Connect with us

Tech

Here's why Amazon and Google are fighting to be 1st in your home – CBC.ca

Published

 on


From Black Friday sales to Boxing Day blowouts, Amazon and Google are offering their smart speakers at a deep discount — even giving them away, in Google’s case. 

But don’t be fooled by the corporate giants appearing to get into the holiday spirit. They’re looking to get you invested in their tech ecosystem and even help train their artificial intelligence.

“They’re all trying to shape your behaviour by giving you these speakers for virtually nothing,” said Marvin Ryder of McMaster’s DeGroote school of Business. He says it reminds him of when Apple debuted its iPods in the early 2000s. 

“I used to argue they should give it away because they’d sell it for $199 but it holds 5,000 songs. If I send you to the iTunes store you’ll spend $5,000 loading it, so why don’t [they] give it to you for free?”

He says the same goes for the Google Nest and Amazon Echo and Alexa devices. To get the full experience, you need more than one.

“Once you’ve got one in your home, your temptation is to go deeper and deeper and deeper,” said Ryder. “I can make more money from you down the road — whoever gets there first is going to mine you the most.”

Amazon dominates smart speaker market

Right now, said Ryder, Amazon dominates, with almost 70 per cent of the market. Google has 25 per cent, with Apple and others fighting for the final five per cent. 

“So if I can get into the home — however I can get that first thing into the home, it’s probably worth it because I can make my money down the road” said Ryder, through app purchases and add-ons like smart thermostats and video doorbells.

They also tap into millions of artificial intelligence trainers, who all contribute in their own way to machine learning and making a better, more reliable product says Graham Taylor, an AI expert who has worked on Google’s Brain Team. He’s now the Canada Research Chair in Machine Learning. 

“This data is pretty valuable,” said Taylor. “The obvious value is in marketing you goods, controlling the retail channel and being able to send you ads for things it thinks you need. But the other less obvious one is around improving the system for all users.”

An ‘essential ingredient’

It comes back to a basic principle of machine learning: you’re interacting with examples provided by a system, and teaching it to learn from those examples. 

“You’re supplying an essential ingredient for the training,” said Taylor. “It’s not just the raw data, it’s the labelling and putting that data in context — something we still really rely on people to do.”

For its part, Google says its goal with the Google Nest Mini giveaway was “to reward YouTube Music subscribers. Nest Mini is a small smart speaker with big sound, making it a fun way for [YouTube Music] subscribers to listen to their favourite songs, hands free.”

A spokesperson didn’t directly respond when asked to what extent user content is used to advance its machine learning, but did say its devices would only send audio to Google when keywords like “Hey Google” are used, and you can delete any of those recordings through the Nest app. 

It also said the microphone can be turned off using a physical button on the device and pointed to its Google Nest privacy principles for more details. 

Let’s block ads! (Why?)



Source link

Continue Reading

Tech

Apple's voice-only Music subscription could boost Siri's accent understanding – TechCrunch

Published

 on


Apple had a slew of interesting announcements at its event on Monday. But one that stood out to me — and I feel didn’t get as much attention — is the new pricing tier of Apple Music. A new “Voice” tier will offer the entire Apple Music library to subscribers at a reduced rate of just $5 per month: The catch is you have to use Siri to access it, eschewing the standard Apple Music visual and typing-friendly in-app user interface.

Apple didn’t share why it is launching this plan, but I think it’s reasonable to speculate that the iPhone-maker is lowering the price barrier and persuading more people to use Siri because it wants to gather more voice data to train and improve its voice assistant.

“We’re excited that even more people will be able to enjoy Apple Music simply with their voice,” Apple chief executive Tim Cook said at the event.

I can’t imagine any other compelling reason why the Apple Music Voice plan exists, especially since Apple is likely offering the new service with much lower margins than the standard plan, as the licensing agreements with labels remain the same to offer up the entire Apple Music catalog.

Again, this is just speculation, but I think given the stiff competition between Apple and Spotify, if the Swedish firm could offer its streaming service at $7-8 a month to beat Apple Music at price, it would. And Apple is taking some loss with the new subscription tier because it really wants to gather vast amounts of data. When I tweeted this theory, my colleague Alex wondered aloud why wouldn’t Apple just make the subscription free? I suppose Apple, a $2.5 trillion company, can technically swallow that much of a hit on the balance sheet, but it doesn’t want to attract more criticism from standalone music streaming firms such as Spotify. It’s already facing scrutiny for anti-competitive behavior on a number of fronts.

Tech firms feed their AI models with vast amounts of data to improve the services’ capabilities. Even as Siri has considerably improved over the years, the general consensus among many people who work in tech and the masses alike is that Amazon’s Alexa and Google Assistant are far superior.

It’s likely that Apple has already been gleaning such voice data from existing Apple Music users, but as a friend suggested, “the point is this — this feature always existed. It’s just that they’d put a high paywall. They’ve lowered that wall now.” In addition to lowering the barrier to entry, making Music voice-only via the new plan means people have to engage with Siri to make use of it; Siri is a feature for standard Apple Music subscribers, but it’s highly likely that most users primarily or exclusively access the content via the app’s UI.

If you want an example of what can happen to voice-powered assistants when you require that users treat it as a voice-first or voice-only service, look at Amazon’s Alexa. Out of the gate, Alexa had to be accessed by voice. This allowed Amazon to not only collect massive amounts of training data for its Alexa algorithms, but also helped train users about how to use it to maximum effect.

Understanding accents and dialects

Another reason why I think my theory works is the markets where Apple plans to offer this new subscription tier first: Australia, Austria, Canada, China, France, Germany, Hong Kong, India, Ireland, Italy, Japan, Mexico, New Zealand, Spain, Taiwan, the United Kingdom and the United States.

Having India, Spain, Ireland and France in the first wave of nations suggests that Apple is looking to amass a wide-range of dialects and accents from across the globe. On a side note, voice search is very popular in many markets, including developing nations such as India, and in markets like China and Japan where text input can sometimes be unnecessarily complex versus spoken word. (A Google executive told me once that the surprising mass adoption of voice searches in India, the world’s second-largest smartphone market and where Android commands about 98% of the pie, helped the company improve Google Assistant and prompted more aggressive approach to innovate on the voice front.)

Siri is often framed as a bit of a laggard in terms of its competence versus the rest of the voice assistant competition, and Apple’s latest move in services could be an attempt to help it close the perceived gap, while offering customers a discounted way to onboard to its music streaming service.

Adblock test (Why?)



Source link

Continue Reading

Tech

PSA: the MacBook Pro 14-inch’s $20 power brick upsell is probably worth it – The Verge

Published

 on


If you’re looking at buying the $1,999 base model MacBook Pro 14-inch, there’s one upgrade that you may really want to make — the $20 one that gets you the 96W power adapter instead of the 67W included power adapter. That’s because, according to some wording on Apple’s MacBook Pro configuration page (spotted by MacRumors), you’ll need the more powerful charger if you want to take advantage of the computer’s fast charging feature, which can charge the laptop up to 50 percent in half an hour.

Is it ridiculous that Apple is basically taxing the people who want to buy its least expensive (but still very pricey!) new MacBook Pro? Yes, absolutely — but you should still probably pay it if you want to charge your laptop up quickly. The exception is if you already have a charging brick capable of 100W USB-PD power delivery: Apple tells The Verge that you can fast charge via Thunderbolt as long as your power brick provides enough power. If you already have a beefy power brick, you can skip the upsell.

I know it probably doesn’t feel great to encourage Apple’s nickel-and-diming, but if you want fast charging, this will likely be the best way to get it. There may be, somewhere in the world, a 100W USB-PD charging brick that sells for $20, but there’s no way I would trust it enough to charge a very expensive computer. (If it was $20, I might not even trust it not to burn down my house). I’d pick the upsell.

The one silver lining is that this is only a problem on the base 8 CPU core / 14 GPU core model — if you do any processor upgrades, you’ll get the 96W brick for “free.” Please note, though, that upgrading just the RAM and/or storage on the base MacBook Pro won’t get you that upgrade, but if you’re in the configuration screen anyways, you should absolutely check that box unless you hate fast charging.

Adblock test (Why?)



Source link

Continue Reading

News

Facebook to pay up to $14.25 million to settle U.S. employment discrimination claims

Published

 on

Facebook Inc has agreed to pay up to $14.25 million to settle civil claims by the U.S. government that the social media company discriminated against American workers and violated federal recruitment rules, U.S. officials said on Tuesday.

The two related settlements were announced by the Justice Department and Labor Department and confirmed by Facebook. The Justice Department last December filed a lawsuit accusing Facebook of giving hiring preferences to temporary workers including those who hold H-1B visas that let companies temporarily employ foreign workers in certain specialty occupations. Such visas are widely used by tech companies.

Kristen Clarke, assistant U.S. attorney general for the Justice Department’s Civil Rights Division, called the agreement with Facebook historic.

“It represents by far the largest civil penalty the Civil Rights Division has ever recovered in the 35-year history of the Immigration and Nationality Act’s anti-discrimination provision,” Clarke said in a call with reporters, referring to a key U.S. immigration law that bars discrimination against workers because of their citizenship or immigration status.

The case centered on Facebook’s use of the so-called permanent labor certification, called the PERM program.

The U.S. government said that Facebook refused to recruit or hire American workers for jobs that had been reserved for temporary visa holders under the PERM program. It also accused Facebook of “potential regulatory recruitment violations.”

Facebook will pay a civil penalty under the settlement of $4.75 million, plus up to $9.5 million to eligible victims of what the government called discriminatory hiring practices.

“While we strongly believe we met the federal government’s standards in our permanent labor certification (PERM) practices, we’ve reached agreements to end the ongoing litigation and move forward with our PERM program,” a Facebook spokesperson said, adding that the company intends to “continue our focus on hiring the best builders from both the U.S. and around the world.”

The settlements come at a time when Facebook is facing increasing U.S. government scrutiny over other business practices.

Facebook this month faced anger from U.S. lawmakers after former company employee and whistleblower Frances Haugen accused it of pushing for higher profits while being cavalier about user safety. Haugen has turned over thousands of documents to congressional investigators amid concerns that Facebook has harmed children’s mental health and has stoked societal divisions.

The company has denied any wrongdoing.

In Tuesday’s settlements, the Justice Department said that Facebook used recruitment practices designed to deter U.S. workers such as requiring applications to be submitted only by mail, refusing to consider American workers who applied for positions and hiring only temporary visa holders.

The Labor Department this year conducted audits of Facebook’s pending PERM applications and uncovered other concerns about the company’s recruitment efforts.

 Facebook is not above the law,” U.S. Solicitor of Labor Seema Nanda told reporters, adding that the Labor Department is “committed to ensuring that the PERM process is not misused by employers – regardless of their size and reach.”

 

(Reporting by Sarah N. Lynch; Editing by Will Dunham)

Continue Reading

Trending