Google Illuminate - creates a radio interview from a research paper

Google Labs has a long history of inviting users to experiment with cutting-edge tech. Gmail was once a private beta project. Illuminate is a project that turns academic papers into AI-generated audio discussions in the style of an NPR podcast. The idea is simple: Google's LLM Gemini generates a paper summary and a Q&A. Two AI-generated voices, a male interviewer and a female expert, will guide you through a short interview describing the paper. You can listen to some of the samples on the Google Illuminate website. This is useful, letting me listen to engaging summaries of the ever-growing stack of research papers I must read as I exercise or drive. It can also be easily adapted to other narration forms for different use cases. Illuminate is in private beta, and you can join the waitlist here.

A good AI story from Google DeepMind

 

GraphCast's forecast for New Zealand Sun 19 Nov

In recent months, we've become used to news stories declaring AI poses humanity an existential threat, that a superintelligence "whose values do not align with ours" may exterminate us all. So, it's nice to see a good AI news story. Yesterday, a team at Google's Deepmind published a paper in Science, Learning skilful medium-range global weather forecasting. They have trained a deep learning model on publically available global historical weather data. They show that their model makes better weather predictions "much faster than the industry gold-standard weather simulation system – the High-Resolution Forecast (HRES), produced by the European Centre for Medium-Range Weather Forecasts (ECMWF)".

Moreover, their system, GraphCast, is fast. They say, "While GraphCast's training was computationally intensive, the resulting forecasting model is highly efficient. Making 10-day forecasts with GraphCast takes less than a minute on a single Google TPU v4 machine. For comparison, a 10-day forecast using a conventional approach, such as HRES, can take hours of computation in a supercomputer with hundreds of machines."

In January, Auckland, New Zealand's largest city, received a year's worth of unforecasted rain in a single afternoon, causing widespread flooding. Cars floated down streets, road were washed away, houses slipped down hillsides, and people died. Better weather forecasting can help prevent this. You can try out GraphCast's 10-day forecast here. I'm going to Waiheke Island on Sunday. It looks like the weather will be Okay.

Classic album covers in Google Street View

We're seeing some increasingly creative ideas for using Google Street View and The Guardian has recently shown us another one - Classic album covers in Street View. This features the album cover superimposed over it's location; from the obvious Beatles' Abbey Road to more modern covers, this is another creative use of Street View.

Time travel with Google Street View

I always knew Google would do this, it was so obvious. Google Street View now has a new feature that lets you go back in time with Street View to see how any location looked right back to when Google's cameras first captured the view. Now when you start Street View you'll see a time stamp on the pop-up Street View window and a time-line with a slider. You can select from any of the points on the slider.
Your street, like mine, probably hasn't changed much. Time magazine has put together a series of time-lapse sequences that show iconic buildings like NYC's One World Trade Center rising out of the ground over the years. This feature will be fascinating to explore in 20 to 30 years time. 

Google Acquires AI Startup DeepMind For More Than $500M

Google will buy London-based artificial intelligence company DeepMind. TechCrunch reports that the acquisition price was more than $500 million, and that Facebook was also in talks to buy the startup late last year. DeepMind has confirmed the acquisition, but hasn't disclosed the deal terms. Some pundits have said that this must be a counter to Apple's Siri, but honestly I think Google already has enough AI smarts to counter Siri. To me this look as if Google sees a could synergy between DeepMind and its own skills and wants to prevent competitors acquiring DeepMind's expertise.

Sweet solution? Google tests smart contact lens for diabetics

The Guardian reports that diabetics could in future avoid painful pinprick blood glucose tests using a smart contact lens being developed at Google which measures glucose levels in tears.

A prototype shown off by the company uses an embedded miniaturised glucose sensor and wireless chip in a contact lens to measure glucose in tears as often as once every second.
That would make monitoring glucose levels - an essential task for diabetics, who are at risk of heightened or lowered blood-sugar levels - faster, easier and less painful because it would remove the need to break the skin and measure blood glucose levels directly.
Diabetes, caused by a deficiency of the sugar-regulating hormone insulin, affects around 5% of people in developed nations. It is a leading cause of kidney failure, blindness and amputations if insufficiently treated.
French scientists discovered that tears carry measurable levels of glucose in the 1930s, but it has taken decades to find a way to exploit that to create a non-invasive test. A team at the University of Michigan showed off a sensor which tested tears in November 2011, and were even then one of a number pursuing the idea. Microsoft Research, then collaborating with Babak Parviz - at the time a professor at University of Washington - unveiled a prototype of the glucose-measuring contact lens in 2011.
Since then Parviz has moved to join Google's experimental X Lab, for which this is its latest product. "We're testing prototypes that can generate a reading once per second," said Brian Otis and Parviz, co-founders of the project which they started at the University of Washington, in a joint blog post. "We're also investigating the potential for this to serve as an early warning for the wearer, so we're exploring integrating tiny LED lights that could light up to indicate that glucose levels have crossed above or below certain thresholds."
Google is currently talking to the US Food and Drugs Administration, which regulates drugs and medical applications, about selling it as a medical device, and seeking partners among medical experts to bring a smart contact lens and apps to market.
"We've always said that we'd seek out projects that seem a bit speculative or strange, and at a time when the International Diabetes Federation is declaring that the world is 'losing the battle' against diabetes, we thought this project was worth a shot," Otis and Parviz said of Google X labs, which also developed Google's self-driving cars and Google Glass, the company's smart glasses. I wonder if Google's long term interest in contact lenses may be as an extension to the Google Glass project?

#Apple's wise maps decision

A year ago Apple was taking a lot of criticism for its decision to kick Google Maps of the iOS platform. I in a post titled "Apple's mapocalypse" took a different view - for competitive reasons Apple had to develop its own mapping and location services. It now seems that Apple (and I) were right; the Guardian in an article titled "Apple maps: how Google lost when everyone thought it had won" shows that a year on Google has lost tens of millions of Google Map users to Apple in the US alone.

Would you eat a Google burger?

You may have seen news reports recently of the first hamburger made from synthetic meat grown in a laboratory. It turns out that Google co-founder Sergey Brin financed this research. Watch the video below to find out why.

IP - "Intellectual" or "Internet" Property

The New Zealand Herald has reported an interesting story that sheds light on the ongoing modern struggle that established copyright and intellectual property laws have with the Internet and the digital distribution of content that may affect you. Do you use the wonderful free media player VLC? I certainly do - it's installed on all four of the computers I regularly use and it is (IMHO) the best video player available. If VLC can't play that file, nothing will. But US media giant HBO has  sent a takedown notice to Google, listing VLC amongst other infringing content they want the search giant to remove from its search results.
   This seems bizarre, VLC isn't infringing anyone's copyright or intellectual property, moreover as open source software nobody could be profiting even if it were. It's as if some lawyer wanted to confiscate your spectacles or contact lens to stop you watching a pirated movie. Has HBO gone mad? Apparently not, it appears that VLC is guilty by association since it often turns up mentioned online around pirated content. HBO's robo-lawyers automatically issue take-down notices and Google is supposed to automatically respond. Fortunately in this case Google has ignored HBO's request. VLC has been tainted by the company it keeps; pirates use VLC therefore VLC must be stopped. VLC runs on computers, perhaps that should be their next target!

Google project Loon over New Zealand

I'm not quite sure why Google chose to test its new X Project, called Project Loon, in New Zealand - but the view is stunning. The idea sounds crazy - WiFi carried by high altitude ballons to provide Internet access to remote areas of the planet - perhaps "loon" is short for "loony." But crazy or not the idea is being tested in the South Island of New Zealand near Lake Tekapo. The ballons fly twice as high as commercial jets and can even be steered a little by increasing or decreasing their altitude to move them into different wind patterns. The idea is to provide constant coverage to remote parts of Africa and Asia without the need for expensive and hard to install and maintain ground infrastructure. Google is looking for "pilot testers" in New Zealand so if you're interested in testing Loon you can sign up here. Needless to say the media reporting this story couldn't resist a few sheep jokes. No hobbits were mentioned though.

Project Loon from Google - Balloon-powered Internet from Trey Ratcliff on Vimeo.

Has Facebook lost its way?

We tend to assume that successful companies like Facebook get large, wealthy and powerful by making better decisions than everyone else. Consider Facebook's recent decision to buy mapping company Waze for a rumoured $1 billion and to hire ex-Apple mapping team leader Richard Williamson. Clearly Facebook has decided it needs its own mapping functionality and not rely on Microsoft's Bing, in it's desktop product, or Apple's maps on iOS and Google on Android. Is this a good decision? Well on the face of it yes, but really Facebook seems to be the last person in the room to wakeup and smell the coffee. As I noted when discussing Apple's mapocalypse back in September 2012 - location services are going to be really important in the future. Apple couldn't afford to gift this market to Google and Facebook can't either.
   Google became interested in maps back in 2004, when it bought the Australian company Where 2 Technologies. Now this is an example of a good strategic decision. Our computing devices weren't even very mobile in 2004 but Google invested heavily in mapping - this implies great long term vision. Apple's lauded leader Steve Jobs actually made a very poor strategic decision in 2007 when the iPhone launched. Google Maps should never have been invited on board; Apple should have had its own mapping service from the start. This they belatedly corrected in 2012 because they realised they had to.
    So where has Facebook been? Have they all been too busy playing Farmville! If Facebook really wants users to spend all their time online in Facebook, and with Facebook Home that is exactly what they want you to do. Why did nobody in their Menlo Park HQ think "hey guys, we like really need our own mapping app." This troubles me. The Facebook execs clearly aren't always that smart.

It's nice to be right

Back in February this year, and earlier in September 2011, I wrote a couple of blog articles titled "Are you selling technology or services?" These posts put forward my opinion that Apple's strategy of tightly integrating its hardware and software had a vital third component; namely the services that people used on their iDevices. I wrote that Amazon seemed to be adopting this approach with the launch of its Kindle Fire. The Software Engineering Services Blog posted last week a piece called "Suddenly everyone wants to follow Apple's integrated hardware-software model," which describes how Google,  Samsung and Microsoft are now also trying to follow Apple's lead  by offering a complete ecosystem of hardware, software and services. The blog post concludes by saying "Steve Jobs may be long gone, but his vision lives on and everyone suddenly wants a piece of it, but just because they each recognize the magic behind Apple's strategy doesn't mean each can successfully copy it or that the market will follow. And that is the real challenge these companies face." It's nice to make a prediction that comes true.

What would you do if you had Glass? #projectglass

Google is inviting people to take part in the Google Glass project. I've blogged about the Google Glasses before and now Google is inviting anyone with a good idea for how to use the augmented reality glasses to tell them about it. The winners will be able to become a Glass Explorer and I assume build their application for use on the glasses. There is a catch, you have to be in the US and have $1,500 (USD) to partake. The deadline for entry is Feb 27. Visit the Glass Project website for more information and read this article from the Verge about the development of Glass and what using one/them is like or watch the video.

Peter Norvig - has the best job in the word

In a recent interview in the Guardian Peter Norvig, Google's director of research, say he has "the best job in the world." In the interview he talks about the future of Artificial Intelligence, mapping, personal computer, and Google's plans for the future. He's extremely well qualified to talk about AI since he co-authored the standard AI text book. If you've taken an AI course, including mine, Artificial Intelligence: A Modern Approach is the recommended text. 

Google "crisis map" for Hurricane #Sandy

Hurricane Sandy looks like it will be the most closely watched storm in history. Google has set us a "crisis map" for Hurricane Sandy to help people who may be in its path. The map features various overlays including: the storm's predicted path, wind speeds, predicted storm surge, public alerts, evacuation routes, refuge centres and more. Of course as the power goes down across the North Eastern seaboard of the US WiFi and broadband will be lost as well. Cell phone towers can work for a time on battery power, so 3G may still be available, but I'd recommend getting any digital information you may require sooner rather than later - the Internet isn't storm proof.

Van Gogh Meets Alan #Turing

Google developers present a video that showcases projects they have been working on that merge art and technology - where Van Gogh meets Alan Turing.

AI - Apple's next killer product

Eric Jackson in a post in Forbes makes the case that Artificial Intelligence "will be the horsepower that bridges the gulf" between our expectations of what apps can do and their current performance. He argues that Apple has "the time, the right AI people, data and money" to make this happen. I would argue that Google is in a similar position; after all Google's Director of Research, Peter Norvig, co-authored the standard AI teaching text (AI: A Modern Approach), and Google has both data and money.
    Despite this though I don't agree with Jackson's thesis. When AI works it becomes invisible, people are unaware of it, therefore it will not be the "next killer product." AI may well make the next generation of smart devices work better but most users will be unaware of it.

#Apple's mapocalypse

Well that's what some journalists are calling it. By a coincidence just a few days ago I blogged about how much effort Google has put into its mapping service. not just effort in terms of satellite imagery and street view, but physical human effort in terms of correcting mistakes and ensuring the overlaid information is correct. Thus, it's no surprise if Apple is finding it hard to hit the ground with a product of similar quality in iOS 6.
    So many people are asking why did Apple remove Google Maps from iOS 6? The answer is simple. Providing location services is so important going into the future that Apple could not abandon this segment of the market to Google. It has to become a major player in mapping and the provision of location sensitive data and services. If that means Apple has to spend a few billion dollars and lag behind Google for a few years it's a price worth paying - mapocalypse now or armageddon latter, that was the choice.

A look inside Google's "Driverless Car"

I've blogged about Google's "Driverless Car" before and the technology, and its implications, features in the last chapters of The Universal Machine. A great new video has appeared on YouTube that shows what the Google cars are capable of and looks inside the technology.

Google AI loves cats

Last week the media was full of stories about a new Google X Lab project that has created an AI which seems to love cats. I couldn't avoid the story because friends kept emailing me links to it - this story in the Financial Times is fairly typical. 
    Google built a huge neural network using 16,000 computer processors to see what it would learn when exposed to 10 million clips randomly selected from YouTube videos. There are basically two types of machine learning; supervised when you say "here's an example of X," "here's an example of Y" and "here's another example of X," and unsupervised learning where there is no instructor. Google's system was unsupervised, it just looked at all the YouTube clips and tried to find interesting patterns. It did - cats! Google's system can now look at a YouTube clip and tell you, with some certainty, if there is a cat in it or not.
   Before we leap to conclusions that AIs like cats or want pets first consider an old experiment conducted with neural nets for the Pentagon. They wanted to find Russian tanks in spy photos; so using supervised learning they showed a neural net hundreds of photos, some of which had tanks in and some which did not. After training the neural net could, again with some confidence, identify tanks in photos it had never previously seen. Success they thought. Later they discovered that most of the photos they had of tanks were taken on cloudy days, whereas most of the photos of countryside without tanks were sunny days. The computer had learnt to see if it was cloudy or sunny - the tanks were a coincidence. Google's network may recognize cats but not as we do.
   Actually the cats were a by-product, Google system can recognise 20,000 different things in the YouTube clips.