26 August 2019

IBM’s quantum-resistant magnetic tape storage is not actually snake oil


Usually when someone in tech says the word “quantum,” I put my hands on my ears and sing until they go away. But while IBM’s “quantum computing safe tape drive” nearly drove me to song, when I thought about it, it actually made a lot of sense.

First of all, it’s a bit of a misleading lede. The tape is not resistant to quantum computing at all. The problem isn’t that qubits are going to escape their cryogenic prisons and go interfere with tape drives in the basement of some datacenter or HQ. The problem is what these quantum computers may be able to accomplish when they’re finally put to use.

Without going too deep down the quantum rabbit hole, it’s generally acknowledged that quantum computers and classical computers (like the one you’re using) are good at different things — to the point where in some cases, a problem that might take incalculable time on a traditional supercomputer could be done in a flash on quantum. Don’t ask me how — I said we’re not going down the hole!

One of the things quantum is potentially very good at is certain types of cryptography: It’s theorized that quantum computers could absolutely smash through many currently used encryption techniques. In the worst case scenario, that means that if someone got hold of a large cache of encrypted data that today would be useless without the key, a future adversary may be able to force the lock. Considering how many breaches there have been where the only reason your entire life wasn’t stolen was because it was encrypted, this is a serious threat.

quantum tapeIBM and others are thinking ahead. Quantum computing isn’t a threat right now, right? It isn’t being seriously used by anyone, let alone hackers. But what if you buy a tape drive for long-term data storage today, and then a decade from now a hack hits and everything is exposed because it was using “industry standard” encryption?

To prevent that from happening, IBM is migrating its tape storage over to encryption algorithms that are resistant to state of the art quantum decryption techniques — specifically lattice cryptography (another rabbit hole — go ahead). Because these devices are meant to be used for decades if possible, during which time the entire computing landscape can change. It will be hard to predict exactly what quantum methods will emerge in the future, but at the very least you can try not to be among the low-hanging fruit favored by hackers.

The tape itself is just regular tape. In fact, the whole system is pretty much the same as you’d have bought a week ago. All the changes are in the firmware, meaning earlier drives can be retrofitted with this quantum-resistant tech.

Quantum computing may not be relevant to many applications today, but next year who knows? And in ten years, it might be commonplace. So it behooves companies like IBM, which plan to be part of the enterprise world for decades to come, to plan for it today.


Read Full Article

Amazon’s free streaming service IMDb TV comes to mobile devices


IMDb TV, the free ad-supported streaming service launched by Amazon-owned IMDb at the beginning of the year (originally called Freedive), is today arriving on mobile devices. With the updated version of iOS and Android IMDb app rolling out now, users can stream from the app’s growing library of free movies and TV series.

Prior to IMDb TV’s launch, the movie website had experimented with video content in the form of trailers, celebrity interviews, and other short-form series. But consumers today are more interested in services where they can stream premium content for free, without a subscription — as they can on IMDb TV competitors like Walmart-owned Vudu’s “Movies on Us,” Tubi, or The Roku Channel, for example.

At launch, IMDb TV offered a collection of TV shows like Fringe, Heroes, The Bachelor and Without a Trace, as well as Hollywood movies like Awakenings, Foxcatcher, Memento, Monster, Run Lola Run, The Illusionist, The Last Samurai, True Romance, and others.

This summer, it expanded its lineup through new deals with Warner Bros., Sony Pictures Entertainment and MGM Studios.

This brought movies like Captain Fantastic and La La Land to the service, the latter which has since become one of the service’s most-streamed movies this summer. Other popular titles included Jerry Maguire, Practical Magic, A Knight’s Tale, Drive, Max, Step Dogs, Zookeeper, Paddington, and NeverEnding Story.

More recent deals with Paramount and Lionsgate have also brought new content to IMDb TV, like Silver Linings Playbook, Age of Adaline, In the Heart of the Sea, and the TV show, The Middle.

The company hasn’t said how many customers IMDb TV has, but the service has benefitted from integrations with Amazon’s Fire TV.

Earlier this year, Marc Whitten, vice president of Fire TV, noted that Fire TV customers’ use of free, ad-supported apps had increased by over 300% during the last year. IMDb TV, is expected to contribute to that, with its placement on the “Your Apps & Channels” row on Fire TV and its availability as a free channel within the Prime Video app.

The updated iOS and Android IMDb app is rolling out starting today, the company says.


Read Full Article

Bi-Tempered Logistic Loss for Training Neural Nets with Noisy Data




The quality of models produced by machine learning (ML) algorithms directly depends on the quality of the training data, but real world datasets typically contain some amount of noise that introduces challenges for ML models. Noise in the dataset can take several forms from corrupted examples (e.g., lens flare in an image of a cat) to mislabelled examples from when the data was collected (e.g., an image of cat mislabelled as a flerken).

The ability of an ML model to deal with noisy training data depends in great part on the loss function used in the training process. For classification tasks, the standard loss function used for training is the logistic loss. However, this particular loss function falls short when handling noisy training examples due to two unfortunate properties:
  1. Outliers far away can dominate the overall loss: The logistic loss function is sensitive to outliers. This is because the loss function value grows without bound as the mislabelled examples (outliers) are far away from the decision boundary. Thus, a single bad example that is located far away from the decision boundary can penalize the training process to the extent that the final trained model learns to compensate for it by stretching the decision boundary and potentially sacrificing the remaining good examples. This “large-margin” noise issue is illustrated in the left panel of the figure below.
  2. Mislabeled examples nearby can stretch the decision boundary: The output of the neural network is a vector of activation values, which reflects the margin between the example and the decision boundary for each class. The softmax transfer function is used to convert the activation values into probabilities that an example will belong to each class. As the tail of this transfer function for the logistic loss decays exponentially fast, the training process will tend to stretch the boundary closer to a mislabeled example in order to compensate for its small margin. Consequently, the generalization performance of the network will immediately deteriorate, even with a low level of label noise (right panel below).
We visualize the decision surface of a 2-layered neural network as it is trained for binary classification. Blue and orange dots represent the examples from the two classes. The network is trained with logistic loss under two types of noisy conditions: (left) large-margin noise and (right) small-margin-noise.
We tackle these two problems in a recent paper by introducing a “bi-tempered” generalization of the logistic loss endowed with two tunable parameters that handle those situations well, which we call “temperatures”—t1, which characterizes boundedness, and t2 for tail-heaviness (i.e. the rate of decline in the tail of the transfer function). These properties are illustrated below. Setting both t1 and t2 to 1.0 recovers the logistic loss function. Setting t1 lower than 1.0 increases the boundedness and setting t2 greater than 1.0 makes for a heavier-tailed transfer function. We also introduce this interactive visualization which allows you to visualize the neural network training process with the bi-tempered logistic loss.
Left: Boundedness of the loss function. When t1 is between 0 and 1, exclusive, only a finite amount of loss is incurred for each example, even if they are mislabeled. Shown is t1 = 0.8. Right: Tail-heaviness of the transfer function. The heavy-tailed transfer function applies when t2 = > 1.0 and assigns higher probability for the same amount of activation, thus preventing the boundary from drawing closer to the noisy example. Shown is t2 = 2.0.
To demonstrate the effect of each temperature, we train a two-layer feed-forward neural network for a binary classification problem on a synthetic dataset that contains a circle of points from the first class, and a concentric ring of points from the second class. You can try this yourself on your browser with our interactive visualization. We use the standard logistic loss function, which can be recovered by setting both temperatures equal to 1.0, as well as our bi-tempered logistic loss for training the network. We then demonstrate the effects of each loss function for a clean dataset, a dataset with small-margin noise, large-margin noise, and a dataset with random noise.
Logistic vs. bi-tempered logistic loss: (a) noise-free labels, (b) small-margin label noise, (c) large-margin label noise, and (d) random label noise. The temperature values (t1, t2) for the tempered loss are shown above each figure. We find that for each situation, the decision boundary recovered by training with the bi-tempered logistic loss function is better than before.
Noise Free Case:
We show the results of training the model on the noise-free dataset in column (a), using the logistic loss (top) and the bi-tempered logistic loss (bottom). The white line shows the decision boundary for each model. The values of (t1, t2), the temperatures in the bi-tempered loss function, are shown below each column of the figure. Notice that for this choice of temperatures, the loss is bounded and the transfer function is tail-heavy. As can be seen, both losses produce good decision boundaries that successfully separates the two classes.

Small-Margin Noise:
To illustrate the effect of tail-heaviness of the probabilities, we artificially corrupt a random subset of the examples that are near the decision boundary, that is, we flip the labels of these points to the opposite class. The results of training the networks on data with small-margin noise using the logistic loss as well as the bi-tempered loss is shown in column (b).

As can be seen, the logistic loss, due to the lightness of the softmax tail, stretches the boundary closer to the noisy points to compensate for their low probabilities. On the other hand, the bi-tempered loss using only the tail-heavy probability transfer function by adjusting t2 can successfully avoid the noisy examples. This can be explained by the heavier tail of the tempered exponential function, which assigns reasonably high probability values (and thus, keeps the loss value small) while maintaining the decision boundary away from the noisy examples.

Large-Margin Noise:
Next, we evaluate the performance of the two loss functions for handling large-margin noisy examples. In (c), we randomly corrupt a subset of the examples that are located far away from the decision boundary, the outer side of the ring as well as points near the center).

For this case, we only use the boundedness property of the bi-tempered loss, while keeping the softmax probabilities the same as the logistic loss. The unboundedness of the logistic loss causes the decision boundary to expand towards the noisy points to reduce their loss values. On the other hand, the bounded bi-tempered loss, bounded by adjusting t1, incurs a finite amount of loss for each noisy example. As a result, the bi-tempered loss can avoid these noisy examples and maintain a good decision boundary.

Random Noise:
Finally, we investigate the effect of random noise in the training data on the two loss functions. Note that random noise comprises both small-margin and large-margin noisy examples. Thus, we use both boundedness and tail-heaviness properties of the bi-tempered loss function by setting the temperatures to (t1, t2) = (0.2, 4.0).

As can be seen from the results in the last column, (d), the logistic loss is highly affected by the noisy examples and clearly fails to converge to a good decision boundary. On the other hand, the bi-tempered can recover a decision boundary that is almost identical to the noise-free case.

Conclusion
In this work we constructed a bounded, tempered loss function that can handle large-margin outliers and introduced heavy-tailedness in our new tempered softmax function, which can handle small-margin mislabeled examples. Using our bi-tempered logistic loss, we achieve excellent empirical performance on training neural networks on a number of large standard datasets (please see our paper for full details). Note that the state-of-the-art neural networks have been optimized along with a large variety of variables such as: architecture, transfer function, choice of optimizer, and label smoothing to name just a few. Our method introduces two additional tunable variables, namely (t1, t2). We believe that with a systematic “joint optimization” of all commonly tried variables, significant further improvements can be achieved in conjunction with our loss function. This is of course a more long-term goal. We also plan to explore the idea of annealing the temperature parameters over the training process.

Acknowledgements:
This blogpost reflects work with our co-authors Manfred Warmuth, Visiting Researcher and Tomer Koren, Senior Research Scientist, Google Research. Preprint of our paper is available here, which contains theoretical analysis of the loss function and empirical results on standard datasets at scale.

The Void’s Curtis Hickman on scaling, creative IP and the future of VR experiences


What can you do with virtual reality when you have complete control of the physical space around the player? How “real” can virtual reality become?

That’s the core concept behind The Void. They take over retail spaces in places like Downtown Disney and shopping malls around the country and turn them into virtual reality playgrounds, They’ve got VR experiences based on properties like Star Wars, Ghostbusters, and Wreck-It Ralph; while these big names tend to be the main attractions, they’re dabbling with creating their own original properties, too.

By building both the game environment and the real-world rooms in which players wander, The Void can make the physical and virtual align. If you see a bench in your VR headset, there’s a bench there in the real world for you to sit on; if you see a lever on the wall in front of you, you can reach out and physically pull it. Land on a lava planet and heat lamps warm your skin; screw up a puzzle, and you’ll feel a puff of mist letting you know to try something else.

At $30-$35 per person for what works out to be a roughly thirty-minute experience (about ten of which is watching a scene-setting video and getting your group into VR suits), it’s pretty pricey. But it’s also some of the most mind-bending VR I’ve ever seen.

The Void reportedly raised about $20 million earlier this year and is in the middle of a massive expansion. It’s more than doubling its number of locations, opening 25 new spots in a partnership with the Unibail-Rodamco-Westfield chain of malls.

I sat down to chat with The Void’s co-founder and Chief Creative Officer, Curtis Hickman, to hear how they got started, how his background (in stage magic!) comes into play here, how they came to work with massive properties like Ghostbusters and Star Wars, and where he thinks VR is going from here.

Greg Kumparak: Tell me a bit about yourself. How’d you get your start? How’d you get into making VR experiences?


Read Full Article

Udacity names former LendingTree executive to CEO post


Online education startup Udacity has hired former LendingTree executive Gabriel Dalporto as its new CEO, an appointment that follows months of layoffs and a restructuring directed by the company’s co-founder and executive chairman Sebastian Thrun.

Dalporto comes to Udacity after seven years at LendingTree, where he served in numerous positions, including chief marketing officer and chief financial officer. Dalporto stepped down as CFO in 2017 to join the company’s board and become executive advisor to the CEO. Dalporto left the executive advisor job in 2018, but remains on the board.

Thrun, who stepped in as CEO after Vishal Makhijani left the top post in October 2018, will stay on as executive chairman.

“He’s extremely strategic and pragmatic,” Thrun said in a recent interview, describing Dalporto.

Dalporto is known for his turnaround skills. But the new CEO says his focus at Udacity won’t be slashing costs and other activities often associated with that skill set.

“I was hired as a growth executive; I was not hired to be a turnaround executive,” Dalporto told TechCrunch.

udacity Team H IMG 5639

Dalporto isn’t ready to provide details of his plans as CEO. Monday is his first day at the startup. But he will likely focus on growth areas such as the startup’s enterprise and government programs, as well as retaining and recapturing students into the Udacity ecosystem. Udacity’s enterprise clients include AT&T, Airbus, Audi, BMW, Capital One, Cisco and the Royal Bank of Scotland. It also has government relationships with Australia, the MENA region and New Zealand.

Dalporto is coming into a startup that is leaner and more productive, in terms of launching new nanodegrees, than it was a year ago.  It’s also cash-flow positive, according to Thrun, who has spent 2019 revamping the company.

When Thrun took over the CEO post, he found a company that had grown too quickly and was burdened by its own bureaucracy. Udacity, which specializes in “nanodegrees” on a range of technical subjects that include AI, deep learning, digital marketing, VR and computer vision, was struggling because of runaway costs and other inefficiencies. Its nanodegree programs, which had grown in 2017, became sluggish in 2018. 

Staff reductions soon followed as Thrun sought to get a handle on costs. About 130 people were laid off and other open positions were left vacant. Thrun then cut further in April. About 20% of the staff was laid off and operations were restructured in an effort to bring costs in line with revenue without curbing growth. The company streamlined its marketing efforts and downsized and consolidated office space. As of April, the startup employs 300 full-time equivalent employees and about 60 contractors.

Other changes included the launch of a global technical mentoring program, switching its direct-to-student business from fixed to monthly subscription pricing to incentivize individuals to move through courses faster. Lalit Singh, who joined Udacity in February as chief operating officer, has been critical to the turnaround, according to Thrun.

Its productivity has also improved. In first six months of 2019, Udacity launched 12 new nanodegree programs compared to just 8 in all of 2018.

“In the three months since we’ve initiated these changes, the consumer business has grown by more than 60%,” Thrun wrote in a blog post Monday announcing the changes.

Udacity’s enterprise and government programs have also grown, with bookings increasing by more than 100% year over year.


Read Full Article

Facebook succeeds in blocking German FCO’s privacy-minded order against combining user data


Facebook has succeeded in blocking a pioneering order by Germany’s Federal Cartel Office earlier this year that would have banned it from combining data on users across its own suite of social platforms — Facebook, Instagram and WhatsApp — without their consent.

Pioneering because the antitrust regulator had liaised with EU privacy authorities during a long-running investigation of Facebook’s data-gathering activities — leading it to conclude that Facebook’s conduct in the German market where it also deemed it to hold a monopoly position amounted to “exploitative abuse”.

The Bundeskartellamt (FCO) order had been likened to a structural separation of Facebook’s businesses at the data level.

Facebook appealed, delaying application of the order, and today’s ruling by the Dusseldorf court grants a suspension (press release in German) — essentially kicking the matter into very long legal grass.

The FCO has a month to lodge an appeal. A spokeswoman confirmed to TechCrunch is will do so. But with the order suspended pending what could be years of appeals there’s little near-term prospect of any change to how Facebook does business based on this particular regulatory intervention.

It’s undoubtedly a major victory for Facebook — to win at the very first appeals layer — and a major blow for regulatory ‘innovation’ (for what of a better word) which sought to evolve the interpretation of current competition law to respond to the outgrowth and dominance of surveillance-based digital business model via applying privacy-focused conditions to data processing.

Europe’s data protection regulators do have the power to order the suspension of infringing data processing, under the bloc’s updated privacy framework (GDPR).

But so far very such orders are as rare as hen’s teeth — barring a recent threat to Google also by a German privacy regulator. (Just the threat of an order in that case triggered a voluntary suspension of the data processing in question.)

This made the FCO’s order against Facebook all the more notable for boldness and forethought. And means Facebook’s success in cutting it down at the first legal hurdle is a depressing result for those in the EU hoping platform power linked to privacy-hostile surveillance of Internet users might be regulated in a meaningful timeframe via an existing antitrust lens.

The European Commission’s own ‘big tech’ antitrust interventions have so far focused their attention elsewhere, in addition to taking years to conclude.

Commenting on the Düsseldorf Higher Regional Court’s decision today in a statement, FCO president Andreas Mundt said: “Data and data handling are decisive factors for competition in the digital economy. The Higher Regional Court of Düsseldorf has today responded differently than the Bundeskartellamt to key legal issues. These legal issues are highly significant for the future state of competition in the digital economy. We are convinced that we can act in this area based on the existing antitrust law. For this reason, we are going to appeal on points of law to the Federal Court of Justice to clarify these issues.”

We’ve also reached out to Facebook for comment.

Professor Rupprecht Podszun, a chair for civil law, German and European competition law at Heinrich Heine University, who has been following the FCO’s intervention, dubs the court ruling a “major blow” for the regulator.

“The FCO had accused Facebook of abusing its dominant position by unlawfully gathering and combining user data. Thus it had ordered Facebook to change its Terms & Conditions within a year. The judges from Düsseldorf have stopped enforcement of this decision now. They have serious doubts as to the lawfulness of the decision,” he said via email. “The case is regarded as a landmark case against the digital giants and it had gained worldwide recognition. To fail at the Düsseldorf court, at the very first step, is a bitter result.”

Podszun said the Düsseldorf court did not accept it follows from a possible violation of privacy rules that it is automatically a violation of antitrust rules if a dominant company is acting. That would require the court to see competitive damage — which it did not in this case.

Additionally, the court took the view that users decide autonomously whether they agree with Facebook’s T&Cs when signing up for the service. It also did not agree that consumers are exploited by Facebook’s data collection since they could continue to make the same data available to other companies.

From here on in he believes legal back and forth is likely to take years — hence, even if the FCO were to prevail at a higher court in future the impact on Facebook’s business at that point would likely be long out of date. (Meanwhile, earlier this year it emerged that Facebook is working on merging the back-end infrastructure of its three social networks — seeking to further collapse cross-platform user privacy, even as its scrambles discrete business units in a way that would complicate any regulatory order to break apart its business.)

“The Cartel Office had shown courage in its decision and had explored new paths. The Higher Regional
Court did not follow this reasoning. The FCO took a long shot by integrating a privacy investigation into the competition assessment. I have a lot of sympathy for that, because data has become a crucial competitive factor. Thus, I think that data collection must be a topic for antitrust law,” said Podszun.

“The law is at its limits with the internet giants. It is too slow. A final decision in a few years on the privacy terms of Facebook is too late either way. Before taking the decision, the FCO had investigated the case for three years. The Google Shopping procedure of the European Commission took seven years. You cannot tame these companies with such proceedings and lengthy litigation in court.”

“The decision is a wake-up call to legislators: If you want to regulate Google, Amazon, Facebook & Co., the existing tools are not enough,” he added. “A new version of the Antitrust Act is currently pending in Germany. This is an opportunity to change the legal bases. Also, the authorities for data protection need to step up their efforts – they seem to lack the bravery of the antitrust watchdog.”

Asked how legal bases need to change to enable local antitrust law to respond intelligently to data-mining platform giants, Podszun suggested four areas of focus — telling TechCrunch:

  • Competition law needs to get away from traditional market definition. There should be a rule that the authorities can interfere with companies like GAFA [Google, Apple, Facebook, Amazon] in cases where they move into new markets where they are not yet dominant but can easily tip the market. Conglomerate effects and digital ecosystems currently are a blind spot in competition law
  • There may be room for a new example of what constitutes an abuse in digital markets
  • The German Competition Office should have powers in consumer law fields (currently, there is no public enforcement of economic consumer protection issues in Germany). An integrated approach with consumer and competition issues could be helpful (including privacy, possibly). Privacy enforcers are particularly weak in Germany
  • Procedures need to be speeded up, e.g. by stricter time limits, less haggling over access to file, more technically savvy staff and more priority-setting by the authorities

“All very difficult – but it’s vital to have some fresh air here,” he added. “Whether this would have helped in the case under debate is a different question.”


Read Full Article

Google falls to third place in worldwide smart speaker market


The global smart speaker market grew 55.4% in the second quarter to reach 26.1 million shipments, according to a new report from Canalys. Amazon continued to lead the race, accounting for 6.6 million units shipped in the quarter. Google, however, fell to the third spot as China’s Baidu surged ahead. Baidu in Q2 grew a sizable 3,700% to reach 4.5 million units, overtaking Google’s 4.3 million units shipped.

China’s market overall doubled its quarterly shipments to 12.6 million units, or more than twice the U.S.’s 6.1 million total. The latter represents a slight (2.4%) decline since the prior quarter.

Baidu’s growth in the quarter was attributed to aggressive marketing and go-to-market campaigns. It was particularly successful in terms of smart displays, which accounted for 45% of the products it shipped.

“Local network operator’s interests on the [smart display] device category soared recently. This bodes well for Baidu as it faces little competition in the smart display category, allowing the company to dominate in the operator channel,” noted Canalys Research Analyst Cynthia Chen.

Meanwhile, Google was challenged by the Nest rebranding in Q2, the analyst firm said.

The report also suggested that Google should to introduce a revamped smart speaker portfolio to rekindle consumer interest. The Google Home device hasn’t been updated since launch — still sporting the air freshener-style looks it had back in 2016. And the Google Home mini hasn’t received much more than a color change.

Instead, Google’s attention as of late has been on making it easier for device manufacturers to integrate with Google Assistant technology, in addition to its increased focus on smart displays.

Amazon, by comparison, has updated its Echo line of speakers several times while expanding Alexa to devices with screens like the Echo Spot and Show, and to those without like the Echo Plus, Echo Dot, Echo Auto, and others — even clocks and microwaves, as sort of public experiments in voice computing.

That said, both Amazon and Google turned their attention to non-U.S. markets in Q2, the report found. 50% of Amazon’s smart speaker shipments were outside the U.S. in Q2, up from 32% in Q2 last year. And 55% of Google’s shipments were outside the U.S., up from 42% in Q2 2018.

table ifnal final

Beyond the top 3 — Amazon, Baidu and now No. 3 Google — the remaining top 5 included Alibaba and Xiaomi, with 4.1 million and 2.8 million units shipped in Q2, respectively.

The rest of the market, which would also include Apple’s HomePod, totaled 3.7 million units.

 

 


Read Full Article

Google falls to third place in worldwide smart speaker market


The global smart speaker market grew 55.4% in the second quarter to reach 26.1 million shipments, according to a new report from Canalys. Amazon continued to lead the race, accounting for 6.6 million units shipped in the quarter. Google, however, fell to the third spot as China’s Baidu surged ahead. Baidu in Q2 grew a sizable 3,700% to reach 4.5 million units, overtaking Google’s 4.3 million units shipped.

China’s market overall doubled its quarterly shipments to 12.6 million units, or more than twice the U.S.’s 6.1 million total. The latter represents a slight (2.4%) decline since the prior quarter.

Baidu’s growth in the quarter was attributed to aggressive marketing and go-to-market campaigns. It was particularly successful in terms of smart displays, which accounted for 45% of the products it shipped.

“Local network operator’s interests on the [smart display] device category soared recently. This bodes well for Baidu as it faces little competition in the smart display category, allowing the company to dominate in the operator channel,” noted Canalys Research Analyst Cynthia Chen.

Meanwhile, Google was challenged by the Nest rebranding in Q2, the analyst firm said.

The report also suggested that Google should to introduce a revamped smart speaker portfolio to rekindle consumer interest. The Google Home device hasn’t been updated since launch — still sporting the air freshener-style looks it had back in 2016. And the Google Home mini hasn’t received much more than a color change.

Instead, Google’s attention as of late has been on making it easier for device manufacturers to integrate with Google Assistant technology, in addition to its increased focus on smart displays.

Amazon, by comparison, has updated its Echo line of speakers several times while expanding Alexa to devices with screens like the Echo Spot and Show, and to those without like the Echo Plus, Echo Dot, Echo Auto, and others — even clocks and microwaves, as sort of public experiments in voice computing.

That said, both Amazon and Google turned their attention to non-U.S. markets in Q2, the report found. 50% of Amazon’s smart speaker shipments were outside the U.S. in Q2, up from 32% in Q2 last year. And 55% of Google’s shipments were outside the U.S., up from 42% in Q2 2018.

table ifnal final

Beyond the top 3 — Amazon, Baidu and now No. 3 Google — the remaining top 5 included Alibaba and Xiaomi, with 4.1 million and 2.8 million units shipped in Q2, respectively.

The rest of the market, which would also include Apple’s HomePod, totaled 3.7 million units.

 

 


Read Full Article

How the West can adapt to a rising Asia | Kishore Mahbubani

How the West can adapt to a rising Asia | Kishore Mahbubani

As Asian economies and governments continue to gain power, the West needs to find ways to adapt to the new global order, says author and diplomat Kishore Mahbubani. In an insightful look at international politics, Mahbubani shares a three-part strategy that Western governments can use to recover power and improve relations with the rest of the world.

Click the above link to download the TED talk.

Downloads need to rank No. 1 on App Store is down 30% since 2016 for apps, up 47% for games


With the App Store’s big makeover in fall 2017, Apple attempted to shift consumers’ attention away from the Top Charts and more towards editorial content. But app developers still want to make it to the No. 1 position. According to new research from app store intelligence firm Sensor Tower, it’s become easier for non-game apps over the past few years to achieve the top ranking.

Specifically, the firm found that the median number of daily downloads required for non-game applications on the U.S. iPhone App Store to reach No. 1 decreased around 34% from 136,000 to 90,000 in 2018, then increased a little more than 4% to 94,000 this year.

At the same time, the number of non-game installs on the U.S. App Store had increased by 33% between Q1 2016 and Q1 2019.

These findings, Sensor Tower suggests, indicate that the U.S. market for the top social and messaging apps has become saturated, with downloads for top apps like Facebook and Messenger decreasing over time. In addition, no other apps have found the same level of success that Snapchat and Bitmoji did back in 2016 and 2017, the report adds.

median downloads no 1 ios

For example, Messenger saw 5 million U.S. App Store installs in November 2016 while Bitmoji and Snapchat passed 5 million installs in August 2016 and March 2017, respectively. And no other non-game app has topped 3.5 million installs in a single month since March 2017.

Meanwhile, the decline in downloads needed to reach the No. 1 spot on Google Play was even more significant.

The median daily downloads for the top non-game app decreased by 65% from 209,000 in 2016 to 74,000 so far in 2019.

Similarly, the store saw a decrease in installs among top apps, including Messenger, Facebook, Snapchat, Pandora and Instagram. Messenger, for example, saw its yearly installs fall by 68% from nearly 80 million in 2016 to 26 million in 2018.

Games

With mobile games, however, it’s a different story across both app stores.

On the Apple App Store, it has taken 174,000 downloads for a game to reach the top of the rankings on any given day in 2019 — 85% more the 94,000 installs required for non-game app to reach the top of the charts.

This figure also represents an increase of 47% compared to the 118,000 median daily downloads required to top the charts back in 2016, Sensor Tower said.

median downloads no 1 google play

In part, this trend is due to the rise of hyper-casual gaming. So far in 2019, 28 games have reached the No. 1 position on the U.S. App Store, with hyper-casual games making up all but 4 of those. And of those four, only Harry Potter: Wizards Unite spent more than one day at the top of the charts. Meanwhile, hyper-casual games like aquapark.io and Colorbump 3D have spent 25 and 30 days at No. 1, respectively.

On Google Play, the median daily installs to reach the No. 1 position increased from 70,000 in 2017 to 116,000 so far in 2019, or 66% growth. Overall game downloads, however, decreased 16% from 646 million in Q1 2017 to 544 million in Q1 2019.

Similarly, 21 out of the 23 games that reached the top spot this year have been hyper-casual titles, like Words Story or Traffic Run.

Breaking the Top 10

While topping the charts has gotten easier for non-game apps over the years, breaking into the top 10 has gotten more difficult. Median U.S. daily installs for the No. 10 free non-game app increased 11% from 44,000 in 2016 to 49,000 in 2019.

median downloads top 10 ios

On Google Play, meidan daily installs for non-game apps fell nearly 50% from 55,000 median daily installs in 2016 to 31,000 in 2019.

For games, the No. 10 game’s spot on the App Store had 25,000 median daily installs in 2016 to 43,000 so far in 2019, and Google Play saw 26% growth from 27,000 to 34,000 during the same period.

median downloads top 10 google play

Categories making the Top 10

In terms of breaking into the top 10 by category, Photo & Video apps on the App Store present the most challenge. The category where YouTube, Instagram, TikTok and Snapchat reside saw a median daily amount of more than 16,000 downloads for the No. 10 app.

This was followed by Shopping (15,300 daily downloads for the No. 10 app), Social Networking (14,500), Entertainment (12,600), and Productivity (12,400).

On Google Play, Entertainment apps — like Hulu, Netflix and Bitmoji — need around 17,100 U.S. installs in a day to reach the top 10. This is followed by Shopping (10,800), Social (9,100), Music (8,200), and Finance (8000).

Beyond the U.S.

Outside the U.S., a non-game app needs approximately 91,000 downloads to reach the top 10 on the App Store in China — higher than the 49,000 installs needed in the U.S. For games, the U.S. is the most difficult to crack the top 10, with a median of 43,000 daily downloads for the No. 10 game.

median downloads top 10 by country ios

On Google Play, India required the most downloads to reach the top 10 with apps needing 256,000 downloads in a day and games needing 117,000 downloads.

median downloads top 10 by country google play

Of course, the App Store’s ranking algorithms — nor Google Play’s algorithms — rely on downloads alone to determine an app’s ranking. Apple takes into consideration downloads and velocity, among other undocumented factors. Google Play does something similar.

But these days, developers are more concerned with showing up highly ranked in app store searches than they are on top charts, where they’ll need to consider numerous other factors beyond downloads — like keywords, description, user engagement, and even app quality, among other things.

 

 


Read Full Article

For the Latest in Tech Trends, Follow MakeUseOf’s IFA Coverage (Sep 4-8)


IFA is Europe’s equivalent to CES in Las Vegas. MakeUseOf will report from the Berlin-based consumer electronics show from September 4 to 8. You can expect to see new TVs, laptops, headphones, smartphones, drones, and smart home products.

What Is IFA?

Originally launched in Berlin, Germany in 1924, IFA (Internationale Funkausstellung; international radio exhibit) is one of the world’s first technology exhibitions and Europe’s biggest trade show today. Every year, IFA presents the latest consumer technology, including product launches from leading electronics brands.

Unlike CES, IFA is open to the public. You’ll see a diverse audience roaming the show floors. Like other exhibitions, IFA forbids the sale of products during the show. It’s a family friendly technology show with a fair-like atmosphere. And it’s settled in Germany’s political and cultural capital.

If you’ve always wanted to attend CES but never got in, give IFA a shot.

What Will MakeUseOf Cover at IFA?

This year, MakeUseOf is sending James Bruce, Ian Buckley, Joe Coburn, James Frew, Tina Sieber, and Kannon Yamada to report from IFA. As members of the press, we get early access to the convention and will start reporting on September 4.

You can expect us to visit and report from the following exhibitions:

  • Home & Entertainment Electronics
    Traditionally, LG Electronics, Samsung, and Sony occupy entire halls at IFA, and we’ll be sure to pay them a visit.
  • Audio Entertainment
    The audio brands exhibiting at IFA include Audio-Technica, Pioneer, and Sennheiser, to name a few. Bose is not represented, this year.
  • My Media
    The halls dedicated to media will include booths of Acer, Asus, DJI, Lenovo, Logitech, Microsoft, and Razer.
  • Communication
    Amongst a selection of lesser known European companies, we’ll definitely check out Huawei.
  • IFA NEXT
    This is IFA’s segment for innovative products and start-ups. We’ll explore the latest robots, drones, and other novelties.
  • Home Appliances
    We don’t typically cover white goods at MakeUseOf, but this year we’ll explore how home appliances are tying into smart home systems.

One brand consistently skipping shows like CES and IFA is Apple. While they’re not exhibiting, we expect their omnipresence, along with Google and Amazon (Alexa).

Apart from the brands mentioned above, we’ll keep our ears open and our eyes peeled for exciting announcements, puzzling features, and cool gadgets. Bookmark this article or our IFA coverage page and check back between September 4 to 8 to see the best new products we’ve discovered at IFA.

Read the full article: For the Latest in Tech Trends, Follow MakeUseOf’s IFA Coverage (Sep 4-8)


Read Full Article