on Acadine succeeding in finding and exploiting gaps where existing OSes are weak then expanding from there to a large user base.
The list of mobile operating systems that have struggled to compete against Android and iOS and gain that large population of users is long: Microsoft’s Windows Phone, Samsung’s Tizen, Jolla’s Sailfish OS, Canonical’s Ubuntu, Hewlett-Packard’s WebOS, BlackBerry’s BlackBerry OS and Mozilla’s Firefox OS. This last project is the one Gong led at Mozilla until he left in April, and it’s the starting point for H5OS.
One hates to say “a fool and his $100m are soon parted”, but it’ll do.
link to this extract
Errata, corrigenda and ai no corrida: I was going to include a link to a video of a male cyclist who was suspected of having a motor in his bicycle (and man, it looked fishy) but realised it is a rabbit hole one would never emerge from.